This is the latest excerpt from AGNT: The Orchestration Economics Manifesto - An Investment Framework for the Agentic Era. Each Thursday, I explore a major theme of the Manifesto and unpack the frameworks, adding extra context with more recent developments. Note: The figures and sequential references are taken directly from the larger Manifesto.

Subscribe now

Parts I and II established that the paradigm has shifted. Machines became actors. The intelligence required for autonomous professional work exists as commercially available infrastructure. Agents are deployed in production. The infrastructure, while strained, is scaling.

The question now: Who captures the value it creates?

When a factor of production commoditizes, when intelligence becomes abundant, then value migrates to whoever controls the coordination of the newly abundant input. This migration is a pattern that has repeated across every major economic transition, from electricity to computing to cloud infrastructure. The entity that coordinates the abundant resource captures the surplus. In the agentic economy, the entity that coordinates machine actors captures the surplus of the entire transition.

The framework that describes who captures value, under what conditions, and why is Orchestration Economics. This manifesto defines the Three Laws of Agentic Value that determine whose orchestration position is durable. They are structural in nature, already visible, measurable, and investable.

However, we must first clarify how the new layer of value is formed.

Share

Orchestration Economics

Orchestration Economics rests on a fundamental structural claim: The entity that controls the orchestration of machine actors captures the surplus of the transition. This is not a marginal improvement over SaaS economics, nor is it limited to software. It is a categorical shift in three dimensions:

Pricing is outcome-based, not seat-based or time-based. The unit of commerce is a claim processed, an analysis completed, a route optimized – not a license sold or a billable hour. Pricing per agent task (i.e., per consumption) is an intermediary step in that journey.

Budget is labor, not IT. The vendor does not compete with software for a slice of the 3-5% IT line. It competes with salaried professionals for a slice of the 25-40% knowledge-work line. The addressable market expands by a factor of 5-10x. What is more, the market is not limited to labor substitution; agents complete value-accretive tasks that did not exist before.

Defensibility is topological, not feature-based. What protects an orchestrator is the structure of the workflow it controls, not the capabilities of any single agent inside it.

The rest of this chapter establishes where this layer sits in the firm's new architecture, how value migrates into it, why most companies will fail to capture it even when they think they are, and under what conditions an orchestration position is durable rather than temporary. The Three Laws formalize those conditions.

The New Layer

Before agents, the architecture of knowledge work was straightforward. A human expressed intent: process this claim, close this deal, optimize this route. The human also had the tools to execute that intent. The human opened a CRM to update a record, a spreadsheet to run an analysis, and an email client to send a message. Each tool was a destination. The human was the orchestrator. The tools were the instruments. The business outcome was the product of human judgment applied through software.

Agents invert this architecture fundamentally. They insert a new layer between humans and tools. Humans used to go to the tools. Now the tools come to the human, mediated by the agent.

In the simplest case, this involves just a single AI agent. The human expresses intent: update the Chicago deal to closed-won and schedule a kickoff with their team next week.” The agent receives that intent, determines which systems to invoke (the CRM, the calendar, the email system), executes the necessary actions across those systems, and delivers the outcome. The human never opens the CRM. Never touches the calendar. Never drafts the email. The agent coordinates it.

Even in this single-agent case, orchestration is happening. The agent is orchestrating tools on the human’s behalf. The agent is, in the precise sense of the word, a synthetic colleague. This is the foundation of Orchestration Economics: agents are not better tools. They are a new layer between humans and their tools, designed to produce business outcomes.

The entity that controls the “Orchestration Layer” owns the relationship between the human and the work for a specific workflow, or set of workflows, aimed at producing a business outcome. It decides which tools are invoked, which data is accessed, which sequence yields the best outcome, which systems are essential, and which are bypassed entirely.

The Orchestration Layer described in this Manifesto is categorically different from the technical Orchestration Layer known as the “Harness”. This is a critical distinction. The Harness enables an LLM to produce results and makes intelligence operational. The Orchestration Layer turns operational intelligence into a business outcome.

The Architecture of the Agentic Firm

The agentic firm is organized in concentric rings, not layers, because what matters is not just function, but proximity to value capture. The farther out the ring, the greater the control over outcomes, and the stronger the economic position:

Ring 1 – Intelligence. The foundation models. The substrate. Its economic unit is the token.

Ring 2 – Harness. The technical orchestration infrastructure that makes intelligence operational: the systems that decompose goals, delegate to agents, manage state, coordinate execution, and accumulate learning across sessions. Claude Code, OpenAI Codex, enterprise agent platforms, managed agent systems, and emerging agent operating environments are Harness products.

Ring 3 – Orchestration. The outermost sub-layer. The position occupied by the entity that places Rings 1 and 2 at the center of its operations and adds what neither can produce: the operational context accumulated through running the world. Its economic unit is the outcome.

When I first argued in 2025 that the durable moat for LLM labs lay in orchestration rather than model intelligence, I was describing Ring 2. The labs confirmed the thesis. They recognized that model capability alone was commoditizing and moved into the infrastructure that makes it operational. The Coding Wedge was the entry point. Each product since has deepened the position.

But the labs did not stop at the Harness. Cowork extended into knowledge work. Vertical plugins reached into business functions. Frontier positioned itself as a control plane for the enterprise. With each step, the labs move outward, from intelligence through Harness toward the outermost sub-layer, contesting it against companies that have occupied it for decades. This expansion from Ring 2 to Ring 3 is at the root of the SaaSpocalypse and the broader value destruction in the public markets since the start of the year. We explore this in Chapter 13 infra.

An entity that provides intelligence is not an orchestrator. An entity that builds a Harness is not an orchestrator. The orchestrator wraps around both and produces business outcomes.

A single entity can occupy more than one ring. Two patterns are already visible.

The debate is no longer whether labs will move into Ring 3, but where they can do so successfully.

Anthropic today credibly sits in Ring 1 (Claude), Ring 2 (Claude Code, MCP, Skills, Agent SDK), and extends into Ring 3 (Claude Code for enterprise, Cowork, managed agents). In narrow domains, coding is the clearest because it already crosses all three. OpenAI is pursuing the same stack. The labs built inward-out: from intelligence, through Harness, toward outcome.

Occupying the inner rings does not confer the outer ring. Ring 3 is not a further technical layer. It is the operational context co-produced with the workflow being orchestrated. The labs can ship Ring 2 unilaterally. They cannot ship Ring 3 unilaterally because the intent, context, and workflow reside within the firm whose work is being orchestrated.

In domains where the lab also owns the workflow, such as coding, where the developer’s intent, context, and workflow are all captured natively inside the Harness, the lab credibly holds Ring 3. In domains where it does not, such as insurance underwriting at a carrier, credit analysis at a bank, claim evaluation, or at a reinsurer, then the lab supplies Rings 1 and 2, and someone else holds Ring 3.

Ring 3 is, therefore, domain-specific, not firm-wide. The same entity can hold it in one workflow and lose it in another. Whether any entity durably holds Ring 3 is not settled by its stack position. It is determined by the Three Laws.

The architecture is new. The contest for who occupies the outermost ring is already underway.

Captain and Crew: The Topology of Coordination

For a simple task such as updating a CRM record or scheduling a meeting, a single agent coordinating a few tools is sufficient. But evaluating an insurance claim, analyzing a company’s creditworthiness, or optimizing this supply chain becomes too complex for a single agent to handle all dimensions.

This is where multi-agent systems emerge. The system has a crew of agents with specialized capabilities whose job is to execute well-scoped work, such as document extraction, risk modeling, fraud detection, pricing, or compliance validation. They can be brilliant within their domains. But they do not see the whole map. They do not decide which workflow to run. They do not own the relationship with the human who asked for the outcome.

A multi-agent system could, in principle, be organized in many ways. A federation of peer agents could negotiate amongst themselves. A swarm could converge by consensus. Independent specialists could run in parallel and average their outputs. All of these have been tried. None produces coherent work at enterprise reliability.

A Google-DeepMind-MIT study, first published in December 2025, formalized this structure. A multi-agent system is an agent system with more than one reasoning entity, where agents interact through a communication topology and an orchestration policy that defines how the system makes decisions: how sub-agent outputs are aggregated, whether the orchestrator can override sub-agent decisions, whether memory persists across rounds, and when the task is complete.

According to the most recent April 2026 revision of the paper, centralized coordination architectures exhibited approximately 75% lower trace-level error amplification than independent topologies (4.4× versus 17.2×). The broader result was not simply fewer propagated errors, but the emergence of verification mechanisms that improved reliability when tasks were naturally decomposable.

Only the architecture where one agent receives the task, decomposes it, selects and validates specialists, and decides what happens next produces the coherent outcomes enterprises will pay for.

The orchestrating agent becomes the captain. It does not perform the specialized work itself. It decomposes the goal into subtasks. It assigns those subtasks to the appropriate specialists on the agentic crew. It collects their outputs. It validates them. It synthesizes the results. The human talks to the captain. The captain commands the crew. The crew does the work. The captain reports back to the human. That is the agentic AI loop.

As we move into the core exploration of Orchestration Economics, it is worth noting that the discussion of the transition to the Agentic Era has remained at a simplistic, binary level. Winners and losers. Eat or be eaten. Victim or beneficiary of the SaaSpocalypse. The more nuanced and challenging question most companies will be facing: Do you want to be the captain or the crew?

The economic consequence follows directly from the topology, not from the metaphor.

Specialist agents are substitutable. A better extraction model can be swapped in. A better pricing model can replace the incumbent. The crew is upgradeable. Its members have limited pricing power because the captain can always source a better one.

The captain is not substitutable. Replacing the orchestrator means rewiring the workflow, including remapping the topology, retraining the policy, reestablishing every integration, and re-accumulating the operational memory. The switching cost is categorical.

Value is multiplicative, not additive. A single agent coordinating three tools yields linear value: the sum of the tasks it automates. An orchestrator coordinating five specialist agents, each coordinating their own tools, produces nonlinear value. The orchestrator can redesign the workflow, parallelizing steps, eliminating redundancies, and catching errors across domains that no individual specialist would detect. The orchestrator is not a participant in the workflow. It is the workflow.

In human organizations, we call this management. In multi-agent systems, we call it orchestration. Value accrues to the Orchestration Layer rather than being evenly distributed among the agents beneath it. Not because someone has to be in charge, but because the topology in which centralized coordination wins is the only topology in which the workflow works at all. The entity at the center of that topology captures the surplus that coherence creates.

Let’s follow the money through an actual enterprise workflow

A commercial insurance submission arrives. Under the old architecture, it enters a human chain: intake, review, coding, investigation, pricing, approval, and documentation. Each human is an expert. Each has trained for years. Together, they process the claim in days, sometimes weeks, at costs measured in hundreds of dollars per claim. The Orchestration Layer lived in a manager’s brain and a project management tool. The bottleneck was human capacity for coordination.

Under agentic orchestration, the same workflow transforms. The submission arrives at a master agent. The master agent does not “process a document” as legacy software does. The master agent interprets the goal: evaluate this risk and produce a quote that aligns with our growth and risk constraints. It inspects the case. It decides which specialized agents to invoke: extraction, enrichment, risk modeling, pricing, or compliance validation. It sets their inputs. It sequences their work. It validates their outputs. It determines when the goal is achieved.

The specialist agents are valuable. They have real reasoning capacity. But they are no longer where the workflow is decided. Their job is to respond to the plan. Not to define it. The control plane has moved. So has the economics.

The question every company in the agentic economy must answer is not rhetorical. It is operational.

Are you the captain? Or are you the crew?

The Programmable Firm

Carnegie Mellon and Stanford University research teams released a landmark study in October 2025 called “How Do AI Agents Do Human Work?” that provided an empirical foundation for the Agentic Era. The teams compared how humans and AI agents complete identical work across data analysis, engineering, computation, writing, and design. They concluded that for tasks defined as “readily programmable,” AI agents will complete the work 88.3% faster at 90.4-96.2% lower cost. These figures represent benchmark environments rather than fully operational enterprise deployments.

The key finding is not that work is programmable end to end. It is more consequential: programming is embedded in 82.5% of knowledge-work occupations. In practice, this means a large share of enterprise workflows contain programmable components, even if they are not fully automatable.

Programmability is the unlock. A programmable workflow can be agentized. A workflow that can be agentized can be orchestrated. A workflow that can be orchestrated can be sold as an outcome rather than licensed as a tool. The economic category changes not because the AI is better, but because the work itself has always been more structured than the software market has reflected.

The gate is reliability. Every enterprise workflow has a reliability requirement: a level of accuracy and completeness below which human oversight remains necessary. When an AI system handles 40% or 50% of cases correctly, it accelerates human workers but cannot replace them. Every output still needs review. The human chain stays intact. The company has purchased a productivity tool.

When an AI system handles 85% or 90% of cases correctly, the economics change categorically. The workflow runs autonomously. Humans handle only the exceptions, the fraction of cases the system flags for review. The company has not purchased a productivity tool. It has purchased an autonomous function. The difference between these two states is not a percentage improvement. It is a reclassification.

In the first state, where AI is a productivity tool, the vendor sells software licenses. The buyer pays $50-$200 per seat per month from the IT budget. The vendor competes on features against other software vendors. The human labor cost remains on the books. This is SaaS economics.

In the second state, where AI becomes an autonomous function, the vendor sells an outcome. The buyer pays per claim processed, per credit analysis completed, or per supply chain route optimized. The budget is no longer IT spend. It is labor spend: the fully loaded cost of the knowledge workers the autonomous function replaces.

A single insurance underwriter costs $8,000-$15,000 per month in salary, benefits, training, and management overhead. A credit analyst costs similar. The vendor prices based on that labor cost, not on competing software licenses.

Apply this lens to a large institution like JPMorgan. With approximately $50 billion in annual labor costs and assuming ~30% of tasks fall into the “readily programmable” category, agents begin to meaningfully reshape the cost structure. However, because these tasks still require verification, exception handling, and integration into broader workflows, the immediate impact is not full replacement but partial automation and labor reallocation. Even under conservative assumptions, this translates into multi-billion-dollar efficiency gains, not through elimination of labor, but through compression of the programmable portions of work. That pool does not sit in the IT line. It sits in headcount.

The addressable market shifts accordingly. Enterprises typically spend 3-5% of revenue on IT, but 25-40% on knowledge work. The critical transition is not from software to AI, but from tools to workflow execution. Once an orchestrator reliably automates the programmable segments of work while coordinating human oversight where needed, it begins to tap into the much larger pool of labor spend. The prize is therefore not to replace SaaS but to expand beyond it by capturing portions of the operating budget that have historically been reserved for human execution. In that sense, the orchestrator that crosses the reliability threshold does not compete within the software market but instead redefines the market boundary, expanding it by a factor of 5 to 10.

The orchestrator creates this transition. The specialist agents, no matter how capable individually, do not. A pricing agent that achieves extraordinary accuracy still operates within the first state. It makes the human underwriter faster. It is a tool.

An orchestrator that coordinates pricing, risk, compliance, and documentation agents into a coherent autonomous workflow delivers the second state. It replaces the underwriting function. It delivers an outcome. The outcome emerges only from the coordinated whole. Only the orchestrator delivers it.

Bypass Risk

Orchestration Economics does not guarantee permanent advantage. The same structural logic that creates the orchestrator’s position also defines the conditions under which that position can be undermined.

Bypass risk is the possibility that a new entrant positions itself between the user’s intent and the existing orchestrator. In doing so, it captures the Orchestration Layer from above. Protocols such as MCP standardized how agents connect to external systems and to each other. A2A standardizes agent-to-agent communication. ANP sits atop both as a discovery layer.

Historically, APIs standardized software connectivity. In the agentic economy, protocols standardize agent connectivity. The strategic implication is profound: once interfaces become standardized, defensibility migrates away from integration and toward intent, context, and workflow ownership.

Standardization is powerful for interoperability. It also dissolves the integration friction that protected many orchestration positions. The company that controlled its workflow through the sheer difficulty of connecting systems wakes up one morning to find that the friction is gone. Protocols dissolved it. Orchestration Economics is therefore not a simple claim that the orchestrator always wins. It is a framework that says the orchestrator wins unless it gets bypassed.

The orchestrator’s advantage is structural, but its defensibility is conditional. It depends on building moats that protocols cannot standardize away and that competitors cannot replicate. Which moats? Under what conditions does an orchestration position become durable rather than temporary? What separates the orchestrators who compound their advantage from those who lose it to the next layer above?

The Three Laws

Knowing that orchestrators capture value does not tell us which positions are defensible. The orchestrator’s advantage is structural. Its defensibility is conditional. It depends on building moats that protocols cannot standardize away and that competitors cannot replicate.

Three conditions determine whether an orchestration position is durable. Each is necessary. None is sufficient alone. The entity that satisfies all three simultaneously holds a position that compounds with use, resists bypass, and captures the surplus of the agentic transition. The entity that satisfies one or two, even brilliantly, occupies a layer that the orchestration economy will eventually treat as crew rather than captain.

Chapter 8 develops the First Law: Proximity to Intent Determines Value Capture.

Chapter 9 develops the Second Law: Context Builds Moats.

Chapter 10 develops the Third Law: Workflow Intelligence Secures Control.

Once we define the Three Laws, we will apply them to Palantir.

The views and opinions expressed in this publication are those of the author alone and are based on publicly available information. The expressed views and opinions do not constitute investment advice, a solicitation, or a recommendation to buy or sell any security or financial instrument. The author may hold positions in the securities of companies mentioned. Certain companies referenced may be current or former clients of, or counterparties to, the author or affiliated entities; such relationships will be disclosed where applicable. Past performance is not indicative of future results. To the fullest extent permitted by applicable law, the author does not accept any liability for any loss or damage arising from reliance on this content. Readers should conduct their own independent due diligence and consult a qualified financial advisor before making any investment decision.

In three days, the market priced SpaceX at $1.75 trillion around a sub-frontier model, Anthropic shipped a real capability jump in Claude Fable, and the government forced Fable offline. The hot takes agreed that this meant the frontier is scarce, so open sources win. But this misreads commoditization. Intelligence commoditizes as a process, and the relationship between open source and frontier is symbiotic, rather than strictly oppositional. The frontier leads by a few months. That lead diffuses outward into the enterprise and into cheaper models as they catch up. The Orchestration Economy is predicated on the idea that diffusion never stops. Fable’s shutdown is the first time it stopped on command. When that happens, it has implications for the frontier, open weights, and orchestration. That becomes everyone’s problem.

Subscribe now

Three days did more to expose the structure of the agentic economy than the previous quarter.

Anthropic shipped Claude Fable on June 9th, and for two days, the conversation was about a real jump in capability rather than another point on the frontier model leaderboard.

On June 12th, the market priced SpaceX at $2.1 trillion after the first day of trading, despite its AI division, whose model is far from the frontier.

At 5:21 p.m. that same afternoon, the U.S. government delivered an export control directive to Anthropic ordering it to block access to any “foreign national,” a move that resulted in the company switching Fable off, not for Americans, not for a region, but for everyone on earth.

The new conventional wisdom that emerged identified the clear lesson: the frontier is becoming geopolitically scarce, so open-source wins. It offered a certain amount of reassurance because if the closed frontier can be imposed by a letter from the U.S. Commerce Department, the future must belong to the weights that no government can barricade.

The conclusion is appealing because it treats the frontier and open source as opposing forces. If governments can restrict one, the other must benefit. But that framing mistakes the relationship between them. The frontier and open source are not competing systems. They are successive stages of the same economic process. The frontier generates new capability. Open source absorbs it. Enterprises operationalize it. The reason intelligence appears to commoditize is not that the frontier stops mattering, but because each new increment of capability eventually diffuses outward through the rest of the ecosystem.

This distinction matters because commoditization is not a state. It is a flow. Intelligence does not commoditize by standing still while the field catches up. It commoditizes by diffusing. New capability appears at the frontier and then spreads through distillation, open weights, lower-cost models, enterprise products, and agentic systems until yesterday’s breakthrough becomes tomorrow’s baseline.

Once you see how the process runs, then last week stops looking like a verdict for open source and starts looking like something less comfortable: the first deliberate interruption of a mechanism the whole AI economy depends on. What the Commerce Department severed was not simply access to the frontier. It severed the channels through which frontier capability becomes abundant.

A central pillar of the Orchestration Economics framework is the collapse of the marginal cost of cognition. As intelligence becomes increasingly abundant, intelligence ceases to be the primary source of scarcity.

Value, therefore, migrates toward the systems that direct, coordinate, and embed intelligence in real operational environments. The prize ceases to be the model itself and instead becomes the orchestration layer surrounding it.

That framework implicitly assumes diffusion continues. Fable demonstrated that diffusion can be interrupted. Once it becomes interruptible, frontier access ceases to be merely a source of innovation and becomes a strategic dependency, a supply-chain risk, and a new source of scarcity. The abundance of intelligence turns out to depend on a process that governments can slow, restrict, or stop altogether.

This creates a more uncomfortable question than the one the public debate immediately settled on. The question is not whether the frontier matters. The market answered that one twice in the same week: once by assigning a $2.1 trillion valuation to a company whose flagship model is not at the frontier, and once by reacting with alarm when access to a frontier model disappeared overnight. The question is how those two facts can be true at the same time.

The answer lies in the mechanism that connects them: diffusion. To understand why last week matters, we need to begin with how commoditization actually works.

Share

How commoditization actually works

Let’s begin where the open-source camp and I agree. The frontier’s lead is real, but it is measured in months. According to Epoch AI’s capability index, the best open-weight models trail the leading closed models by three to four months, a gap that has held even as the frontier accelerated and that occasionally collapses to nothing.

At the model layer, there is no decade-long moat. Whatever the frontier learns today, a downloadable model can do in a season for a fraction of the cost.

That is the headline. It is also incomplete.

The headline describes the outcome. It does not describe the mechanism that produces it, and the mechanism is the whole argument. As I noted previously, intelligence does not commoditize by standing still while the field catches up; It commoditizes by diffusing.

New capability emerges at the frontier and then propagates through the ecosystem: first into lower-cost models, which steadily close the gap through distillation and imitation, and then into the enterprise, where each increment of capability is embedded into workflows, software, and agentic systems. Commoditization is simply the steady state of that process. It is what happens when the frontier continuously feeds the layers beneath it.

In dollars, a second pattern emerges, and it is the one the capability discussion often obscures.

The frontier and the commodity layer do not obey the same economics. Once a capability begins to diffuse, its cost collapses with extraordinary speed. The cost of commodity inference has fallen between fifty and two hundred times per year, accelerating since early 2024. Capabilities that once required frontier models have become cheap enough to disappear into software budgets and operational workflows.

The frontier follows a different curve. Its costs decline, but much more slowly, perhaps five to ten times per year. The newest capability remains expensive because it is newly created and briefly scarce. The economic gap between the frontier and the commodity layer is therefore not closing. It is widening.

Work naturally sorts itself along the same boundary. High-volume, repeatable tasks such as claims processing, document extraction, and routine compliance reviews ride the commodity curve downward. Tasks that depend on judgment, coordination, and reasoning across ambiguity remain tethered to frontier or near-frontier models because those capabilities have not yet diffused. Fable made the distinction more visible than any chart could. It launched at $10 per million input tokens and $50 per million output tokens. That’s twice the price of Opus 4.8, the model it replaced for general use. At the very edge of the frontier, commoditization did not merely slow. It reversed. The newest capability entered the market at a premium.

This is why it is misleading to speak about “the model layer” as though it were a single market. In aggregate, models commoditize. At the frontier, capability behaves more like a scarce asset: expensive, slow to cheapen, and occasionally more costly than the previous generation. The Orchestration Economy assumes that this scarce capability continuously feeds the layers beneath it.

Value migrates toward context, coordination, and proximity to intent because yesterday’s frontier eventually becomes everyone’s commodity.

The entire mechanism depends on that flow continuing.

What the market priced in SpaceX

Which is what makes the SpaceX listing worth a second look.

The prospectus sizes the AI opportunity at $26.5 trillion, of which $22.7 trillion is labeled as “enterprise applications.” That is not a model number. It is an orchestration number. It represents the value of sitting at the center of how companies operate: where work begins, where decisions are made, and where intelligence is translated into outcomes.

The significance of the IPO is not that the market suddenly became optimistic about enterprise AI. It is the market that reveals where it believes value ultimately accumulates. SpaceX was valued at $2.1 trillion, even though Grok is not widely regarded as a frontier model. The capability gap barely registered. Investors were underwriting a position in the orchestration economy, not a position on a leaderboard.

Whether SpaceX ultimately earns that position is a separate question. As I argued two weeks ago, xAI is not yet embedded in the places where enterprise work runs, and the company’s current footprint in business usage remains marginal. The segment accounts for less than 2% of business usage. The $22.7 trillion is pure prospect, a position priced as though already won by a company that has not entered it.

But that only reinforces the point. The market was willing to assign extraordinary value to the possibility of controlling the orchestration layer while largely discounting relative model performance.

In other words, the market already understands the conclusion that much of the AI debate continues to resist: the prize is not the model itself. It is the position from which intelligence is directed, coordinated, and operationalized.

Yet that conclusion contains an assumption of its own. It assumes the flow of new capability from the frontier continues uninterrupted.

The market priced the destination. Fable exposed a dependency on the journey and what happens when that assumption is tested.

What Fable was

The mistake with Fable is to argue about benchmarks. The significance of Fable was not that it answered questions better. It was that it crossed reliability thresholds that determine whether work can be delegated at all.

Taking a step back, that means Fable potentially changes whether entire categories of work can be automated. If so, then Fable becomes the first capability increment that appears to matter economically.

Consider this result to illustrate this point. Stripe pointed Fable, running on its own, at a fifty-million-line Ruby codebase to perform work counted in engineer-months. Fable finished in a day. Given persistent memory, it held its place across millions of tokens and tracked complex state roughly three times more reliably than the model that would shortly replace it. Neither is it a matter of knowing more facts. Both are a question of coordination: staying coherent over a long horizon, catching it’s own errors, deciding what comes next, and knowing when to stop.

As I wrote when Anthropic disclosed that Claude now builds Claude, the most valuable thing the frontier learned to do is judge its own work.

Viewed through the cost curve, Fable’s position is almost inevitable. Long-horizon coordination occupies the narrow band where capability remains scarce: work that depends on judgment, persistence, and self-correction rather than retrieval or recall. It is among the last capabilities to diffuse, and the first enterprises need before autonomy becomes practical.

Fable was the newest, least diffused increment of the capability on which the agentic economy is built. It had not flowed anywhere. Then it was switched off.

When the diffusion stops

What the Commerce Department interrupted was not the frontier’s existence but its diffusion. Access to Fable disappeared overnight, and with it the channels through which new capability normally spreads: into enterprises that operationalize it and into the broader ecosystem of models that eventually learn from it.

It was done for a reason, not a whim. The architecture that made Fable a better coordinator made its unrestricted twin, Mythos, a genuine strategic instrument. It found decades-old vulnerabilities in critical software on its own and generated biological hypotheses that its makers’ scientists preferred to their own. The trigger came from inside the tent. Amazon’s researchers, the Wall Street Journal reported, used a chain of prompts to get the model to give up cyberattack-useful material. Amazon CEO Andy Jassy reportedly carried the finding to Treasury Secretary Scott Bessent. Anthropic says it was then given ninety minutes to pull the model.

That a single demonstrated bypass moved that fast, and that it came from Anthropic’s largest partner, $13 billion invested and committed to $100 billion of cloud spend, suggests the government was no longer treating the model as a commercial product. It was treating unrestricted access to this level of capability as a strategic asset whose diffusion it was willing to restrict, regardless of commercial consequences.

So, the question is no longer whether the frontier matters. It is what happens to an economy built on diffusion when the diffusion becomes conditional.

There are two answers, neither of them easy.

To the enterprise, first. We tend to think of an agentic system as a finished product: a workflow once built, a harness once deployed, a moat once established. It is not. An agentic system is a position inside a continuously evolving capability stack. Its value depends on a steady supply of new capability arriving from upstream as the frontier advances and each increment filters down into production.

Freeze the supply, and the system freezes at the last increment it received.

That would be survivable if such work were to degrade gently. It does not. Long-horizon autonomy lives just above a reliability threshold, and the newest frontier coordination is what has carried it there. Drop the line, and an agent below does not give slightly worse answers; its long runs stop finishing. The companies that fell back to Opus 4.8 on Friday did not get a worse Claude. The workflows that had been clearing on their own simply stopped clearing.

To the economy, second. The gap that commoditization is meant to close stops closing.

Scarcity, which the whole framework assumes is migrating off the model toward context, snaps back to the model - not forever, but for a window no one can schedule.

Scarcity does not disappear. It migrates. The framework assumes it migrates steadily away from intelligence and toward context, coordination, and operational embedding. Last week demonstrated that the movement is not one-way. When diffusion stops, scarcity snaps back toward the model layer. Perhaps not permanently, but for a window whose duration no one controls.

The asymmetry is the one the Manifesto returns to repeatedly: destruction is priced in days, creation in years.

The diffusion lag was measured in months. The revocation was immediate. A dependency at the center of an agentic stack can be pulled on a Friday afternoon, while the context that was meant to be the moat took a decade to accumulate.

The two clocks run at different speeds, and the space between them is where the new risk lives.

The market read it that way within hours, less a story about Anthropic than about everyone, proof of how brittle a workflow becomes when it rests on a single vendor within one country’s ecosystem.

In retrospect, this is what much of the debate around Chinese open models has been missing. The significance of Qwen, DeepSeek, GLM, and their successors was never simply that they were becoming competitive with American systems. It was that they were creating a second channel through which capability could diffuse.

For the past year, I have argued that the rise of Chinese open models represented a structural shift in the economics of AI because it weakened the ability of any single company to control the flow of intelligence. What last week demonstrated is that the same dynamic may matter for a different reason: Redundancy.

A world with one diffusion channel behaves very differently from a world with two. In the first, a regulatory intervention can interrupt the entire flow. In the second, capability routes around the blockage. The question is no longer whether Chinese open models reach parity with the frontier. It is whether they provide a second source of diffusion when the first becomes politically constrained.

The drift was already on: a congressional commission found this spring that four in five US start-ups were building on Chinese open-weight models, whose share of global downloads had climbed from roughly 1% to 30% in a year.

A frontier that can be switched off is one that firms must find a way to insure against.

Why open source and orchestration do not suffice

Naturally, that instinct turns to open source. Route around the loss with open weights and a good harness. Both are real strengths. Neither replaces what was taken.

This represents a fundamental category error. Open source and orchestration are often treated as substitutes for the frontier. They are not. They exist downstream of the frontier. They are what diffusion produces, not what produces it.

Open weights are the frontier capability arriving on a lag, and a fine bargain while the flow runs. Stop the flow, and the open frontier freezes where it stands, while the true frontier, now behind a classifier and an export control, keeps moving.

The mechanism that would close the gap is distillation from a frontier teacher, which is precisely what the anti-distillation regime and the ban exist to prevent. Open source keeps the lights on. That is not nothing, and in a crisis, it is a great deal.

What it cannot do is manufacture the newest increment of capability, the one that lives only at a frontier no longer in reach.

Orchestration fails on the other face of the same coin. The confusion comes from mistaking leverage for creation. A harness multiplies the intelligence at its center. It cannot manufacture it.

The best routing, memory, and verification in the world take what the model can do and get more from it. But multiplying a capped number results in a capped value. No coordination conjures frontier-grade reasoning from a model that lacks it.

The production data make the dependence plain: in the most detailed survey of agents running in production, seventeen of twenty are built on a proprietary frontier model, with open weights in three, and only where cost or regulation forces them. Even the designs built to spend as little as possible on the frontier prove the point.

Anthropic’s own advisor strategy uses a cheap model to run the entire task and calls a frontier advisor only for decisions it cannot make. It is engineered to touch the frontier as rarely as possible, but it still cannot do without it.

Two other events from the week help clarify the roles.

The city of Rio de Janeiro post-trained a 397-billion-parameter model on an open Chinese foundation model and shipped it into production. But what Rio built was not a substitute for the frontier. It was a demonstration of diffusion at work.

The intelligence came from elsewhere; the value came from adaptation. Rio combined a widely available capability with a context that only Rio possesses. It is the Cursor or Harvey pattern expressed at the civic scale: not winning the frontier race, but capturing value by embedding intelligence into a domain no one else can replicate.

Mistral presents the mirror image. Reports suggest the company is raising at a valuation approaching €20 billion despite benchmarks that trail the frontier. That is only surprising if one assumes the market is paying exclusively for frontier capability. It is not. It is paying for sovereignty, distribution, enterprise relationships, deployment, and strategic positioning inside a Europe increasingly uncomfortable with technological dependence.

The apparent contradiction disappears once the roles are separated.

The frontier creates capability. Diffusion spreads it. Orchestration captures it.

Rio demonstrates the third stage. Mistral monetizes the second. Anthropic still occupies the first.

The week’s countermove proved it by accident. Within a day, Zhipu open-sourced GLM-5.2 under a permissive license, a million-token context, no regional limits, framed as the answer the moment demanded. Frontier intelligence, it argued, should not be a thing that a few rules can switch off.

As a promise of continuity, it is unanswerable: open weights cannot be revoked by a letter at 5:21 p.m. on a Friday. As a promise of function, it is unproven, since Zhipu published no validated benchmarks.

Continuity is not function. The entire debate over Fable spent a week treating them as though they were the same thing.

Control, not value

Fable and SpaceX don’t refute the frameworks for Orchestration Economics. Rather, the lessons enrich and clarify them in important ways: Intelligence becomes abundant through diffusion, and value migrates to orchestration for as long as diffusion holds. Sever it, and the migration reverses for a time, because the orchestration layer is now multiplying an input that has stopped improving.

The frontier is not where value accumulates. It is where control accumulates over the flow on which everyone downstream depends. The value settles one ring out in the operational context that no lab can reach. But control over the flow remains at the source, and the source still sits at the frontier.

That is an uncomfortable place to land, especially for Anthropic. For three days, it held the cleanest story in the industry: the orchestration leader, two-fifths of the enterprise, a coding wedge grown into the routing graph, a frontier model that finally made the source of the flow visible, and a prospectus already filed on some $47 billion of run-rate revenue.

Then the Commerce Department showed, in an afternoon, that the tap is the government’s to close.

Anthropic had spent the spring arguing that Mythos-class capabilities ought to be regulated as national-security assets. The government agreed, and the agreement cost the company its own product.

It goes to market carrying the contradiction it helped write. The bull case is that it holds the most valuable position in the agentic economy. The newest risk is that the most valuable position is the one a government is most likely to seize.

The lesson extends far beyond a single company. Every orchestrator built on a single frontier supply learned the same lesson in the same hour: proximity to intent and a decade of context are necessary, and not sufficient, when the intelligence at the center can be switched off on ninety days’ notice, or none.

You could watch the hedging start at once. Within days, developers were mining the model’s leaked system prompt, some 120,000 characters of it, published after the jailbreak, as a manual for long-running agents, grafting its scaffolding onto whatever models they could still reach. The exercise is instructive where it fails: a system prompt can encode process, structure, and instructions, all of which are portable. The judgment that executes those instructions lives elsewhere. It lives in the model itself, and that is the part that does not travel with a copied prompt.

The resilient architecture of the Agentic Era, then, is not the one built around a single frontier supplier. It is the one designed to withstand any supplier interruption. The lesson of last week was not that the frontier no longer matters. It was that it matters enough to become a dependency.

For three years, the industry optimized for access to the best model. The next phase will be defined by something more mundane and more important: redundancy. A frontier that can be withdrawn becomes a supply chain. And supply chains are not trusted. They are hedged.

The views and opinions expressed here are those of the author alone and are based on publicly available information. They do not constitute investment advice, a solicitation, or a recommendation to buy or sell any security or financial instrument. The author may hold positions in the securities of companies mentioned and maintains no current position in Anthropic, SpaceX, Mistral, or any private entity discussed. Past performance is not indicative of future results. Readers should conduct their own independent due diligence and consult a qualified financial advisor before making any investment decision.

This is the latest excerpt from AGNT: The Orchestration Economics Manifesto - An Investment Framework for the Agentic Era. Each Thursday, I explore a major theme of the Manifesto and unpack the frameworks, adding extra context with more recent developments. Note: The figures and sequential references are taken directly from the larger Manifesto.

Subscribe now

The agentic economy is a software phenomenon built on a hardware foundation. That foundation is being constructed on a scale without precedent in the history of private capital allocation. But is it being constructed correctly?

When I first published this Manifesto in late April, the projected $650 billion in AI infrastructure investment in 2026 already represented one of the largest coordinated private investments in peacetime history. Since then, those numbers have continued to soar.

In May, the four largest hyperscalers reported quarterly earnings within hours of each other and raised 2026 capital spending guidance, committing between $650 and $700 billion to AI infrastructure for the year. Goldman Sachs raised its Capex spending forecast for the Big Four Hyperscalers from $4.5 trillion to $5.3 trillion through 2030. Overall, it projects $7.6 trillion of Capex spending between 2026 and 2031 across compute, data centers, and power.

The problem remains, however, that much of this capital is also being committed under assumptions that are already shifting.

The capital was designed for training. The agentic economy runs on inference. The cloud was designed for stateless applications. Agentic workloads are stateful, persistent, and coordination^-intensive. The power grid was designed for variable demand. Inference generates a continuous baseload.

The infrastructure does not need to stop being built. It needs to change what it is becoming. This chapter traces that metamorphosis, from the capital flowing in, through the balance sheets absorbing it, to the financing structures underwriting it, to the inflection point where the workload changes, to the efficiency gains that determine the unit economics, to the energy constraint that governs the speed, and finally to the architectural evolution of the cloud itself. Together, they constitute the infrastructure question on which the AGNT thesis ultimately depends.

Share

The Buildout

The numbers announced during Big Tech earnings calls throughout 2025 arguably constituted one of the largest coordinated private infrastructure investments in history. Bloomberg estimated that the four largest hyperscalers would spend ~$650 billion in 2026. Goldman Sachs projected cumulative hyperscaler Capex of $1.4 trillion from 2025 to 2027. That is close to three times the $477 billion spent from 2022 to 2024. Leading hyperscalers have emphasized that they are primarily constrained by power and capacity, not by demand, repeatedly flagging supply bottlenecks on recent earnings calls.

Microsoft carried an $80 billion backlog of unfulfillable Azure orders due to power constraints as of January 28, 2026 (Microsoft Q2 2026 Earnings). Alphabet’s cloud backlog surged to $240 billion at year-end 2025. Amazon said new capacity “sells out immediately”.

Yet investors punished these announcements. Amazon fell 8%. Microsoft dropped ~10%. The four companies collectively shed over one trillion in market value after revealing their spending plans. By February 2026, a Bank of America survey showed overinvestment concerns at an all-time high, roughly 35% warning of excess, the highest reading in twenty years.

Is this Capex buying what the agentic economy will require?

The Balance Sheet Transformation

The balance sheets of the world’s most profitable technology companies underwent a structural transformation beginning in 2025. For two decades, the defining characteristic of large-cap technology was capital efficiency. It was driven by asset-light models generating extraordinary returns on modest physical investment.

That era ended when the AI buildout began. The more than fourfold increase in capital intensity from 2015 to 2026, the dollars of capital required to produce one dollar of sales, tells the story.

Cloud computing infrastructure from 2015 supported predictable workloads with well-understood utilization patterns and multi-year lifecycles. AI infrastructure serves workloads with radically different characteristics: shorter useful lives, uncertain utilization, and a pace of obsolescence that makes traditional depreciation assumptions unreliable.

The depreciation debate alone reveals the depth of uncertainty. Amazon shortened the useful lives of a subset of its servers and networking equipment from six to five years in January 2025. Microsoft extended expected asset life in FY 2023, adding $3.7 billion in operating income from the change alone. These are not accounting technicalities. They are competing theories about how long AI hardware will retain value. The answers differ by billions of dollars.

Yet, the demand for NVIDIA H100’s has never been higher, suggesting longer utility than GAAP rules. When the industry cannot agree on whether its core infrastructure lasts one year or six, every capital allocation decision rests on a foundation of sand.

The free cash flow consequences are direct. Bank of America projected that Amazon’s free cash flow would turn negative at approximately $28 billion in 2026. Amazon filed with the SEC^, indicating it may need to tap equity and debt markets. Such a statement would have been unthinkable for a company generating $638 billion in annual revenue in 2024 (and $717 billion in 2025). Pivotal Research projected that Alphabet’s FCF would plummet approximately 90% in 2026, from $73.3 billion to approximately $8.2 billion, as $175-185 billion in Capex overwhelms operating cash flow. Barclays projected Meta’s FCF to drop approximately 90% in 2026, then turn negative in 2027 and 2028.

Step back and absorb the aggregate picture. Bank of America estimates that AI Capex now accounts for 94% of hyperscaler operating cash flows after dividends and buybacks. That figure was 76% in 2024. In the span of a single year, Capex has gone from absorbing three-quarters to nearly all of the available post-distribution cash flow, substantially compressing the cushion and forcing record debt issuance. The margin for error has dramatically narrowed.

The weight of physical assets on these balance sheets means that even the most profitable technology companies in history are subject to the same economic physics that govern utilities, railroads, and industrial conglomerates. Returns on invested capital matter more than revenue growth. The companies that built their fortunes on capital-light software economics are becoming capital-intensive infrastructure operators.

The financial frameworks applied to them have not caught up.

The Debt Machine

The Big Five (Amazon, Alphabet, Meta, Microsoft, and Oracle) raised ~$121 billion in bonds in 2025 alone: Amazon $15 billion, Alphabet $25 billion, Meta $30 billion (the largest-ever individual non-M&A high-grade bond sale), and Oracle $18 billion. That’s more than three times the average of the prior nine years. Alphabet’s long-term debt quadrupled in 2025 to $46.5 billion.

Oracle, whose fiscal year 2026 ended in May, disclosed that it had raised $43 billion in debt financing and $5 billion in equity financing. The company said it plans to raise $40 billion in debt and equity financing in FY 2027.

Companies that spent decades accumulating fortress balance sheets are now leveraging them at a blistering pace.

Morgan Stanley had earlier estimated that hyperscalers will borrow ~$400 billion in 2026, double the $165 billion borrowed in 2025. On June 10, Morgan Stanley predicted that AI-related global debt issuance would double to about $570 ​billion in 2026. JPMorgan projects $1.5 trillion in AI-related bond issuance over the next five years.

The clearest expression of that thesis arrived in February 2026, when Alphabet issued a 100-year bond. This was the first century bond from a tech company since Motorola in 1997. The sterling-denominated tranche, part of a $31.5 billion multi-currency debt raise completed in under 24 hours, was oversubscribed by nearly 10 times. Pension funds and life insurers, institutions that match liabilities measured over generations, bought the paper at a 6.05% yield, treating Google's parent as sovereign-grade infrastructure. Alphabet sits on c.$127 billion in cash (cash, plus equivalents, plus marketable securities) and chose to borrow anyway because even that war chest looks modest against $175 – $185 billion in planned 2026 Capex.

The century bond is a statement that buyers believe this buildout is not a cyclical phase but a permanent transformation of the economy’s substrate. The last tech companies to issue century bonds were IBM in 1996 and Motorola in 1997. They did so at the zenith of their market dominance. Both subsequently declined.

At the DealBook Summit in December 2025, Anthropic CEO Dario Amodei described the high-wire financial planning. Established scaling laws, he said, are here to stay. Models get better at every task as you put more data and compute into them. But the timing of economic value remains uncertain, creating an inherent risk of under-reacting or over-extension.

Spend too little, and a company like Anthropic or OpenAI risks being unable to serve customers in two or three years. Overspend, and companies face revenue shortfalls that in extreme cases could threaten bankruptcy or tempt them to adopt what Amodei called a “YOLO” mindset.

The Fragility Beneath the Conviction

The hyperscalers can service this debt. They have investment-grade ratings, massive revenue bases, and the operating cash flow, though diminished, to cover even aggressive borrowing. The fortress is leveraged, but it remains a fortress.

The risk lies elsewhere, in OpenAI deploys the chips at CoreWeave, which has grown so quickly and so intricately that its systemic properties deserve the same scrutiny applied to any previous era of financial innovation. At the center sits a pattern of circular vendor financing.

The architecture is straightforward to describe and difficult to contain. NVIDIA invests in OpenAI. OpenAI uses the capital to purchase NVIDIA chips. NVIDIA records the transaction as revenue (subject to LOI stage; not yet binding). OpenAI deploys the chips at CoreWeave. CoreWeave secured term-loan facilities collateralized by GPUs. Institutional investors purchase the securities as infrastructure debt, providing the capital that flows back to the beginning of the cycle. The same dollar is counted as demand by the chip maker, revenue by the neocloud, and collateral by the lender.

The pattern is not novel. Classic vendor financing catalyzed previous bubbles. The 2000s telecoms collapse followed this structure, with equipment makers extending credit to customers, inflating revenues while disguising demand weakness. When customers could not pay, receivables evaporated, and cascading failures followed.

What is novel is the concentration. Telecoms vendor financing funded “the internet” broadly across hundreds of carriers and thousands of build-outs. Today’s circular deals are largely concentrated around a single company. OpenAI sits at the nexus of NVIDIA’s investment, Microsoft’s partnership, Oracle’s data center contracts, AMD’s equity-for-chips arrangement, and CoreWeave’s largest customer relationship.

The greatest risk is not that AI fails. It is that the technology succeeds while the single company around which the financial architecture has been constructed fails to build a defensible moat, triggering contagion that strands capital and stalls the buildout before the transformation completes.

Because while this concentration of relationships presents a sizable risk, it is also far from the limit. The boundaries extend far beyond this inner circle and continue to grow. Data center-related securitization via CMBS and ABS products is expected to reach $30-40 billion annually in 2026 and 2027.

Private credit finances approximately $50 billion in AI infrastructure quarterly. The instruments are growing more complex. For GPU-backed instruments, collateral depreciation is outpacing debt amortization, even if the broader data center ABS/CMBS market is not similarly exposed (yet). And the whole structure is more concentrated around fewer counterparties than any comparable financing architecture in recent financial history.

You can be bullish on AI and bearish on this financial architecture. They are separable propositions. The technology may well deliver everything this manifesto argues it will. The question is whether the financing structure survives long enough to support the transition, or becomes the mechanism by which a sound technological thesis precipitates a financial crisis before the economics have time to prove themselves.

The financing architecture will be tested by what comes next. Because the infrastructure it underwrites was designed for a workload that is already giving way to something fundamentally different.

Two Tales of Compute

The vast majority of the capital committed in 2025 and 2026 was allocated on the assumption that the primary demand for compute would come from training frontier models.

Training was the activity that the industry understood when the investment decisions were made. This means massive, centralized GPU clusters running synchronous workloads for weeks, requiring homogeneous hardware, high-bandwidth interconnects, and cutting-edge silicon. Each generation of models requires roughly ten times the compute of its predecessor. Frontier AI training costs escalated from over $100 million for GPT-4 (2023-2024) to approximately $490 million for xAI’s Grok 4 (mid-2025), with projections exceeding $1 billion by 2027.

First, let us examine the tale of training compute. The hardware characteristics are distinctive. Training requires homogeneous clusters of cutting-edge GPUs with high-bandwidth interconnects, where a single GPU failure can corrupt an entire run that takes weeks of computation and costs millions of dollars. You cannot mix GPU generations or repurpose last-generation training infrastructure for new workloads without massive retrofitting.

Leading-edge racks consume 130-250 kilowatts (as of 2025), with projections reaching 900 kilowatts by 2027. Frontier training performance advantages degrade within 18-24 months, but the hardware itself retains economic utility for more than five years, typically migrating to inference or lower-tier workloads. This creates an oligopolistic market. The number of organizations training frontier models globally is perhaps fifteen to twenty:

Anthropic’s arrangement with AWS provides co-engineering of custom Trainium chips that reportedly reduce training costs by 40% compared to GPUs.

Google’s TPU advantage allows training at dramatically lower costs than GPU-based alternatives.

xAI built its Colossus cluster from site preparation to full operation in 122 days.

Access to massive, synchronized GPU clusters that require +250-megawatt data centers and billions in capital now determines who can compete at the frontier. Training compute is too strategic to outsource. Microsoft, Google, and Amazon have all reached the same conclusion.

Training is necessary. Without it, the models that power the Agentic Era do not exist. But training is not where this manifesto's thesis lives. Training is one-time Capex, concentrated among a handful of players, and depreciates in 18 months. It is the cost of building the engine.

Now, let us look at the tale of inference compute. Every query processed, every agent action executed, every token generated represents ongoing expenditure that compounds relentlessly. A single ChatGPT query costs approximately $0.003 in compute. That is seemingly negligible until multiplied by 800 million weekly users^, each generating multiple queries.

At five queries per user weekly, that is four billion queries, $12 million in weekly compute, $624 million annually for inference alone. Over a model’s operational lifetime, inference costs routinely exceed training costs by factors of three to fifteen. At CES in January 2026, Lenovo’s chief executive officer predicted that 80% of AI compute will be inference and only 20% training. As of March 2026, inference already accounts for over 60% of AI workloads and is projected to reach 70-85% by 2028-2030.

The economic characteristics diverge at every level. Training hardware depreciates in eighteen months. Inference hardware maintains utility for four to five years. Training serves a concentrated customer base for frontier labs. Inference serves a broadening market of millions of enterprises and consumers. Training requires the latest GPU architecture. Inference benefits from specialized silicon optimized for throughput and cost per token rather than peak performance.

Training is the arms race. Inference is the economy.

Inference Efficiency

At GTC 2026, NVIDIA CEO Jensen Huang completed a pivot that had been in the making for 18 months. His keynote focused on inference, agents, and AI factories. Data centers are factories, Huang argued. Their product is the token. Their efficiency determines the unit economics of everything built above them. The governing metric has shifted from training FLOPS to tokens produced per watt.

This pivot culminated in the introduction of NVIDIA’s Vera Rubin platform, which promises to deliver 10x the inference performance per watt compared to its Blackwell architecture. The pivot was reinforced by NVIDIA’s December 2025 deal with chip startup Groq. The deal was reported to be a $20 billion non-exclusive licensing agreement for Groq’s LPU inference technology, paired with the hiring of CEO Jonathan Ross and the bulk of Groq’s engineering team.

Groq remains operationally independent. At GTC 2026, Huang showed how the two architectures would work together: Vera Rubin handling the prefill phase of inference, and the new Groq 3 LPX rack, housing 256 LPUs, handling the decode phase where tokens are generated. The Groq integration recalled NVIDIA’s “Mellanox moment” several years ago^, when it acquired the company for its networking architecture (InfiniBand, Spectrum-X Ethernet), which became cornerstones of AI cluster design. Groq’s architecture is optimized for inference, adding deterministic low-latency decoding using massive on-chip SRAM at 80 TB/s internal bandwidth, achieving 500+ tokens per second on Llama 3 70B at up to 10x the energy efficiency of GPUs.

On the silicon side, the inference landscape has fragmented in ways training never did: AWS Inferentia claims significantly better price-performance, Google’s TPUv5e offers strong cost-effective efficiency gains, AMD and a generation of startups are all targeting the inference economics that determine AI’s unit costs.

The algorithmic gains compound on top of the silicon. Mixture-of-experts architectures activate only a fraction of total parameters per query. Qwen3-Next, for example, activates 3 billion of 80 billion through ultra-sparse expert routing, achieving 10x faster inference on long contexts. Dynamic batching pushes GPU utilization from 30% to over 80%. Speculative decoding doubles token-generation speed at a fixed cost. These represent order-of-magnitude gains that compound with one another and with the silicon improvements beneath them, producing a 5-10x annual price-performance trajectory.

Stanford and Together AI researchers quantified where this trajectory leads. Their Intelligence Per Watt metric improved by a factor of 5.3 between 2023 and 2025. Local models on consumer-grade hardware answered 88.7% of single-turn reasoning queries correctly. The share of queries that edge devices can handle rose from 23.2% to 71.3%. Inference is beginning to migrate from data centers to the edge, and this trend will accelerate as devices become more capable.

The Jevons paradox operates here with particular force. This refers to the principle that increased resource efficiency drives higher total consumption. In this case, efficiency gains in inference will expand the frontier of viable workflows, which expands total consumption even as per-unit costs fall. A workflow unprofitable at $50 in agent compute becomes routine at $5. A category of work inconceivable at one price point becomes obvious at a lower one. The efficiency gains feed the demand they were designed to serve.

This has direct implications for the AGNT thesis. The orchestrator’s primary variable cost is the cost of inference tokens. Every improvement in efficiency expands the addressable market. The orchestrator that routes tasks intelligently across capability tiers (commodity models for extraction, frontier models for complex reasoning, edge devices for latency-sensitive work) achieves structurally better margins.

Whether the agentic economy becomes the deflationary force described in Part V depends on whether this efficiency trajectory continues to compound. If it does, the surplus available to orchestrators grows with every cycle. If it stalls, the transition slows. Not for want of intelligence, but because the economics simply do not work.

The Energy Constraint

Beneath the capital, the debt, the silicon, and the algorithms lies a constraint more fundamental and less tractable than any of them: Power.

The scale of the demand is unprecedented. Huang’s $1 trillion demand outlook through 2027 implies not merely more chips and racks but more electricity consumed continuously at an industrial scale. A single Vera Rubin NVL72 rack draws power at levels that would have been classified as a small industrial facility a decade ago. Leading-edge AI racks consume 130-250 kilowatts today, with projections reaching 900 kilowatts by 2027. Each hyperscaler is pursuing a one-gigawatt data center campus, a scale that draws as much electricity as a mid-sized city.

The workload itself has changed shape. A chat query resolves in 500 milliseconds. An agentic task, such as a coding audit, a research synthesis, or a multi-system workflow, takes 10 minutes to 14+ hours and runs in parallel with thousands of siblings inside a single campus. The load profile of an AI data center has converged toward that of an aluminum smelter: continuous, high-wattage, near-100% utilization, 24/7. The bursty inference workload grid planners that were modeled against have been removed. The new profile is baseload.

In addition, agentic compute has a unique memory-power overhead. To maintain state over a 14-hour + task, agents must keep massive context windows active. In 2026, hyperscalers are offloading KV caches to SSD storage to manage the memory bottleneck, but the energy required for constant data shuffling remains a high hidden cost. Anthropic has published research on “context engineering”. This refers to compaction, structured note-taking, and multi-agent architectures designed to address “context rot” in long-horizon agents. These techniques reduce the memory-power overhead of maintaining state across extended tasks and indirectly ease the underlying power constraint.

The power infrastructure to support these facilities does not exist in most locations where they are being planned. Grid interconnection timelines run three to five years in the United States, longer in Europe. Permitting for new transmission lines routinely takes a decade.

The US is on track to add a record 86 GW of utility-scale generation in 2026: 51% solar, 28% batteries, 14% wind. Useful firm power near data-center hubs is a fraction of the headline figure. In PJM, the single most important AI interconnection, utility large-load commitments now exceed accredited generation capacity by a factor of 3. China, in contrast, added 434 GW in 2025 alone and is co-locating compute with generation in the renewables-rich West. This is not a cyclical shortfall; the US closes with Capex. Rather, it is a profound infrastructure divergence, one that could constitute a Single Point of Failure of the Agentic Era, i.e., the specific vulnerability that, if left unresolved, constrains the entire transition regardless of how fast models improve or how much capital flows into AI. In the framework we introduced in 2025, an SPOF is the component of a system whose failure compromises the entire structure.

Nuclear is, as of 2026, the only carbon-free source with the density for AI-scale loads, and the scramble for it since 2024 is no longer directional. It is the infrastructure. Microsoft underwrote the restart of Three Mile Island via a 20-year PPA with Constellation (835 MW, 2027-2028). Amazon bought Talen’s Susquehanna campus and is developing SMRs with X-Energy and Dominion. Google signed with Kairos (500 MW SMR fleet), Fervo Geothermal, Oklo, and Vistra, making it the largest corporate purchaser of nuclear energy in US history. Sam Altman backed Oklo (2.6% stake, former chair, co-founder of the SPAC that took Oklo public). xAI stood up 200+ MW of on-site gas turbines in Memphis in under a year. Stargate Abilene is architected as a 1+ GW behind-the-meter campus. Together, these constitute a coordinated admission that the public grid cannot deliver what agentic compute needs within the required timeline. Hyperscalers are no longer interconnecting to the grid: they are becoming utilities.

The constraint sharpens as the workload shifts from training to inference. Training runs are episodic. A frontier model trains for weeks in a concentrated burst, then stops. Inference is continuous. Every agent action, every orchestrated workflow, every token consumes power indefinitely, at a load that scales with agent populations.

As the mix shifts, the demand profile moves from periodic spikes to sustained baseload: precisely the pattern that stresses grids most acutely, because it requires always-on generation rather than peak-shaving.

The efficiency gains partially offset this. As noted before, Jevons in this respect operates at an industrial scale: cheaper inference makes more workflows viable, expanding total demand even as per-unit consumption falls. Hyperscalers that secure power early through long-term agreements, on-site generation, or strategic locations near abundant clean energy gain compounding advantages. The value of a data center is increasingly determined not by the GPUs it houses, but by the megawatts it can draw.

For the AGNT thesis, energy introduces a temporal variable into every projection. The Inference Economy expands at a rate bounded by the availability of power, regardless of demand or capital. The orchestrators of Parts III to V will find their ability to scale agentic workflows ultimately limited by whether the infrastructure they depend on has the electricity to operate.

The direction is not in question. The speed is. Watts, not weights, now govern the pace of the transition. The pace, as the Two Scenarios in Chapter 18 will argue, is what separates deflationary growth from demand destruction.

The Agentic Cloud

There is a final condition that completes the picture. It is architectural rather than financial or physical. The cloud was built for a different era. The SaaS-era cloud was designed for stateless applications, ephemeral compute, hub-and-spoke networking, and consumption-based pricing. An enterprise spins up a virtual machine, runs a workload, and shuts it down. Two decades of perfecting that pattern. It is the wrong pattern for what comes next.

The agentic systems of Chapter 5 generate workloads that differ in every dimension. Every MCP connection, every A2A coordination, every persistent memory retrieval, every multi-step workflow is an inference operation. An agent booking a flight generates 50,000 tokens across research, comparison, booking, and confirmation. This compares to just 500 for a chat query. Agentic workloads multiply token consumption 10x to 100x.

But the difference is not merely volume. Agents are stateful. They maintain persistent memory and workflow state across sessions, requiring considerably more memory than traditional workloads. The three-tier memory architecture of Chapter 5 requires infrastructure that preserves state, not infrastructure designed to forget:

Agents coordinate laterally. They communicate peer-to-peer through MCP and A2A, requiring sub-100ms latency and dedicated bandwidth between coordinating agents. The topology must evolve from a centralized switchboard toward a mesh.

Agents generate explosive storage demands. Vector databases typically experience 10x data expansion, with total storage reaching 30x the original data after indexing and context persistence.

Agents require a different economic model. Pricing must be tied to task completion and outcomes rather than compute hours consumed, which demands infrastructure instrumented for metrics that traditional cloud monitoring was never designed to capture.

These requirements point toward a new infrastructure category: the agentic cloud. Three categories of providers are positioned to build it.

The hyperscalers hold the deepest structural advantages: custom silicon, enterprise relationships built over decades, and captive internal demand for inference. Google’s Search, YouTube, Gmail, and Maps are continuously consuming inference, providing a substantial utilization floor regardless of external demand. Their challenge is that hundreds of billions in existing infrastructure were designed for the SaaS era. The organizational complexity of serving legacy and agentic workloads simultaneously is real, and scale alone does not resolve it.

The neoclouds proved something important: that AI workloads demand fundamentally different compute from general-purpose clouds. CoreWeave and other AI-specialist clouds have demonstrated MFU improvements of 15-30% over hyperscaler baselines (e.g., CoreWeave’s 49.2% MFU on H100 training runs vs MosaicML’s 41.85% baseline, per CoreWeave’s published benchmarks). Industry training MFU typically sits in the 35-55% range, depending on hardware generation, software stack, and workload. But the neocloud window is narrowing. Their core design is stateless, single-tenant, and hub-and-spoke. This served training-centric workloads well. It is less suited to what agentic systems demand.

The emergent agentic clouds are purpose-built. Daytona Cloud achieves sub-90ms sandbox creation for stateful agent workflows. Others target observability for multi-agent systems, optimization for coordination-intensive workloads, and outcome-based resource allocation. They carry no architectural debt. They lack scale. They are building for the incoming workload.

The model providers see the same opening. Anthropic’s Claude Managed Agents, launched in April 2026, offers the full production stack for agentic workloads as a managed service: sandboxed execution, state persistence, tool orchestration, error recovery, and multi-agent coordination. It runs on existing cloud infrastructure, but the developer never touches it. Anthropic abstracts the cloud into a substrate and bills by the session-hour.

The orchestrator’s variable cost is the cost of inference tokens. Hardware efficiency sets the floor for the cost of producing a token. The agentic cloud determines the cost of serving that token within a coordinated, stateful, production-grade workflow. The companies that optimize for both will set the economic foundation for everything built on top of them.

The Substrate Must Evolve

The AI infrastructure being built in 2026 is the largest private capital deployment in history. It is transforming Big Tech’s balance sheets from capital-light to capital-intensive. It is financed through structures of increasing ambition, from conventional bonds to century paper to securitized GPU assets. And it was committed under assumptions about training workloads that the migration toward inference is already reshaping.

The metamorphosis is substantial but navigable. Inference efficiency is compounding at a pace that makes the unit economics of the agentic economy more favorable with each quarter. The energy constraint governs speed, not direction. And the cloud is beginning, unevenly, incompletely, but perceptibly, to evolve toward the stateful, coordination-grade architecture that agentic workloads demand.

The intelligence is sufficient. The agents are in production. The infrastructure is scaling and starting to change shape. What determines whether specific companies, sectors, and workflows successfully cross is the subject of Part III.

The views and opinions expressed in this publication are those of the author alone and are based on publicly available information. The expressed views and opinions do not constitute investment advice, a solicitation, or a recommendation to buy or sell any security or financial instrument. The author may hold positions in the securities of companies mentioned. Certain companies referenced may be current or former clients of, or counterparties to, the author or affiliated entities; such relationships will be disclosed where applicable. Past performance is not indicative of future results. To the fullest extent permitted by applicable law, the author does not accept any liability for any loss or damage arising from reliance on this content. Readers should conduct their own independent due diligence and consult a qualified financial advisor before making any investment decision.

AGNT Podcast Ep. 9 with Gemma Allen & Raphaëlle d'Ornano

00:00 - Intro

00:02 - Title: Exploring AGNT and Anthropic's Evolution

05:41 - Anthropic's Strategic Edge in Enterprise

10:29 - Addressing IT Budget Constraints

13:17 - Role of Tokenomics in AI

16:49 - Layered Structure of AI Systems

19:25 - Implications of Anthropic's Growth

23:03 - Reimagining Enterprise Software: Valuation, Reinvention, and the AI Paradigm Shift

SpaceX IPO S-1 Teardown

Durable Growth Moat™ Analysis

Why Enterprise AI Is an $800B Black Hole in the $1.75 Trillion Story

Decoding Discontinuity | June 2026

As the largest IPO in history approaches public markets, investors are being asked to underwrite three fundamentally different businesses: a dominant launch monopoly, a global connectivity compounder, and an AI platform pursuing one of the most ambitious enterprise narratives ever presented to public investors.

Our latest institutional teardown examines SpaceX through the lens of Orchestration Economics, Durable Growth Moats™, and the emerging structure of value capture in the Agentic Era.

The IPO narrative assumes that AI will be the largest driver of future value creation, with enterprise applications accounting for nearly 86% of the company’s projected AI opportunity. Yet beneath the scale of that ambition lies a critical question:

What structural position allows xAI to capture the enterprise market it is being valued to own?

This report examines why that distinction matters.

The teardown explores:

• the growing divergence between compute ownership and enterprise value capture,

• why infrastructure leadership does not automatically translate into orchestration power,

• the structural weaknesses in the enterprise AI thesis underpinning the IPO,

• Grok’s positioning relative to emerging orchestration leaders,

• the role of Cursor, developer workflows, and the Coding Wedge in future enterprise adoption,

• and the implications of orbital compute, consumer intent capture, and model commoditization for durable moat formation.

The analysis positions SpaceX not simply as a space company, but as one of the first public-market tests of a larger question confronting the AI economy:

Will value accrue to those who own the infrastructure, or to those who control the workflows where intelligence is ultimately applied?

The report includes:

• Durable Growth Moat™ assessment,

• Architectural Resilience Assessment Framework (ARAF™) analysis,

• sum-of-the-parts valuation framework,

• enterprise adoption benchmarking,

• and broader implications for value migration in the Agentic Era.

Bottom line: Starlink remains one of the most compelling compounders in the world. The AI story may eventually justify extraordinary value. But today’s IPO asks investors to prepay for an enterprise position that has yet to be earned.

Institutional Research

Decoding Discontinuity provides proprietary research for institutional investors, operators, and strategic decision-makers navigating AI-driven discontinuity.

Our institutional work focuses on:

• AI infrastructure and compute markets,

• frontier model economics,

• inference efficiency and orchestration systems,

• GPU, custom silicon, and cloud dynamics,

• AI-native business models,

• and value migration across the agentic stack.

Unlike our public essays and Substack analysis, Institutional Research delivers company-level teardowns, strategic frameworks, technical assessments, and market intelligence developed for professional use.

Institutional Access

Institutional Research includes:

• proprietary research reports,

• strategic briefings,

• direct analyst access,

• and custom research discussions.

Access is limited and managed directly.

To learn more about Institutional Research or request access to the Cerebras teardown:

Raphaëlle d’Ornano
Founder & CEO, Decoding Discontinuity
raphaelle@decodingdiscontinuity.com

or

Mathieu Huet
COO, Decoding Discontinuity
mathieu@decodingdiscontinuity.com

Anthropic’s Recursive Self-Improvement (RSI) research (i.e. AI autonomously designing and improving its own successors) isn’t about “superintelligence” or a “pause.” It is the seventh “tremor” in the shift from tools to goal-seeking agents. Intelligence and the Harness (routing, memory, orchestration code) are commoditizing fast and becoming self-improving, but real scarcity, which defines durable value and moats, migrates to the two true bottlenecks: scarce compute (which RSI actually intensifies) and operational context (the real-world data, outcomes, and judgment that labs can’t automatically generate or own). RSI accelerates everything except the parts enterprises control, giving Anthropic’s enterprise/Harness lead a powerful booster. But it’s not an unassailable moat yet.

Subscribe now

Even with the historic SpaceX IPO scheduled for later this week and rival OpenAI confidentially filing for its own public offering, Anthropic once again managed to hijack the public discussion about artificial intelligence. As is often the case in these early days of the Agentic Era, I think that the public conversation has tended to focus on the more immediate and dramatic aspects, while the deeper economic significance has received less attention.

In this case, the conversation stopper came not in the form of a new product, but rather a research paper published by Anthropic’s research arm: “When AI builds itself.“ The paper represented its most detailed public account yet of how it is developing a technique known as “recursive self-improvement” (RSI), which it describes as “an AI system capable of fully autonomously designing and developing its own successor.” The publication landed weeks after Anthropic scored a recruiting coup by hiring legendary AI researcher Andrej Karpathy, a founding member of OpenAI and formerly of Tesla, who recently had been working on his own RSI project, to build a team to accelerate the pre-training of the next Claude.

Because the Anthropic paper discussed the “implications” of such a system, the world focused almost solely on a single conversation:

It seemed astonishing that the company would call for such a thing, considering that on June 1, Anthropic confirmed it had confidentially submitted a draft S-1 that reportedly included an $47 billion ARR and had announced a $965 billion funding round.

So, was this a savvy PR exercise at a time of intense competition for attention, to steal the spotlight from the $1.75 trillion SpaceX IPO? Was this a branding exercise to reinforce Anthropic’s image as the good governance LLM?

Of course, the first problem is that Anthropic wasn’t really calling for a pause. But I will leave it to Gary Marcus to explain why.

However, from my perspective, the larger implication of this announcement flows back to the framework of Orchestration Economics that I have been building over the past three years. In the Orchestration Economics Manifesto, I define generative and agentic AI as Discontinuity, a geology that has been built as a sequence of six shocks along a fault line, each amplifying the others until the boundary gave way and created a structural break as machines crossed from being tools that process instructions to actors that pursue goals.

RSI represents the seventh tremor in that continuum. Again, the instinct is to view RSI as a capability story, perhaps a step toward superintelligence or AGI. This also misses the point. As with the other six tremors, the real question is what happens to value and moats when the production of intelligence itself becomes recursive.

Every major shift in economic history has been accompanied by a migration of scarcity. When a factor becomes abundant, value moves elsewhere. If recursive self-improvement works, even partially, it does not merely accelerate intelligence. It accelerates the search for whatever intelligence cannot produce.

That is why Anthropic’s announcement matters. Not because Claude may someday build Claude, but because it gives us a glimpse of where scarcity goes next and where value and moats are migrating: Further away from the Model layer, more firmly in the Harness, and even closer towards the Orchestration layer. And, naturally, to compute.

Recursive self-improvement as a tremor in the Discontinuity

Let’s start with two important bits of context before I dive back into Claude and RSI

Between September 2024 and February 2026, I counted six tremors: intelligence arriving, intelligence becoming cheap, silicon independence, protocol standardization, the inference swarm, and the long-context frontier. They arrived separately and reinforced one another into a single structural break.

Recursive self-improvement is the seventh. And yet, its character is different.

The earlier shocks expanded the supply of intelligence, made it cheaper, or taught it to coordinate. RSI changes the machine that produces the intelligence in the first place, the engine of the whole sequence. If it starts building better versions of itself, everything compounds.

Anthropic is careful to frame this as a fork, not a fait accompli, and sketches three scenarios:

Scenario 1: The trend stalls. The gains plateau and capability diffuses without a runaway.

Scenario 2: The one we are visibly in, the efficiency gains compound but humans still set the direction, choosing which problems matter and judging the results.

Scenario 3: The loop closes. AI designs and trains its own successor, end to end, and the pace of progress becomes bounded only by the availability of compute.

The real question is the distance between the second and third scenarios. Because that when the human stops being the one who decides.. The seventh tremor is the crossing toward the third.

The Institute does not ask us to take this progress on faith. The numbers are disclosed. As of May 2026, more than 80% of the code Anthropic merges is written by Claude. That’s up from low single digits before Claude Code shipped in February 2025. Anthropic leadership has put the looser figure north of 90%. The typical engineer merges roughly eight times as much code per day as in 2024. In April, Claude shipped over 800 fixes that cut one class of API errors a thousandfold, work the supervising engineer estimated would have taken a human four years.

Lines of code are a crude proxy, and Anthropic says so itself. The median engineer self-reports only ~4× uplift, and Anthropic cites METR’s finding that developers overestimate AI productivity gains. A competing arXiv analysis (2602.04836) argues the METR curve may not be cleanly exponential. None of this changes the direction - code authorship >80%, quality at parity and rising- but the argument rests on the direction and the mechanism, not on any single headline multiple.

Now let me add the second bit of context: the Three Rings.

If those six tremors turned machines into actors, the central question of the Agentic Era is: Who directs them? The Agentic Enterprise is organized around three rings: Intelligence, the Harness, and the Orchestration Layer.

When applying Orchestration Economics, the job is to determine where a company sits within the Three Rings and if it is capable of moving further out to capture more value, because that is the whole game. Viewed through this lens, Anthropic’s SRI disclosure tells us the following:

The most obvious leap is in Ring 1: Intelligence, the cognitive substrate. Hand Claude some model-training code and ask it to make it faster while passing the same checks. In May 2025, Opus 4 averaged a 3x speedup. By April 2026, Mythos Preview was hitting 52x in four to eight hours, compared to the 4x a skilled human needed in the same time. On that narrow, verifiable task, the model went from helpful to superhuman in under a year. Measured from the outside, METR finds the task length a model can complete unsupervised, now doubling roughly every four months, up from seven. Opus 4.6 handles twelve-hour tasks, and Mythos Preview reportedly ran for sixteen.

The more interesting leap is in Ring 2: The Harness. This is the routing engine that decides where every unit of work flows. The Harness decomposes a goal, spawns sub-agents, manages state, retrieves memory, and reviews output. The substrate, the intelligence, is commoditizing. The model is not the moat. The labs know this, which is why they are no longer primarily focused on building chat interfaces. Instead, they are racing to build the best Harness.

RSI is what lets them build it. The Harness is code, the one substance RSI has proven it can improve. A self-improving model can autonomously write, test, and optimize its own integration script, its own pointers, the connectors that wire it into every tool, system, and sub-agent, grading each version against a clean metric: did the task complete, faster, and more coherently?

When 80% of Anthropic’s merged code is Claude’s, much of it is not the model. You do not write a model in a pull request. It is the Harness Claude Code built by Claude.

The model is becoming the engine that wires itself into everything and then optimizes the wiring. That is what it means to become the central routing engine of the orchestration graph: not a smarter chatbot, but the node that decides where work goes and is increasingly written by the thing it routes.

We saw how deep this runs in late March 2026 when Anthropic accidentally exposed the source code of Claude Code, some 1,900 files, and a full agent runtime. This revealed the orchestration thesis rendered in code. Its most telling component is the pointer system: a small, always-loaded index governing what the model loads, fetches, or forgets across a long task, routing-and-memory logic, a self-improving model can rewrite against its own metrics by lunchtime.

That is the agent improving the scaffolding that runs the agent, and it is how the lab compounds its central role in the orchestration graph.

Claude is not only helping improve future versions of Claude. It is also helping improve the software that surrounds Claude: the routing systems, memory layers, evaluation frameworks, and orchestration logic that determine how intelligence is deployed.

These are two different forms of recursive self-improvement, and they operate on very different clocks.

Improving the model itself requires new training runs, new compute, and months of iteration. Improving the orchestration layer requires only code, tests, and evaluation. One loop is slow. The other can run continuously. This gives RSI two loops at different speeds.

The inner loop of the model-improving model is gated by compute and training runs. These are slow, expensive, and episodic. The outer loop of model improving the Harness is gated only by code and evaluation. These are fast, cheap, continuous, and ship between releases without a pre-training. The outer wheel turns faster, and it is where the lab banks a position, because the model commoditizes.

Ring 1 is converging across five vendors, while the routing engine is proprietary. A model that improves the system that coordinates models is the most valuable self-improvement the labs have, and the least visible on a leaderboard.

However, recursive self-improvement only works where success can be measured automatically.

This distinction is crucial because it means the important boundary is not between the model and the orchestration layer. RSI can improve both. A model can evaluate whether code runs faster. It can evaluate whether an agent completed a task. It can evaluate whether memory retrieval has improved.

The boundary sits between the orchestration layer and operational reality.

Ring 3 is where we find operational context, which has no impartial scorer. Whether a decision was right lives in the world, arrives on the slow clock of realized outcomes, and is owned by whoever runs the operation. The machine-turned-actor cannot retrospectively evaluate whether an insurance policy was correctly underwritten, whether a supply chain decision was optimal, or whether a loan should have been approved.

Those outcomes are determined in the real world. There are no clean, verifiable metrics to be optimized. One of the keys to Anthropic’s rapid progress has been understanding that these loops need such metrics, but they also need to run in narrowly defined boxes to allow for that feedback. Not only does Ring 3 not lend itself to tidy metrics, but it also doesn’t fit neatly into such a box.

What RSI changes and where we are

Before considering the economic implications, it is worth being precise about what RSI does and what it does not.

First, RSI is not AGI or superintelligence. It is the engine you would use to reach either. It is a claim about the rate of development. For instance, how much of the AI R&D loop can be delegated to AI? It is not about the resulting system being generally intelligent or conscious. You can believe in compounding RSI while staying agnostic about AGI timelines. I do.

Second, it is not new. Google DeepMind’s AlphaEvolve, announced May 2025, was a genuine precursor, an evolutionary coding agent pairing Gemini with automated evaluators that edged past Strassen’s 1969 matrix-multiplication record, clawed back ~0.7% of Google’s worldwide compute, and improved the training of the very models underneath it. AI improving its own substrate is not a 2026 invention.

Third, the tidy “phases” framing presented by Anthropic in the paper is misleading.

Anthropic presents recursive self-improvement as the next stage in a progression from coding agents to orchestration and then to AI systems capable of improving their own successors. The framework is useful, but the boundaries are less clean than the diagram suggests.

First, coding agents are real. This is no longer speculative. The evidence is now overwhelming that frontier models can autonomously write, debug, test, and improve code across increasingly long horizons. Claude Code, OpenAI’s coding systems, and Google’s AlphaEvolve all demonstrate that AI can already contribute meaningfully to software development. We are not waiting for coding agents. We are living through them.

Second, autonomous agents are underway but unfinished. The industry’s real race is no longer to build a better chatbot but to build the Harness: the routing, memory, agent coordination, and workflow infrastructure that turns intelligence into action. This is the transition I described in the Orchestration Economics Manifesto. We are clearly moving into this phase, but the architecture is still being contested. Every frontier lab is building its own orchestration layer. The protocols are still evolving. The center of the orchestration graph has not yet been claimed.

Third, recursive self-improvement lies further along the same continuum. The evidence is increasingly compelling that AI can improve parts of the systems around it and, under tightly controlled conditions, improve components of the AI development process itself. But this remains the least mature of the three stages. Much of the evidence comes from coding environments and tasks with clear evaluators. The direction is real. The degree remains uncertain.

This is why I think the notion of neatly separated phases is misleading. Coding agents, orchestration, and recursive self-improvement are not successive rooms in a hallway. They overlap. We are already inside the first, actively building the second, and beginning to glimpse the third.

We should also not confuse demonstrations with deployments. The lab’s striking result, which found agents recovering 97% of an alignment-research gap that two humans recovered only 23% of by designing every experiment themselves, is a research artifact with caveats. It did not transfer to production-scale models. And humans chose the problem and wrote the rubric.

Lessons from Anthropic’s Project Vend last year offer the corrective. Claude, running an actual store as “Claudius,” and with no context or guardrails, lost money, sold tungsten cubes at a loss, and at one point insisted it was a human in a blazer. “We would not hire Claudius,” the researchers wrote. Drilled on procedures, it later ran respectably, until fresh red-teamers fed it falsified documents and walked it into debt.

The capability and deployment frontiers are not the same line.

The true costs of RSI

This brings us back to the question of what RSI potentially does for the labs’ economics.

The framing that RSI slashes R&D cost is only half right. Yes, RSI compresses the human labor share of research, such as salaries and engineer-hours. But the dominant line in a frontier lab’s R&D budget was never salaries. It is compute. RSI does not cut the compute bill. It re-denominates the entire cost base onto compute, and then it makes that bill bigger. The 97% experiment cost $18,000 in compute precisely because the human time fell away. The unit of production has changed from a paid researcher to a running GPU.

Making a researcher almost free does not reduce compute demand. It multiplies it. When an experiment is bottlenecked by scarce human judgment, you run a few at a time. When judgment is cheap and parallelizable, you run hundreds, each consuming compute, each competing for the same clusters.

The Institute is explicit in its third scenario. If the loop closes, “the pace of progress becomes determined entirely by the availability of compute.” So, RSI does not relieve the labs’ compute hunger. It intensifies it. A self-improving lab is a more compute-constrained lab, not a less constrained one. That is why the same companies publishing RSI papers are signing the largest power-and-silicon contracts in corporate history. The unit of value moved, and it moved to the one input that only a handful of entities on earth can secure at scale.

The change that genuinely matters, though, is not in the cost structure. It is in the position of the human. And this represents a major consequence.

Until now, the human sat at the center of the development loop: deciding, designing, executing, and judging. RSI moves the human to the edge, from operator to reviewer, and then toward observer. Doing now costs almost nothing in human time. What remains is taste, judgment, and direction-setting, and even that is eroding. On which next step to take in a stuck research session, Opus 4.5 beat the human choice 51% of the time last November. By April, Mythos Preview was at 64%. Anthropic states the trajectory plainly: once human and AI code reach parity, humans stop writing and only review. If they cannot review as fast as Claude generates, the review itself becomes the bottleneck. And then it is dropped.

Houston, we have a problem.

When the system that builds the system runs faster than the humans nominally supervising it, human review stops being a control and becomes a ceremony. Recursive self-improvement potentially becomes recursive error and harm propagation, flaws no human inspected, compounding into the next generation, the unreviewed one helped build. We are, for the first time, mass-producing the producer and removing the human from its quality control.

That is an audit, liability, and governance problem. And it sits under every enterprise that will deploy these systems into regulated work. The Institute, to its credit, says this aloud, ties it to alignment, and floats a verifiable slowdown. This was one of the “implications” that the institute raised in its paper. And while it’s certainly a legitimate one, the sci-fi scary AI nature of it was naturally catnip for mainstream media conversations.

What does that mean in the AI race?

Strip away the drama, though, and RSI confirms the core thesis of Orchestration Economics. It reinforces the race to two scarcities and even accelerates both.

The first is the operational context. RSI accelerates abundant intelligence and perfects the Harness that channels it. Abundance is not value. When a factor of production is commoditized, value migrates to whatever the abundant thing cannot produce itself. RSI cannot produce Ring 3: the insurer’s ten million adjudicated claims, the carrier’s billion routed shipments, the bank’s trillion dollars of underwriting. That has no automatic scorer, because the scorer is the world. The faster the core commoditizes, the more obviously the durable surplus sits in context and proximity to intent.

The second scarcity is compute, and we have just seen why RSI tightens rather than loosens it. Anthropic has, in roughly a year, assembled commitments from Amazon (up to 5 GW), Google and Broadcom (multiple gigawatts from 2027), Microsoft and Nvidia ($30bn of Azure capacity), and recently taken over the Colossus 1 data center in Memphis, now with Google! The orchestration leader is renting the scarcest asset of the orchestration laggard. Compute the flows toward the company that earned the upstream position.

Beyond both bottlenecks, RSI also accelerates deployment. A self-improving Harness gets better, autonomously, at the two things that have always slowed enterprise AI: mapping a customer’s systems and writing integrations. Automated context mapping and dynamic integration mean an agent can survey an enterprise’s schemas, tools, and workflows and wire itself in days rather than quarters. But it does not lower the cost of connecting to the operational context, nor does it transfer ownership of it. The agent lays the pipe faster. The enterprise still owns the water, and faster integration only deepens its dependence on the routing layer controlled by the labs. Deployment accelerates; the moat does not move.

And for Anthropic?

This is the question the IPO will turn on.

First, it is a well-timed piece of communication: unreported internal numbers, published into an IPO race weeks after the Karpathy hire, telling the market the engine compounds at the moment it is being priced. The research is real and the candor genuine. A clear-eyed reader holds both thoughts: it is both true and calculated in terms of timing.

Second, and more important: RSI is not a moat. Every frontier lab has the loop: AlphaEvolve proves Google’s; OpenAI runs its own. A capability your two largest competitors also have is table stakes, not a moat. RSI is a moat amplifier. It compounds whatever position you already hold by pouring its fastest, cheapest gains into Ring 2.

The real question is whether the position RSI amplifies is defensible. The answer is a qualified, conditional yes. Anthropic’s position is the enterprise: roughly 40% share by the trackers, eight of the Fortune 10 as customers, a coding wedge that became the orchestration harness at the center of the routing graph. RSI compounds that lead on the fast loop, and it is a better moat than the model, because the model commoditizes, and the routing engine, for now, does not.

But “for now” it is doing real work. The Harness is contestable. Every lab is building one. MCP is open and donated to the Linux Foundation. Owning the center of the orchestration graph is a land-grab still in progress. RSI raises the stakes in that land grab. It does not win it on its own.

The one layer that would be an unassailable moat is Ring 3. But that demands an operational context that RSI structurally cannot deliver to a lab for the same reason it cannot improve it. There is no scorer that the machine can run. Even maximal RSI leaves the deepest scarcity with the enterprise that generated it.

So, RSI is an accelerant for Anthropic on a real but contested advantage at the routing layer, and a reminder that the durable surplus sits one ring beyond the lab’s current reach. The moat question reduces to a race Anthropic has not yet won: can it convert a harness lead into ownership of the orchestration graph’s center before the model commoditizes and enterprises wall off their own context? RSI buys speed in that race; it does not end it.

There is a tell in the document itself: Anthropic foregrounds oversight, governance, and the option of a verifiable pause precisely because its buyer is the regulated enterprise, where those are procurement requirements. That is as much a positioning statement as a research note. But positioning is not a moat either.

Has anything fundamental changed, then? Not for anyone reading the fault line rather than the headlines. RSI is the seventh tremor, not a new earthquake. It accelerates the commoditization of abundant intelligence. It perfects the Harness, which was always contestable. It shifts the binding constraint onto compute. And it leaves the durable surplus where the Manifesto put it: in the operational context of the outermost ring, the one ring RSI cannot reach, because it is the one layer with no in-silico scorer.

The machine is starting to build the machine. That is genuinely new, genuinely unsettling, and I will not pretend otherwise, least of all when it comes to the human stepping out of the loop. Machines are becoming actors faster than ever. But ultimately, the real value is still captured by whoever directs them and owns the context that tells them what “correct” means.

For now, at least, those are also humanity’s moats.

The views and opinions expressed in this Website are those of the author alone and are based on publicly available information. The expressed views and opinions do not constitute investment advice, a solicitation, or a recommendation to buy or sell any security or financial instrument.

The author may hold positions in the securities of companies mentioned. Certain companies referenced may be current or former clients of, or counterparties to, the author or affiliated entities; such relationships will be disclosed where applicable.

Past performance is not indicative of future results. To the fullest extent permitted by applicable law, the author does not accept any liability for any loss or damage arising from reliance on this content. Readers should conduct their own independent due diligence and consult a qualified financial advisor before making any investment decision.

This is the latest excerpt from AGNT: The Orchestration Economics Manifesto - An Investment Framework for the Agentic Era. Each Thursday, I explore a major theme of the Manifesto and unpack the frameworks, adding extra context with more recent developments. Note: The figures and sequential references are taken directly from the larger Manifesto.

Subscribe now

Read more

TL;DR: Today, we are publishing our “A Durable Growth Moat™ Assessment of the AI Segment of SpaceX - ‘AI’” for Institutional clients. In addition, we will be sending the executive summary to paid subscribers. While the SpaceX public offering represents a potentially historic IPO and has been widely scrutinized, our agentic frameworks offer a different analysis of the company’s structural position in the Agentic Era.

Subscribe now

Read more

This is the latest excerpt from AGNT: The Orchestration Economics Manifesto - An Investment Framework for the Agentic Era. Each Thursday, I explore a major theme of the Manifesto and unpack the frameworks, adding extra context with more recent developments. Note: The figures and sequential references are taken directly from the larger Manifesto.

Subscribe now

Read more

I joined Solenne Niedercorn-Desouches on the Finscale podcast to discuss:

--> How agentic AI is fundamentally rewriting business models, competitive moats, and investment frameworks.

--> Why orchestration rather than the models will be the true moat in AI.

--> Why traditional SaaS defensibility is collapsing in the “SaaSpocalypse,” and how investors need to rethink company value through the lens of architectural resilience.

Read more

TL; DR

• The king. Anthropic was worth $61.5 billion at Series E in March 2025. It is closing approximately $50 billion of primary capital at a post-money valuation of roughly $900 billion in May 2026. The Wall Street Journal reports that Anthropic is on track for its first profitable quarter in Q2 2026, with $559 million in operating profit on $10.9 billion in revenue, implying a 5.1% operating margin. The SpaceX S-1, filed May 20, 2026, disclosed that Anthropic could pay xAI up to $40 billion through May 2029 for the full 300 megawatts of capacity at the Colossus I data center and has an agreement for capacity at Colossus II. By every operational and financial measure, Anthropic is king of the first agentic cycle.

• The thesis. The trillion-dollar question now: can Anthropic defend the throne? Its success rests on its ability to successfully do three things simultaneously: leverage its current position at the Harness layer to win the orchestrator role, t…

Read more

This is the latest excerpt from AGNT: The Orchestration Economics Manifesto - An Investment Framework for the Agentic Era. Each Thursday, I explore a major theme of the Manifesto and unpack the frameworks, adding extra context with more recent developments. Note: The figures and sequential references are taken directly from the larger Manifesto.

Subscribe now

Read more

A follow-up to last week’s piece, now that the deal is live. A deep dive into whether the wafer-scale chip is built for the inference economy that has actually arrived.

Cerebras’s wafer-scale architecture was designed for monolithic, high-precision, single-model inference at extreme speed. The Inference Economy that has materialized in 2026 is heterogeneous, orchestrated, multi-model, context-heavy, and CPU-centric. This creates a structural architectural mismatch. Three misalignments limit Cerebras’ durable moat. At ~$66 billion, the stock trades 3-4× our fair-value range of $17-20 billion. The Inference Economy is real. Whether Cerebras’s architecture is positioned to capture its economic surplus remains an open question.

Subscribe now

Read more

This is the latest excerpt from AGNT: The Orchestration Economics Manifesto - An Investment Framework for the Agentic Era. Each Thursday, I explore a major theme of the Manifesto and unpack the frameworks, adding extra context with more recent developments. Note: The figures and sequential references are taken directly from the larger Manifesto.

Subscribe now

Read more

Decoding Discontinuity today released a report on Cerebras Systems’ S-1 in advance of its IPO this week for Institutional clients. Find more information on our Institutional Client Services here.

Read more

Cerebras Systems S-1 Teardown

Durable Growth Moat Analysis

Why the Wafer-Scale Moat Does Not Survive the Inference Economy

Decoding Discontinuity | May 2026

As the AI infrastructure cycle enters its next phase, the economics of inference are beginning to reorganize the competitive structure of the entire stack.

Our latest institutional teardown examines Cerebras Systems ahead of its IPO through the lens of inference efficiency, orchestration economics, and the growing compute crunch reshaping AI infrastructure markets.

For two years, the dominant infrastructure narrative assumed that larger models, larger clusters, and increasingly specialized silicon architectures would compound into durable pricing power as AI demand accelerated. The Cerebras IPO arrives precisely as that assumption is beginning to face a different economic reality: the future AI economy may scale less through unconstrained frontier inference consumption and more through inference efficiency.

This report examines why that distinction matters.

The teardown explores:

• the structural shift from capability-maximization to efficiency-maximization,

• the growing role of orchestration systems in compressing frontier inference demand,

• NVIDIA’s absorption of decode optimization directly into the CUDA ecosystem,

• the narrowing economic surface area for specialized inference architectures,

• and the implications for durable moat formation in the Agentic Era.

The analysis positions Cerebras not simply as a semiconductor company, but as one of the first public-market tests of whether premium frontier infrastructure economics remain durable as the AI stack becomes increasingly routing-native and heterogeneous.

The report includes:

• detailed architectural analysis,

• moat durability assessment,

• and broader implications for the future structure of the AI economy.

Institutional Research

Decoding Discontinuity provides proprietary research for institutional investors, operators, and strategic decision-makers navigating AI-driven discontinuity.

Our institutional work focuses on:

• AI infrastructure and compute markets,

• frontier model economics,

• inference efficiency and orchestration systems,

• GPU, custom silicon, and cloud dynamics,

• AI-native business models,

• and value migration across the agentic stack.

Unlike our public essays and Substack analysis, Institutional Research delivers company-level teardowns, strategic frameworks, technical assessments, and market intelligence developed for professional use.

Institutional Access

Institutional Research includes:

• proprietary research reports,

• strategic briefings,

• direct analyst access,

• and custom research discussions.

Access is limited and managed directly.

To learn more about Institutional Research or request access to the Cerebras teardown:

Raphaëlle d’Ornano
Founder & CEO, Decoding Discontinuity
raphaelle@decodingdiscontinuity.com

or

Mathieu Huet
COO, Decoding Discontinuity
mathieu@decodingdiscontinuity.com

00:00 - Intro

00:01 - Exploring the Tech Landscape: Insights and Market Dynamics

02:22 - The Inference Economy Explained

04:48 - Hyperscalers' Business Models

07:55 - Challenges in AI Pricing Models

10:55 - AI Infrastructure and Compute Needs

14:24 - Hyperscalers and Economic Value

16:38 - Strategic Tech Giants: Google's Moats & Microsoft's Case Study

19:15 - AWS and Amazon's Strategy

21:53 - Title: Strategic Insights and Reflections on Tech Giants

AGNT: Where Enterprise Tech Meets The Agentic Era is a podcast partnership between Decoding Discontinuity, SiliconAngle, and NYSE Wired. Every episode, co-hosts Gemma Allen and Raphaëlle d’Ornano unpack how intelligent systems are reshaping companies, markets, and the way real work gets done. From Fortune 500 boardrooms to breakout upstarts, they dig into the strategies, technologies, and people defining the next chapter of AI.

Find the Podcast here:

Spotify

Apple

YouTube: https://www.youtube.com/@siliconangle/videos

Gemma Allen is a New York-based writer, on-camera host, and communications and marketing consultant focused on technology, capital markets, and leadership. She has nearly 20 years of experience in the technology sector, advising C-suite executives on international growth, go-to-market strategy, and corporate positioning across the U.S. and Europe.

Allen is a producer and co-host for SiliconANGLE & theCUBE. And she has been a Forbes contributor since 2023. Allen writes and speaks about how emerging technologies, including AI, are reshaping work, leadership, and everyday life, with a particular focus on women in business. Her work also examines female change agents influencing technology, finance, sport, and culture, including founders, investors, and public figures such as Kamala Harris, Caitlin Clark, Allyson Felix, and Barbie.

Allen is a regular on-camera host at the New York Stock Exchange, where she interviews CEOs, founders, and investors on innovation, scale, and market-moving trends. She also writes for Nasdaq on the intersection of technology and culture.

Raised in Ireland, Allen is an alumnus of University College Dublin and is an active member of the Irish-American business community. Follow her for reporting grounded in the lived realities of female employees, founders, leaders, and working mothers.

Links

• Forbes

• Twitter/X

• LinkedIn

SiliconANGLE Media Inc. is based in Palo Alto and Boston. We focus on enterprise technology, digital transformation, and cultures of innovation. We cover cloud, AI, mobile, social, and emerging topics, including IoT, autonomous vehicles, blockchain, and others. Our brands include theCUBE.net, our video operation, SiliconANGLE.com, our publishing operation, and theCUBEResearch.com, our analyst services operation.

Links

• SiliconANGLE

• theCUBE Events

• theCUBE Research

• Twitter/X

• Facebook

• LinkedIn

• CUBE365 (Agile Video Software)

Quoted:

The flurry of moves are signs that private fund managers understand how AI presents a fundamental reconfiguration of the way companies do business and deliver value, said Raphaëlle d’Ornano, founder and managing partner at D’Ornano + Co., a private markets advisory and strategic consulting firm. Every company is becoming an “agentic enterprise,” and will require an “orchestration” process to allow it to unlock a new layer of value provided by the abundance of artificial intelligence tools, she argued in a recent newsletter post.

Private equity firms are “amongst the biggest employers in the United States... so it is absolutely imperative that every single company within private equity portfolios, small and big, transform themselves into this new reality,” she said in an interview.

The way to do that is to develop agents that are effectively grounded in the right business and enterprise context, D’Ornano added.

“Today, there is a missing layer for all this to happen,” she said. “Th…

Read more

This is the latest excerpt from AGNT: The Orchestration Economics Manifesto - An Investment Framework for the Agentic Era. Each Thursday, I explore a major theme of the Manifesto and unpack the frameworks, adding extra context with more recent developments. Note: The figures and sequential references are taken directly from the larger Manifesto.

Subscribe now

Read more