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. “There’s a gap in the market.”

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 software industry entered 2024 in what looked like a renaissance. AI launches pushed SaaS valuations higher. ServiceTitan’s IPO succeeded. Salesforce announced Agentforce. Investors believed the next growth cycle had begun. Beneath the optimism, something else was happening.

The software industry that began 2025 celebrating AI-driven growth ended the year confronting the narrative that AI-driven growth reflected in their top line mattered less than the perceived vulnerability to their fundamental business model. Over the course of early 2026, the surface cracked open. ServiceNow, Salesforce, Workday, and Adobe each lost between 6% and 15% of market capitalization in a single week of trading in February 2026.

That fallout continues. Each new release by Anthropic sends shudders through the markets. This week, its new financial plugins hit financial stocks. Meanwhile, Anthropic experienced 80x growth in Q1 2026, compared to its own 10x projection. That created a problem of compute constraint that suddenly seemed to be solved by a shocking deal with SpaceX-xAI to lease compute.

These events seem almost incomprehensible compared to everything the tech industry has experienced over the past 30 years. That was the primary motivation for publishing AGNT: The Orchestration Economics Manifesto, a framework for what is becoming the largest value dislocation within the S&P 500 in modern memory.

The transition into the Agentic Era is not a faster version of the software cycle we already know how to analyze. It is a discontinuity, a break in the operative logic by which work is organized, coordination is conducted, and economic power becomes durable. Disruption rearranges positions inside an existing order. A discontinuity alters the architecture itself.

This Manifesto is my attempt to make that new architecture legible, and to name where durable value will accrue in the era now opening.

Share

Discontinuity is not Disruption

The era of generative and agentic AI is not disruption. It is discontinuity.

I know it is tempting to dismiss these as buzzwords or shallow consultant shorthand. Disruption. Discontinuity. Paradigm. Several decades of overuse have stretched and flattened such words, robbing them of meaning. But these words once signified very precise concepts that shaped how we understood technological change. That is why, when I began writing three years ago about generative and then agentic AI, I chose the word discontinuity. Not because it sounds dramatic. But because it is technically accurate, and the distinction between disruption and discontinuity is load-bearing for everything that follows. I believe we must reclaim that meaning to help navigate the present and guide us toward the future as it unfolds.

Disruption follows predictable patterns. New entrants target overlooked segments, gradually moving upmarket until incumbents fall. An existing curve steepens or flattens, but the curve continues. Leaders and investors can deploy existing tools to analyze the shift. This was the foundational insight of Clayton Christensen’s Innovator’s Dilemma. He first coined the phrase in 1995, one year after the Netscape browser was released, and his framing for thinking about technological disruption became a cornerstone for the next three decades of innovation and entrepreneurship.

Discontinuity goes beyond disruption. If disruption changes the slope of a curve, Discontinuity breaks the curve entirely. All previous strategies, tools, and historical lessons are ripped away. The break can lead to extraordinary value creation or extraordinary destruction.

When cloud computing disrupted on-premises software, it was a radical change in terms of pricing and technology. And yet, it was still CRM or payroll. People using it and installing it had reference points from their experiences. That is disruption.

Generative and agentic AI broke this pattern. Companies are not being handed a new CRM or payroll platform. Someone is giving them a magic wand and saying, “Here, it can do anything. Imagine what you can do with this”. Nothing in their experience has prepared them for the magic-wand scenario. They have no mental models.

So, they turn back to the old ones and stick the magic wand into the old tech stack, wishing for increased ROI that never materializes. Only 5% of enterprises reported any ROI from their initial generative AI projects last summer when MIT conducted a first-of-a-kind survey. As the study’s authors noted, the failure was not “driven by model quality or regulation but seems to be determined by approach”.

This is not surprising. Most investors and executives optimize for disruption because they can deploy familiar analytical frameworks, frameworks that project cash flows, calculate returns, and assess competitive dynamics using tools refined over decades. Discontinuity places them in a landscape where everything they once knew has lost its power to guide them.

French philosopher Gaston Bachelard described “epistemological rupture” as moments when progress requires abandoning old frameworks entirely, because accumulated knowledge has become an obstacle to understanding.

This is the core difficulty. The greatest obstacle to understanding discontinuity is not the absence of new information. It is the presence of old frameworks that interpret new evidence as confirmation of the regime that is ending.

This discomfort produces two failure modes. The first is paralysis: waiting for clarity that will arrive too late. The second is pattern-matching to previous cycles: applying frameworks that are no longer relevant and making catastrophic allocation decisions.

Frustration grows because the discontinuity has not been recognized. Meanwhile, technology's capabilities are rapidly advancing. Those who have recognized the discontinuity are somehow, almost inexplicably to others, transforming their businesses at unimaginable velocity.

This inability to build new mental models is a defining characteristic of discontinuity. Those unable to reorient themselves are at grave risk of being on the wrong side of a historic economic valuation shift.

The Human Challenge and the Scale of Time

This is about tech. But it’s a mistake to think that it’s just about tech. Such moments present a fundamental human challenge: People’s brains are not framed for discontinuity.

Navigating discontinuity requires an abstraction capability, the ability to see how all the pieces might eventually align and what consequences could be unleashed. Generative and agentic AI have the power to turn many non-technology companies into technology companies, an evolution many industries should start recognizing now.

To understand and adapt to discontinuity requires an understanding of the underlying components driving this phenomenon, a clear vision of how that impacts an existing business model, and a sophisticated view of the structures and assets that are realigning or emerging to enable something new.

This has been at the very core of my work for the past three years. And if there is one critical aspect that must be recognized to begin building the right mental model for this discontinuity, it is understanding the scale of time. The generative and agentic AI transformation is far more profound and sweeping than even many sophisticated technology leaders have truly grasped. That means it will play out over a much longer time frame than investors and entrepreneurs are accustomed to tolerating. This shift requires everyone across the innovation economy to fundamentally reset the way they think about building new companies, making investments, and defining the shape of markets. This moment demands a long-term view to map out which investments and technologies will make sense at which moment and why.

Amid the rubble of the dot-com bust, a certain defeatism set in. Maybe all the internet hype had really been just that: hype. But many Silicon Valley VCs began discussing a book written a decade earlier by Science Historian David Nye called “Electrifying America: Social Meanings of a New Technology”. The book traces the many ways in which electricity transformed the nation.

Cities installed electric streetlights, which expanded nightlife and led to a wave of new restaurants opening. To increase demand for electricity in homes, they designed appliances such as refrigerators and washing machines. No one could have envisioned most of these innovations when Thomas Edison made his first lightbulb in 1879. It took decades to put the infrastructure in place, raise investment capital, develop use cases, and refine business models.

The real transformation played out over a much longer timescale as all the pieces were put into place. Such was the case with digital transformation. Though it may have taken longer than initially predicted, today we are living in the digital world many envisioned in 2000, thanks to smartphones, 4G wireless, broadband, and a host of business-model innovations enabled by these waves of infrastructure.

When dealing with discontinuity, the scale of change is so massive that timelines are impossible to predict. So many pieces must align before exponential potential is unleashed.

Then, when it happens, the speed of change is startling.

This pattern repeats. A breakthrough technology appears. Euphoric predictions about sweeping changes are made. They fail to manifest. Doubt appears. Progress seems to stall. And then, in a blink, everything accelerates with little warning. We are in that phase now.

Anyone in the tech industry who experienced 2025 felt every extreme of this cycle. One moment, there was anxiety about an AI bubble. Then there was fear of overspending on infrastructure. Then the pendulum swung, and the market believed a niche product announcement crushed billions of dollars in market value for incumbent SaaS companies. Now, in reverse, we are all panicked that we are running out of compute and assessing whether compute constraints constitute or not a single point of failure (SPOF) for the AI labs.

OpenAI CEO Sam Altman captured the difficulty of reconciling these extremes when he said in August 2025: “Are we in a phase where investors as a whole are overexcited about AI? My opinion is yes. Is AI the most important thing to happen in a very long time? My opinion is also yes”.

The bubble skeptics are correct that many companies in this space will fail, just as many internet companies failed in 2000. They are wrong to conclude that the transition itself is speculative.

The Capability Curve Broke First

This discontinuity creates the foundation for something new: abundant intelligence, accessible through APIs, improving exponentially.

Companies that relied on linear improvement in AI capabilities, expecting gradual progress from models that could barely complete sentences to models that could write paragraphs, now suddenly faced step-function changes that break all projection curves.

This is not about doing existing things 10% better. When ChatGPT was first released in November 20226, it could describe a concept. By 2025, AI could architect a solution, write the code, build the interface, and deploy a functional application autonomously. By March 2026, AI could coordinate teams of sub-agents that divide complex professional work across domains, execute in parallel, sustain reasoning across a million tokens of context, operate computers through their own interfaces, and correct their own errors over hours of unsupervised operation. Three years. Three different worlds.

Importantly, this capability was proven and reproducible across multiple vendors. It was not one model, one company, or one approach. Frontier performance converged: multiple models achieve 80% on SWE-Bench (Software Engineering Benchmark), which evaluates large language models by testing their performance on real-world software engineering tasks. Multiple coding tools are capturing billions in revenue. The leaderboard positions for the latest frontier models change from day to day, week to week. The progression is accelerating in a way fundamentally different from previous AI winters.

The gap between capability announcement and production deployment has been compressed from years to months. GPT-4’s initial capabilities were rolled out to a limited set of users over several months in 2023 with tight restrictions, and then gradually expanded to broader availability over the next year. In sharp contrast, when Google released Gemini 3 Pro in November 2025, it was available globally and already baked into many of its most widely used products. What once required a year of cautious rollout now ships in a morning.

Models are reaching PhD-level reasoning as a foundational capability. Not as a research artifact, but as a commercially available infrastructure.

Then the Substrate Commoditized

As AI capabilities advanced rapidly, a group of researchers s in late 2023 proposed the graduate-level Google-Proof Q&A (GPQA) benchmark, based on 448 PhD-level questions from multiple disciplines. A more rigorous variation based on a subset of 198 questions emerged, known as GPQA Diamond, where human experts scored 69.7%²⁴. The first commercially available model did not surpass that human threshold until December 2024. Eighteen months later, multiple models from competing laboratories exceeded 90%.

Meanwhile, the cost per query had fallen to fractions of a cent. Peer-reviewed research from Epoch AI found that the price to achieve a given level of frontier performance was declining by a factor of 5-10x per year. Roughly a third of that decline was attributable to algorithmic efficiency alone.

The convergence was as significant as the cost trajectory. On the Massive Multitask Language Understanding (MMLU) benchmark, the standard measure of general knowledge across academic subjects created in 2020, the gap between proprietary frontier models and their open-source alternatives collapsed from approximately 15 percentage points to less than 1 percentage point by late 2025.

Each month brought a new release that narrowed the distance further. DeepSeek, Kimi K2, and Llama achieved near-parity with systems that cost orders of magnitude more to build. Intelligence that was once the exclusive province of three or four laboratories became available to anyone with an API key or a sufficiently capable laptop. When the substrate becomes uniform, differentiation migrates to what is built upon it. This is not a theoretical observation. It is the mechanism that drives everything that follows.

Abundant intelligence is the necessary condition. The discontinuity is what that abundance made possible: machines that act.

Then Machines Became Actors

The break occurred when intelligence became sufficient for autonomous action, when machines crossed from processing instructions to pursuing goals. This was a qualitative change in what software is. Every previous generation of software, including every previous generation of AI, operated within the same architectural relationship to human work: the human decided, the tool executed.

The tool could be astonishingly capable. AlphaGo defeated the world Go champion ten years ago. It did not decide to play Go. GPT-4 could draft a legal brief. It did not decide which brief to draft, did not select the relevant precedents from the firm’s case history, did not coordinate with the associate handling discovery, and did not file the brief with the court. The human performed every coordination function. The AI performed one step, brilliantly, and waited for the next instruction.

Between late 2024 and early 2026, that architecture broke.

The evidence is not ambiguous. When Claude Code receives a task such as “resolve this GitHub issue”, it reads the codebase, identifies the relevant files, reasons about the bug, writes a patch, runs the test suite, discovers the patch introduced a regression, rewrites the patch, reruns the tests, and submits a pull request. No human selected the files. No human ran the tests. No human identified the regression. The agent received a goal and autonomously determined how to achieve it.

That is not a better tool. That is a new kind of actor.

The distinction matters because it determines the correct analytical framework. If machines are better tools, the old frameworks apply with modest adjustment: productivity improves, margins expand, the curve steepens. The analyst’s DCF model works. The board’s strategic plan holds. The investment bank’s sector thesis survives with updated assumptions.

If machines are actors, the frameworks break. The unit of production changes. The cost structure of work changes. The competitive dynamics of every industry that employs knowledge workers change. The analyst needs a new model, not updated assumptions.

That is what Kuhn meant by a paradigm shift. For those who still associate “paradigm shift” with Kuhn, the phrase can almost conjure a mental image of gliding from one framework to another. What is often overlooked is an equally important part of Kuhn’s insight into what catalyzed these shifts. Within scientific communities, an existing paradigm would begin to exhibit a series of “anomalies”. These observations were beyond the framework's ability to explain, plunging it into a period of crisis.

The “paradigm shift” often involved a messy, heated middle period full of disagreement over direction, rules, and ways to explain and interpret the new information and dynamics. Many schools of competing thought would emerge trying to define the rules of the new paradigm, a necessary, if sometimes brutal process. While Kuhn spoke of “anomalies” in this, we have chosen a different, precise word to guide us: discontinuity.

To navigate through discontinuity and into the new paradigm, it is important to recognize that the old observations are wrong. They are incommensurable. The old framework cannot produce the new observations. The data that would explain the shift do not fit into any field in the old model.

Consider the shock when software companies lost hundreds of billions in value in early February 2026. Entire sectors were repriced on the assumption that something fundamental had changed. They were right. But the shift was not caused by a single model release or a single company. It emerged from a cascade of changes across the AI stack in cost, capability, infrastructure, coordination, and deployment that arrived within months of one another. Each was significant alone.

Together, they produced the first structural signal of the Agentic Era: intelligence behaving like a commodity. When intelligence commoditizes, value does not disappear. It migrates.

But to where, to whom, and on what terms?

Those questions are the core of this manifesto. Coding was the first domain to cross, but its lessons generalize to any domain involving structured reasoning, tool use, and verifiable outputs. The crossing order is predictable. The crossing itself is not in question.

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. 7 with Gemma Allen & Raphaëlle d'Ornano

00:00 - Intro
00:01 - Exploring Recent Developments: The Altman-Musk Case and Beyond
02:10 - IPOs and Market Impacts
04:46 - OpenAI Internal Conflicts
07:34 - Navigating the Intersection of Technology and Geopolitics
13:58 - SaaS Market Challenges
18:38 - The Impact of Agentic AI on Financial Markets
23:58 - Future Horizons: Opportunities, Expectations, and Concluding Insights

TL;DR: The infrastructure cycle is still in the acceleration phase. But for most hyperscalers, Capex is buying a utility that risks commoditization, while the agentic transition is simultaneously destroying the moats of the high-margin businesses those same companies are trying to protect. Meta, Amazon, Google, and Microsoft represent c.20% of the S&P 500.

Subscribe now

The four largest hyperscalers reported quarterly earnings within hours of each other last week. They each beat consensus and raised 2026 capital spending guidance, committing between $650 and $700 billion to AI infrastructure for the year. That’s nearly double what these companies spent in 2025 and arguably the largest concentrated infrastructure cycle in the technology sector's history. Morgan Stanley extended the count to five (adding Oracle) and raised its projection up to $805 billion this year, up from $765 billion, an amount one analyst noted would be equivalent to all spending by non-tech S&P 500 companies in 2025.

This spending has now persisted long enough to confirm an unprecedented cash-flow crunch. Amazon’s trailing twelve-month free cash flow collapsed from $26 billion to $1.2 billion, a 95% decline. CEO Andy Jassy used his annual shareholders’ letter to defend the trajectory in cash-cycle terms. AWS, he wrote, “has to lay out cash for land, power, buildings, chips, servers, and networking gear in advance of when we can monetize it,” with Capex growth meaningfully outpacing revenue growth in the high-growth phase, while the underlying assets have useful lives of five to thirty-plus years.

The bull case for Capex spending ultimately rests on the argument that returns arrive later, after monetization outpaces the buildout. As Jassy reminded investors in that same letter, AWS investment was being questioned internally as recently as the 2014 operating plan review, when a senior leader opened the discussion by asking, “Tell me again why we’re doing this business?”

The problem is that it’s hard to see when that “after” arrives. In my view, these Capex numbers are set to keep growing as the Inference Economy accelerates. The question of “will the spend pay off?” will be a perennial one.

In this context, the market needs to scrutinize the multiples applied to companies whose valuations represent ~20% of the S&P 500 to determine whether they are consistent with the new reality of hyperscalers in the Agentic Era. That starts with stepping back and recognizing the broader circumstances facing these companies.

The Agentic Era is recomposing the tech stack and challenging the value pockets of the hyperscalers. As a result, they are getting squeezed from above and below.

The AI labs are cannibalizing the value that accrues at the infrastructure layer by adding a new layer, the Harness, atop the intelligence itself. All that money being poured into this layer will only produce commodity returns.

Meanwhile, the intelligence they produce threatens the core of the hyperscalers' application value. The agentic paradigm is dismantling the application-layer advantages that have justified their historic valuations.

The architectural and financial profile that earned hyperscalers their historic multiples is mutating. This implies a structural mispricing of a large portion of the market.

However, it would be a mistake to assume that all hyperscalers are doomed or can even be evaluated in the same way. Because that assumes they are spending on the same thing, building the same kind of moat, and facing the same competitive landscape. They are not. The agentic transition is acting on each layer of the technology stack in opposite directions, simultaneously, sorting these companies into different valuation regimes that the consensus has not absorbed.

Share

The financial mutation

For two decades, the defining characteristic of the largest technology companies was capital efficiency, with asset-light platforms converting modest infrastructure investment into extraordinary returns. The cloud was a virtuous cousin of the software business. It was more capital-intensive than pure SaaS but operated at gross margins of 70% and produced the cash flows that made these the most enviable corporate financial profiles in modern markets. This is the financial profile that earned hyperscalers their multiples.

That profile is now in transition. 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, and have repeatedly flagged supply bottlenecks on recent earnings calls.

Operating cash flow is no longer sufficient to fund the buildout. Combined hyperscaler debt issuance was $121 billion across the Big Five in 2025, compared to an annual average of $28 billion between 2020 and 2024. Morgan Stanley estimated that hyperscalers will borrow ~$400 billion in 2026, double the $165 billion borrowed in 2025. JPMorgan projects $1.5 trillion in AI-related bond issuance over the next five years. In February 2026, Alphabet issued a 100-year bond, the first century bond from a technology company since Motorola in 1997. The $32 billion multi-currency raise was nearly tenfold oversubscribed. Pension funds and life insurers, institutions that match liabilities over generations, bought Google paper at 6.05 percent as if Alphabet were sovereign-grade infrastructure rather than the parent of a search engine.

The hyperscalers have, so far, avoided negative free cash flow. The technical fact is increasingly trivial. Pivotal Research projects Alphabet’s 2026 free cash flow will fall roughly 90 percent, from $73 billion to perhaps $8 billion. Barclays projects Meta’s free cash flow will drop 90 percent in 2026, then turn negative in 2027 and 2028. The accounting reveals the strain. Amazon shortened the useful life of a subset of its servers from six to five years in January 2025, while Microsoft extended the same assumptions in fiscal 2023, adding $3.7 billion to operating income from the change alone. The industry cannot agree on whether its core infrastructure lasts one year or six.

What was once the most capital-efficient sector in modern history has become one of the most capital-intensive. The question of the moat that results from this spend is now acute: who is benefiting from this spend?

Decomposing the value stack

Each of these firms is, in functional terms, a stack of four layers operating under a single corporate roof.

Layer 0 is silicon.

Google’s TPUs are the most advanced (now in their 8th generation), Amazon’s Trainium and Inferentia are gaining traction for inference workloads, while Meta’s MTIA, in partnership with Broadcom, and Microsoft’s Maia remain solid but less differentiated to date. Amazon’s Jessy insisted on the positive contribution of the company’s booming chip business in his latest investor letter: “Our chips business is on fire, changes the economics for AWS, and will be much larger than most think (…) Having our own hotly demanded AI chip opens up many possibilities, but perhaps none larger than the ability to lower costs for customers and secure better economics for AWS. At scale, we expect Trainium will save us tens of billions of Capex dollars per year, and provide several hundred basis points of operating margin advantage versus relying on others’ chips for inference.

Our annual revenue run rate for our chips business (inclusive of Graviton, Trainium, and Nitro-our EC2 NIC) is now over $20 billion, and growing triple digit percentages YoY. To dimensionalize this versus other chips companies, that run rate is somewhat understated by our currently only monetizing our chips through EC2. If our chips business was a stand-alone business, and sold chips produced this year to AWS and other third parties (as other leading chips companies do), our annual run rate would be ~$50 billion. There’s so much demand for our chips that it’s quite possible we’ll sell racks of them to third parties in the future.”

Layer 1 is intelligence.

The foundation models. Here, Google is the only company producing frontier intelligence, matching the AI labs and the Chinese open-source competitors. Gemini 3.1 Pro Preview is in the top 3 of Artificial Analysis Intelligence Index, while Meta’s most recent model, Muse Spark, ranks in the top 5, and Amazon’s Nova only ranks 22^nd.

Layer 2 is the technical orchestration substrate.

This is what the industry now calls the Harness. The infrastructure that decomposes goals into tasks, delegates them to specialist agents, manages state, and accumulates learning across sessions. Functionally, it is the AI compute layer plus the orchestration runtime that sits immediately above it. The hyperscalers historically assumed they would own this layer, as they owned the SaaS-era developer toolchain. They are discovering that the Harness is parting ways from the compute beneath it and that the margin is traveling with the Harness, not the compute.

Layer 3 is the “application.”

These are the user-facing surfaces where intent originates and value is captured. Search, Office, advertising, retail, consumer hardware, enterprise SaaS.

The agentic transition is acting on these layers in different directions. Layer 0 is becoming structurally important for cost and latency reasons. Layer 1 is consolidating into a small number of frontier providers and commoditizing rapidly underneath them. Layer 2 is the layer absorbing all of the announced Capex, essentially. For three of the four hyperscalers, it is becoming a utility. Layer 3 is splitting between surfaces that survive the transition and surfaces that do not.

The two sections that follow examine each layer.

Layer 2 becomes a utility

The hyperscaler thesis was that the cloud business would absorb AI workloads the way it absorbed every prior generation of software workloads. Rent the compute, sell the developer toolchain that sits on top, capture both layers as a single economic surface. In the SaaS era, this was correct. The cloud and the orchestration software were sold together. The high-margin orchestration tooling was bundled with the lower-margin compute. The hyperscaler captured the full vertical and earned a software multiple on the result.

Layer 2 is the orchestration substrate where the bulk of the announced Capex is landing. It is now being pulled apart from both sides. The model labs are taking the runtime from above. NVIDIA is taking the layer just above the silicon from below.

The labs first: As I argued in Anthropic’s Digital Labor Tax, Claude Managed Agents, launched in early April 2026, is an architectural pivot away from token-based pricing entirely. Anthropic now bundles compute, reasoning, and execution into what I call the Agent Unit: a synthetic colleague billed by the hour. The product runs on existing cloud infrastructure, but the developer never touches it. Anthropic abstracts the substrate into a primitive, charges by the session-hour, and captures the margin between raw compute cost and orchestrated outcome. By owning both the model and the runtime, Anthropic positions itself as the tax collector of digital labor, regardless of whose cloud the labor runs on. OpenAI is building the same architecture under Frontier and Codex.

This is the only path through which AI labs build a durable moat. Intelligence alone is commoditizing. Gemini, Claude, GPT, and the open-source frontier are converging on benchmarks at near-parity. The lab that owns the runtime through which intelligence becomes operational captures the coordination surface, the customer relationship, and the pricing power.

And the labs are not stopping at the runtime. Cowork, Frontier, and the vertical plugins (Claude for Finance, Claude Design) are the labs’ move further up the stack into Layer 3, where intent is captured, and outcomes are delivered. The Harness is a beachhead, not a destination. And this remains true as frontier labs push prices higher.

NVIDIA second: As I argued in NVIDIA’s CUDA Playbook Has a Second Act, the March 2026 Agent Toolkit, which included Nemotron, OpenShell, and AI-Q Blueprint, is the second-act CUDA play, providing a portable, open-source runtime layer immediately above the silicon. The toolkit is model-agnostic and compatible with MCP without depending on it, preventing Anthropic from establishing protocol-level lock-in. NVIDIA is making the bet that if it controls the runtime and the chips, it does not need to own the model or the application.

As a result, the AI compute sold by the hyperscalers, and that is enabled by the unprecedented Capex allocation, could start to resemble a utility as the AI labs and Nvidia capture the actual “AI” margin, with the valuable orchestration layer.

The bull case for the April 29 Capex hike rests heavily on Q1 commentary that AI margins are better than legacy cloud margins. The financial statements support this for now. Microsoft’s commercial cloud gross margin decreased 100 basis points sequentially, AWS operating margin printed at 38%, and Google Cloud at 33%.

This is short-lived. The strong AI margins notably reflect HBM scarcity passed through as price. These are scarcity rents on a window that closes mechanically as new HBM capacity arrives in 2027 and 2028. Strip out the scarcity rent, and AI compute starts to resemble regular compute. The equation does not hold.

There is one structural exception: a hyperscaler that owns its own Layer 1. In that case, the lab and the cloud are the same entity, the Harness is captured internally, and the margin migration that pressures the others does not occur. This is precisely Google’s position. Gemini, TPU, and Vertex Agent Builder all sit inside Alphabet, so the same capital allocation that is being structurally weakened at the other three is structurally reinforced at Google.

Microsoft, Amazon, and Meta are the three hyperscalers without their own Layer 1. Their strategic path forward is to demonstrate value at the Harness level rather than the compute level. That means building orchestration products credible enough to capture the session, the outcome, and the agent-hour, rather than just renting the GPU underneath. Microsoft’s Copilot Studio and AWS’s Bedrock AgentCore are first attempts. Neither has yet displaced the labs in workflows where the lab also offers a Harness.

The arithmetic is unforgiving. Compute infrastructure earning commodity returns supports utility multiples, perhaps 12 to 15 times earnings. The hyperscalers without Layer 1 are currently priced as software companies in the 20-to-30-times range. That premium rests substantially on what we term Layer 3, the application franchise. Which is precisely the layer at risk in the Agentic Era.

Layer 3: when the operator becomes an agent

The application layer was supposed to be the safe ground. Even if Layer 2 was fragmented, Layer 3 was where the historical moats lived: the Office franchise, the Search index, the ad-targeting machine, the retail relationship. The consensus defense of hyperscaler multiples rests substantially on this assumption.

It is correct for some Layer 3 surfaces and demonstrably wrong for others. The dividing line is precise: the moats that survive are those whose value lies beneath the interface. The moats that break are the ones whose value lives in the interface.

The agentic transition changes who operates the system. For the entirety of the prior era, every interface, every workflow, every UX investment was architected for a human operator. The user knew the keyboard shortcuts, navigated the menu, and understood the conventions. The user was trained on Word in school and could not switch to a different word processor without paying retraining costs. The historical Layer 3 moats were anchored in the human operator’s tacit knowledge of the interface, which earned a software multiple precisely because of that lock-in.

When the operator becomes an agent, the moat depends on what sits behind it.

Agents do not need to be trained, do not learn keyboard shortcuts, and are not certified. They operate on any available interface by screenshotting, parsing the DOM, or calling the underlying functionality programmatically. The infrastructure that makes this concrete is now in production: OpenAI’s Operator, Anthropic’s Computer Use, and Google’s Project Mariner (now integrated into other Google products vs. standalone) are general-purpose agents that operate web interfaces directly. Every SaaS application accessible through a browser is now automatable by any party, whether or not the vendor consents. The browser is now a universal, consent-free automation surface for general-purpose AI agents.

Where the moat is the interface itself, that is the end of it. Where the moat is a physical or operational substrate that the interface only renders, agents expand the volume flowing through it.

Microsoft is down 14% year-to-date. It now trades at 24.65x trailing earnings. The consensus sees a cyclical correction and a buying opportunity.

The consensus may be right about the cyclical part and profoundly wrong about the structural part. To understand why, you have to grasp what is actually dying. It is not Microsoft’s software. It is the user interface as a workspace.

Consider how a financial model gets built today. The old way: you open Excel, you build the model, you format the output. Microsoft captures your intent because your hands are on its keyboard. The entire value chain, from input to processing to output, lives inside one company’s interface.

Now consider the emerging alternative. You give an orchestrator a high-level objective. It spawns a swarm of specialized agents. One fetches the data from the SQL database. One runs the analysis in Python. One formats the deck. These agents interact with Microsoft’s APIs, not its interfaces. Excel may still exist in this workflow, but as a rendering layer rather than a work surface. The user interface becomes a status dashboard for work happening elsewhere.

This distinction between a source system, where work begins, and a target system, where finished work is deposited, is the crux of the structural challenge.

Microsoft’s moat was never primarily about the quality of Word or Excel. It was that Microsoft was the source system for enterprise knowledge work. Every workflow started there. What Claude Code, Cowork, and NVIDIA’s OpenShell runtime represent is the migration of source-system authority to the orchestration layer.

GitHub still stores the code, but it has become a target system, the place where the agent deposits the pull request, not the place where the developer forms the intent. Office still formats the presentation, but it is a target system, the place where the agent saves the .pptx, not the place where the analysis happened. The hundred billion dollars Microsoft has invested over decades in user experience becomes a legacy cost rather than a competitive advantage once work happens in the runtime layer, not the interface layer.

It is an architectural shift in where intent is captured. Claude Code reached $2.5 billion in annual recurring revenue within nine months. Cursor crossed $2 billion in three months. These are enterprise products that are redirecting source-system authority away from the incumbents and toward the model providers.

The economic consequences compound under the per-seat pricing model. Agent swarms are designed for autonomous, parallel execution. NVIDIA’s AI-Q hybrid architecture demonstrates that these swarms can cut execution costs by 50% while doing the work of multiple human roles. If a swarm orchestrated by a third party replaces a department of fifty people, Microsoft loses fifty subscription seats. The per-seat model becomes inversely correlated with the technology’s success: more productive AI means fewer humans, which means fewer licenses.

That is a structural headwind that compounds over the next few years, and changes the source of the company’s moat.

The counterweight is distribution

Microsoft’s enterprise distribution surface includes Azure Active Directory as the identity layer, Teams as the communications fabric, Windows on the endpoint, the global enterprise sales force, and the existing CIO relationships. This remains unparalleled and is harder to disintermediate than the Office work surface. Whatever wins at Layer 2 still has to be sold, deployed, and governed inside enterprises that already run on Microsoft’s identity and security plumbing. Copilot is an interface play and will be contested. Azure-as-the-distribution-substrate-for-other-people’s-agents is a different proposition, and one that an Anthropic locked into an $30 billion Azure contract through 2029 is materially funding. The question for Microsoft is whether the distribution surface compounds fast enough to offset the work-surface erosion. And whether the business model changes that result from it prove to be convincing.

Take Google now, for example.

Google splits internally. The algorithmic infrastructure beneath search includes the index, the crawl, and the rank. It is agent-hardened because the agent has to query the index to answer the user’s question. The advertising business model layered on top is structurally exposed.

Pew Research found that Google’s AI Overviews had already reduced link click-through rates from 15% to 8%. Zero-click searches now account for 56% to 69% of all queries since AI Overviews launched. When an agent reads the AI Overview and synthesizes an answer, the sponsored link below it is never seen. The advertising auction prices the click. The click is disappearing.

For now, AI features are actually helping drive more queries and longer sessions (AI Mode queries are ~3× longer). Google has also started placing ads above, below, and sometimes within AI Overviews and testing them in AI Mode. So the ad business isn’t collapsing. It’s adapting and still printing money. The structural risk, however, remains unchanged.

The bifurcation produces an unusual investment characteristic: Google is the only hyperscaler with both the structurally strongest Layer 1 position (Gemini at the frontier, TPU integrated) and a structurally exposed Layer 3 revenue line (search ads). The capital allocation for the first is being underwritten by the second's cash flow. The transition risk is whether Gemini-based monetization (Vertex, agentic commerce, enterprise) ramps fast enough to offset the slow erosion of CPCs in core search. The bull case is that it does, and that Google ends up with a Layer 1+2 franchise that no other hyperscaler can match. The bear case is that the erosion runs faster than the offset.

The right question

The hyperscalers built the infrastructure used to train the models. They provided the cloud capacity, the data pipelines, and the GPU clusters. They also invested billions in the AI companies that now threaten their application-layer businesses. Amazon and Google are among Anthropic’s largest backers. Microsoft funded OpenAI with $13 billion. They are, in the most literal sense, raising the children who are pushing them out of the house.

And yet, Anthropic is one of Microsoft’s largest Azure customers. The emerging orchestrators depend entirely on the hyperscalers’ infrastructure. Anthropic alone has contracted approximately $80 billion in cloud costs through 2029. If Anthropic succeeds in capturing the orchestration premium from Microsoft’s application businesses, Microsoft simultaneously loses high-quality application revenue and becomes increasingly dependent on lower-quality infrastructure revenue. The money circulates but changes character. It enters as orchestration. It exists as plumbing. The parent funds the child who displaces it. The child pays rent on the way out.

NVIDIA’s Agent Toolkit highlights the paradox. It makes cloud infrastructure fungible for agent swarms by enabling any workload on any cloud, routed on millisecond-level spot pricing. Yet, NVIDIA erodes the switching costs that justify premium cloud pricing. Good for NVIDIA in the short term: more orchestrators, more GPU demand. Potentially catastrophic in the medium term: if the infrastructure built with those GPUs earns commodity returns, the incentive to keep buying GPUs at premium prices evaporates. Everyone is building the road that leads to someone else’s destination.

The Capex debate, as currently conducted, asks whether the spending is too much or too little. It is the wrong frame. It assumes the spend is buying the same thing across companies. Once the spend is decomposed by layer, the question becomes whether the layer being purchased holds its margin in an era where agents, not humans, are the operators. And whether the company doing the spending owns the adjacent layers needed to make the spending defensible.

The answer is not the same for the four hyperscalers. Google owns Layer 1, captures the Harness internally, and pays for the buildout with a Layer 3 revenue line that is eroding but not collapsing. Amazon owns a fulfillment substrate that agents amplify, and its Layer 0 chip business is now monetizing internally at scale. Meta owns an attention substrate (Layer 3) that is largely agent-resilient, with a narrower exposure to commerce and an open-source hedge. At Layer 3, Microsoft owns a distribution surface that compounds and a work-surface franchise that does not.

The market is currently pricing all four as if Layer 2 commoditization and Layer 3 erosion hit them symmetrically. They will not.

The recognition the framework predicts has likely been delayed by one specific gap: there is no clean public-market comparable for an integrated Layer 1+2+3 platform that is not also a hyperscaler. Anthropic’s reported preparation for an October 2026 IPO matters here. A liquid, priced comparable for an orchestration company expanding from the Harness into the application layer, will create a reference point that does not currently exist. When it arrives, the hyperscaler multiples will be repriced against it, not in isolation.

The Capex was never the question. The question, properly asked, is which layer the Capex is buying, whether that layer holds its margin once the operator shift completes, and whether the company doing the spending owns the adjacent layers needed to make the spending defensible. For most of the announced Capex at three of these four companies, the answer is: only partially. That is what the multiple has to absorb.

The recognition has not arrived in full. The conditions for its arrival are now in place.

For the broader framework: Orchestration Economics lays out the value-capture logic; The AGNT Manifesto places this hyperscaler reckoning inside the wider thesis on how the Agentic Era reorders enterprise value.

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.

Today, I am publishing AGNT: The Orchestration Economics Manifesto, a framework for what is becoming the largest value dislocation within the S&P 500 in modern memory.

For three years, I have been writing about a fault line forming beneath the software economy.

At first, the signals arrived as separate events: a model crossing a capability threshold, a cost curve breaking open, a defensible business becoming structurally exposed in a single quarter. The instinct was to read each shock locally. But the pattern was never local. The shocks were expressions of the same pressure moving through the same boundary.

That boundary is what this Manifesto is about.

The transition into the Agentic Era is not a faster version of the software cycle we already know how to analyze. It is a discontinuity, a break in the operative logic by which work is organized, coordination is conducted, and economic power becomes durable. Disruption rearranges positions inside an existing order. A discontinuity alters the architecture itself.

This Manifesto is my attempt to make that new architecture legible, and to name where durable value will accrue in the era now opening.

It rests on five claims:

1. The marginal cost of cognition is collapsing. The collapse is non-linear.

A task that costs $100 in human cognitive labor today will cost $10 in 2027, $1 in 2028, and pennies in 2029.

2. As cognition becomes abundant, machines shift from tools to actors.

For the first time, software receives goals, takes autonomous action, and delivers outcomes. The unit of economic production changes from the human labor hour to the orchestrated workflow.

3. Orchestration creates a new economic layer. The companies that control it capture the surplus of the entire transition.

The entity that reorganizes itself around a new architecture - with intelligence at the core of workflows - and delivers outcomes from irreplaceable operational context through the Orchestration Layer captures the surplus of the transition. That entity is the AGNT.

4. Three structural laws predict who wins, who loses, and who the market is mispricing.

Proximity to intent. Context depth. Workflow intelligence. Satisfy all three: you are an AGNT, the captain of the Agentic Era. Miss one: you are just a member of the crew.

5. Every company carries an embedded option that the market cannot yet price.

Every company holds two value curves inside a single stock price: the fundamentals curve and the orchestration curve. The gap between them is an embedded option that runs in both directions. The destruction will be priced in days. The creation will be priced in years.

These five claims are the spine of the Manifesto. Each is developed across the full text: the substrate that made cognition collapse, the shift from tools to actors, the emergence of the Orchestration Layer, the laws of durable value capture, and the hierarchy of firms, sectors, and positions that will fracture, fuse, or be repriced in the crossing.

The stakes

This is not the software-versus-semis trade the public market has settled into. The dislocation runs through every sector of the economy. The same lens reprices software and semis themselves on terms the consensus has not yet considered.

Of the 391 companies in the S&P 500 exposed to the agentic transition, the Manifesto identifies approximately 125 emerging as durable orchestrators and roughly 250 facing structural value compression. Twice as many losers as winners, sorted not by sector but by architectural position.

The first violent expression of this sorting was the SaaSpocalypse of late January/February 2026: $285 billion in software market capitalization was erased in 48 hours. That was not the event. That was the first negative options resolving. The positive options resolve over years, not days, which is precisely why the dispersion is becoming structural rather than cyclical.

Inside that dislocation sit three mispricings the income statement has yet to reveal. 

The unrecognized long: a company carrying a large positive orchestration option, the market has priced at zero, because the architectural position is strengthening beneath a sector-wide selloff. 
The unrecognized short: a company whose financials still print clean even as agents route around its system of record and the seat-compression cascade approaches. 
The phantom orchestrator: a company priced for orchestration, it does not have the architectural position to deliver.

These three mispricings generate longs, shorts, and pair trades in equal measure. They cannot be identified through financial data, which lags. They can be identified through structural assessment, which leads.

And then there is the largest pricing question of the cycle: the AI labs themselves, headed for IPO.

Anthropic last priced at $380 billion against $14 billion ARR. That’s a 27x revenue multiple. If the labs are model companies selling intelligence into a market where intelligence is converging, that multiple is an aggressive bet on a commoditizing asset. If they are platform companies selling orchestration, accumulated enterprise context, and switching costs, 27x may be the entry price for a generational franchise. Revenue from API consumption and revenue from platform orchestration appear identical on a financial statement. They are not identical in their implications for margin trajectory, competitive durability, or terminal value. The Manifesto provides the diagnostic. It also names the two structural single points of failure - open-source compression from below, the compute trap from above - that determine whether the labs cross before model commoditization erodes the intelligence advantage that gives them the right to orchestrate. This is a race against the clock.

For the investor, the allocator, and the operator, this is the moment the framework begins to matter. The architecture is already shifting. The repricing has already started. The question is whether the next move is read through old instruments or new ones.

The Manifesto is now live here.

What comes next

Beginning next Thursday, and for the weeks that follow, I will open this argument here on Substack. I will proceed through a guided unfolding of the framework, presenting one structural theme at a time.

Some of the ideas first appeared in this publication over the past eighteen months. What matters now is that they can be seen from within a more complete structure. A thought first published as an essay is not the same thought once the architecture around it has been built.

The regular newsletter will continue its usual rhythm. Alongside it, every Thursday for the coming weeks, you will receive a piece of the Manifesto, excavated from within the whole, sharpened by the vantage point the whole now makes possible.

We will begin with the concept of discontinuity and explore the difference between a world disrupted and one whose governing architecture has already begun to break.

- Raphaëlle

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.

• Six major model releases over ten days reordered the AI landscape: Anthropic’s Opus 4.7 and OpenAI’s GPT-5.5 traded blows with Kimi K2.6, DeepSeek V4 Pro/Flash, Qwen3.6-27B, and Xiaomi MiMo. The closed labs still hold the apex on reliability and frontier iteration. The substrate has decisively gone open.

• The cost-quality ratio has inverted. DeepSeek V4 Pro is roughly one-seventh the blended cost of GPT-5.5 and one-sixth of Claude Opus 4.7, an order-of-magnitude shift in deployment economics.

• Kimi K2.6 shipped swarm orchestration trained into the weights via Parallel-Agent Reinforcement Learning, not implemented in the runtime above them. Capabilities that lived in the proprietary Harness layer twelve months ago are now native to open-weight models.

• The runtime harness still matters more than ever for production reliability, auditability, and compliance. But it matters now for operational rigor rather than capability scarcity. The engineering tax that protects it is collapsing.

• The moat has moved. As intelligence converges and commoditizes, value migrates to the orchestrator: the entity that holds operational context, workflow ownership, and the relationship between humans and work. The AGNT Manifesto, to be published on Thursday, builds the investment framework for that migration.

Subscribe now

A flurry of major new model releases over the past ten days has scrambled the AI landscape once again.

Anthropic opened the cycle on April 16 with Claude Opus 4.7. Moonshot answered on April 20 with Kimi K2.6, the first open-weight model to top the Artificial Analysis Intelligence Index. Two days later, Alibaba’s Qwen team released Qwen3.6-27B, and Xiaomi released MiMo-V2.5-Pro, the latter tying with K2.6 at the top of the open-weight tier, though it remains API-only for now. OpenAI shipped GPT-5.5 on April 23. DeepSeek closed the cycle on April 24 with V4 Pro and V4 Flash, both Mixture-of-Experts models under the MIT license, that bring the lab within striking distance of the frontier.

Fifteen months ago, when DeepSeek-R1 arrived at one-fiftieth the cost of OpenAI’s o1, I argued that what looked like a Chinese price cut was a punctuated-equilibrium event: long periods of stability interrupted by rapid, structural change. R1 commoditized intelligence. What just shipped does something more consequential: it commoditizes the layer above it. The cost of training a frontier model fell in 2025. The cost of deploying autonomous agents in production has fallen over the past 10 days.

The standard reading of these releases is that “open-source caught up.” That is correct on the leaderboard and misleading on the substance. Open-weight models have not closed the gap on long-horizon agentic reliability, multimodal grounding depth, or instruction-following discipline under adversarial conditions. The Stanford AI Index 2026 reports that the closed-versus-open Chatbot Arena gap partially reopened to roughly three points by March 2026, after compressing from eight in early 2024 to under two a year later. The closed labs still hold the apex. But two things are now simultaneously true, and their interaction is the structural news of the cycle.

The first is that open-weight models are absorbing the architectural innovations that defined the closed agentic stack twelve months ago: long-context inference at economically viable pricing, multi-agent coordination trained into the weights rather than implemented in the runtime above them. The Harness is the proprietary runtime layer that I argued last spring was the durable moat in the Agentic Era, and that I described more recently as Anthropic’s Metalayer. This is being absorbed, capability by capability, into the open-weight substrate.

The second is that the cost-quality ratio has inverted. DeepSeek V4 Pro is priced at $1.74 per million input tokens and $3.48 per million output. That’s roughly one-seventh the blended cost of GPT-5.5 ($5 input, $30 output) and about one-sixth that of Claude Opus 4.7 ($5 input, $25 output).

The frontier advantage still belongs to the closed labs. The economic advantage increasingly does not.

Together, these shifts undermine the labs’ defensibility in raw intelligence and compress the window in which they can convert it into platform durability. Anthropic and OpenAI are now forced to compete on harness quality and orchestration sophistication at exactly the moment those capabilities are being absorbed into the open-weight substrate.

Share

What actually shipped

First, the long-awaited DeepSeek V4. V4 Pro is a 1.6-trillion-parameter Mixture-of-Experts with 49 billion active per token. Flash is the lighter sibling, with 284 billion total and 13 billion active. Both ship a one-million-token context window, a number already available in proprietary models. What is new is that DeepSeek makes it economically viable at open-weight pricing, by shipping two memory-side innovations: Compressed Sparse Attention and Heavily Compressed Attention. These build on V3.2’s DeepSeek Sparse Attention and solve the KV-cache memory wall that previously made long context prohibitive on commodity hardware. Both models are released under the MIT license with no field-of-use restrictions. Note that inference is slower than Western frontier models on a per-token basis, and the model remains text-only. DeepSeek’s research team has signaled multimodal is in flight, but it is not there yet.

Then, Kimi K2.6. This is the more important release for the agentic argument. One trillion total parameters, 32 billion active per token, with a 384-expert MoE topology of which eight are active, and one is shared per token. Native multimodal input - text, image, video - outputting text. A 256,000-token context window, smaller than DeepSeek V4 but more than enough for the median production agent workload. The architectural step that matters is what Moonshot continues to bake into the weights since K2.5 in January: Parallel-Agent Reinforcement Learning, a training regime that produces an orchestrator agent capable of decomposing tasks, instantiating frozen specialist sub-agents, and coordinating concurrent execution at scale: up to three hundred sub-agents and four thousand coordinated steps per session in K2.6.

These are capabilities that, until recently, lived in the Harness layer above the model and are now increasingly trained into the model itself.

Last, Qwen3.6-27B and Xiaomi MiMo-V2.5-Pro. Qwen3.6-27B is a dense 27-billion-parameter model. Every token routes through every parameter under a clean Apache 2.0 license. It scores 46 on the Artificial Analysis Intelligence Index, far above where dense open models sat eighteen months ago, and it is sized to run on a high-end M-series MacBook Pro or a Mac Studio. That makes it the local development loop primitive of the moment. Xiaomi MiMo-V2.5-Pro currently ties with Kimi K2.6 at the top of the AAII open-weight composite at 54, but ships closed-weight via API only. Xiaomi has promised to release an open-source version eventually, but has not committed to a date.

The trajectory of these releases matters a lot here.

Benchmark convergence is not agentic parity. The cost-quality ratio is.

Benchmark convergence is real, but it is not agentic parity. On Chatbot Arena, the gap between proprietary frontier models and their open-weight alternatives compressed from roughly eight points in early 2024 to under two by early 2025, then partially reopened to about three points by March 2026 as the closed labs accelerated. That is the convergence. The frontier-closed labs have not stopped pulling away at the very edge of capability. They have stopped pulling away enough to defend their per-token pricing on the dimensions that matter for production agents.

Benchmarks like MMLU and AAII measure capability on bounded tasks. Agentic deployment is bounded by something different: long-horizon reliability under tool use, coherent state across a thousand calls, adherence to system prompts under adversarial inputs, and graceful recovery from tool failures rather than the unbounded looping that killed AutoGPT in 2023. On these axes, the closed frontier still leads by margins that matter. On Meta’s Gaia2 benchmark, GPT-5 leads Claude 4 Sonnet by eight points on adaptability under ambiguity. On τ-bench, even the strongest models succeed on fewer than 25 percent of retail agent tasks at pass-at-eight. As Princeton’s Narayanan and Kapoor noted in March: “AI agents are getting more capable, but reliability is not improving at the same pace”.

Multimodal grounding tells the same story. DeepSeek V4 is text-only; Kimi K2.6 has vision and video, but vision-grounded reasoning lags; and GPT-5.5 posted a jump on MRCR-v2 from 36.6 to 74 percent at 1 million tokens in a single generation.

Apex parity, then, is not here. UC Berkeley’s Measuring Agents in Production study, published in December, found production teams overwhelmingly default to the best-performing closed-source models regardless of inference cost.

What matters for the substrate question is narrower: sufficiency for the median agentic workload at an economic ratio that flips ROI. That ratio has inverted.

DeepSeek V4 Pro delivers near-frontier performance on long-horizon coding tasks at a fraction of GPT-5.5’s per-token output pricing. Kimi K2.6 ships orchestration in the weights at a similar discount. The cost asymmetry is 10 to nearly 100 times cheaper for inference, depending on the tier and routing. That means even a five-to-ten-per-cent quality gap flips the deployment math for most production workloads.

Domains where reliability tail risk matters more than unit economics will continue to default to the closed frontier. That includes sectors such as high-stakes legal, regulated medical, and autonomous engineering, which will continue to default to the closed frontier. But these have already absorbed the cost. The growth is in the median workload, where a model that is materially worse and fifty times cheaper would not have been built at all under closed-frontier economics.

The calibration follows. The labs hold the apex. The economy will increasingly run on the substrate. Both are simultaneously true. The substrate is what gets built on, not what gets paid for. As of this cycle, the substrate is open.

Share

The architectural shift behind the cost ratio

The cost-quality ratio is the headline. The underlying architectural shift determines whether the ratio holds.

Two innovations matter most: long context at economically viable open-weight pricing, and the migration of orchestration capabilities from the runtime layer into the model itself.

Below the one-million-token threshold, agentic coding is assisted editing. The model helps modify files one window at a time. The human provides the architectural overview, and the work is stitched across context boundaries. Above the threshold, with reliable retrieval, the model holds an entire codebase in working memory, including every file, every dependency, every test, and every configuration. That is a different thing. The model no longer operates on fragments. It understands the system. It refactors across files without losing coherence. It traces failures end-to-end without decomposing them into disconnected steps.

That is autonomous engineering, and until this cycle, it was available only on Claude Opus 4.x and Gemini 3 Pro at frontier prices.

With DeepSeek V4 and CSA/HCA making a million-token context economically viable on commodity hardware, it becomes a primitive that any startup can deploy. The economic threshold for putting coding agents into production drops by an order of magnitude. Adoption stops flowing through enterprise sales cycles and starts flowing through developers from the bottom up.

There is a second-order effect: every long-context interaction becomes in-context learning. The agent that has processed a thousand pull requests for a specific repository develops institutional knowledge that no general-purpose model can synthesize from training data. The system becomes the training ground.

The second shift is more consequential for the labs. Kimi K2.5, released in late January, was the first open-weight foundation model with built-in swarm orchestration: a Parallel-Agent Reinforcement Learning architecture that decomposes complex tasks, dynamically instantiates specialist agents, and coordinates their concurrent execution. K2.6 extends it with native multimodality and the higher ceilings already mentioned.

The point is where the orchestration lives. In the closed-labs’ architecture, swarm orchestration is a runtime concern. The Harness layer above the model handles task decomposition, agent instantiation, and coordination. In Moonshot’s architecture, those capabilities are trained into the weights. The orchestration is not sitting on top of the model. It is fused with it.

That distinction matters for the labs’ durable moats. Twelve months ago, the runtime moat was defensible because building a production-grade Harness required months of backend engineering, frontier-grade reliability, and a primary intelligence layer no open model could match.

Each pillar is now compromised. Long context at open-weight pricing eliminates the reliability premium for context-heavy workflows. Swarm orchestration in the weights eliminates a meaningful share of the runtime engineering. The convergence of intelligence on a uniform capability frontier eliminates the third leg.

The architectural innovations are not auxiliary to the cost story. They are what make the cost story durable.

In-weights orchestration carries its own costs. The capability trained into the model is, by construction, less observable than the capability implemented at runtime above it. Debugging is harder. Domain-specific guardrails are harder to attach. The auditability requirement that every step of a regulated workflow be reconstructable for compliance review is harder to engineer.

The point is not that in-weights orchestration replaces the Harness. The engineering moat, in terms of the difficulty of building one well, is no longer load-bearing on capabilities that the open model itself cannot replicate. The Harness still matters. The reasons it matters have shifted from capability scarcity to operational rigor.

The labs under pressure, and the moat that keeps moving

Twelve months ago, I argued that orchestration is the durable moat in the Agentic Era. I defined this to include the runtime layer that coordinates tool use, manages state, and decomposes goals.

Two weeks ago, I described how Anthropic operationalized that thesis with Claude Managed Agents, launched on April 8: a managed runtime billed by the session-hour at $0.08 above standard token rates, with sandboxed execution, checkpointing, credential vaulting, and end-to-end tracing. I termed the product category the Metalayer.

Both arguments need to be sharpened in light of these announcements.

A year ago, the Harness was a moat because primary-grade reliability, frontier intelligence, and months of backend engineering were each scarce. This cycle compromises each one. Kimi K2.6 carries swarm orchestration in the weights. DeepSeek V4 delivers the long-context capability that was the load-bearing assumption of the closed Harness. Both ship under near-permissive licenses. The runtime advantages of Managed Agents include multi-agent coordination, persistent state, and tool orchestration. These are increasingly available as native model behaviors in customer-controlled environments at a fraction of the cost.

The Harness is still a moat. It is just no longer the moat. It is being squeezed from two sides.

From below, open-weight models are absorbing orchestration capabilities that the labs spent two years building as proprietary infrastructure. Every quarter, the gap between what the closed Harness can do and what an open model in a customer’s VPC can do narrows. The Metalayer’s pricing logic is that the runtime adds enough value to justify a tax above the token layer. However, that depends on the runtime’s defensibility. As that defensibility erodes, the tax becomes contestable.

From above, the entities such as insurers, banks, carriers, industrials, and software incumbents that hold operational context are under no obligation to route through a lab’s Harness when an open-weight equivalent ships every quarter. They can build their own runtime against an open model and own the orchestration outright.

Every operational context holder that opts for open weights becomes a Harness that the lab cannot tax.

This is why the labs cannot stand still. Anthropic moved at an extraordinary cadence across the first quarter: Cowork, multiple model upgrades, Managed Agents, the $30 billion raise at $380 billion, and the Amazon five-gigawatt compute deal. OpenAI shipped GPT-5.3-Codex, 5.4, and then a fully retrained 5.5. These are platform plays. They demonstrate how the labs are climbing further out of the model layer, deeper into the runtime, and reaching for the workflow itself. The alternative is to become a very expensive API in a market where APIs are converging towards free.

The two structural vulnerabilities we identified last summer during a private assessment of Anthropic and OpenAI were open-source compression and the compute trap. These are no longer future risks. They are present-tense pressures.

DeepSeek V4 and Kimi K2.6 just shortened the timer on the first. The flywheel runs faster, the rented ground gets thinner, and the urgency to convert the runtime moat into a workflow moat compounds. The platform multiples that justify $380-billion-class valuations (and up to $1 trillion per the latest data for Anthropic) now require winning the workflow layer against incumbents whose operational context the labs cannot replicate by shipping a better model.

The wave of open-weight releases from Chinese labs is not only a pricing problem for OpenAI and Anthropic. It is the visible expression of a strategic divergence. Hugging Face’s State of Open Source Spring 2026 report shows that Chinese-developed models accounted for 41 percent of platform downloads through February, ahead of US-developed models at 36.5 percent. Singapore’s AI Singapore initiative built its sovereign regional model on Qwen. African enterprises adopt DeepSeek at two to four times the rate of other regions. Permissive licensing via Apache 2.0 from Qwen, MIT from DeepSeek, and modified MIT from Moonshot is the policy lever, not just the pricing lever, reducing dependence on a small number of US providers for sovereign and regulated workloads.

xAI framed Grok-1 around the goal of maximum truth-seeking with broad access. In practice, the cadence of permissive releases now belongs to China, and OpenAI’s gpt-oss release in August 2025 was a response rather than a lead. The substrate going open is broadening the surface of who can build with frontier-class intelligence. That’s good for the field, even where it is uncomfortable for incumbents.

The labs are running uphill on ground that is moving faster than they are climbing. A year ago, orchestration was the moat. Today, orchestration is the contested terrain. Tomorrow, the moat sits one layer further out, where operational context lives, workflow ownership accrues, and intent originates.

The moat keeps moving. The framework has to move with it.

Where value migrates

DeepSeek V4 and Kimi K2.6 are not a story about open-source nationalism, MoE architectures, or the SaaS selloff. They are the visible expression of a structural shift. The substrate of the Agentic Era, already cheap and ubiquitous, has gone open. The architectural innovations that defined the closed stack twelve months ago are arriving in open weights faster than the labs can replace them.

As intelligence converges and commoditizes, value migrates to the orchestrator, to whoever coordinates the abundant resource and captures the surplus released by abundance. The pattern repeats across every major economic transition: electricity, compute, cloud.

Twelve months ago, the orchestrator looked like the Harness. This cycle proves the Harness itself is being absorbed into the substrate alongside the model. Value cannot stop there. It continues until it reaches the layer that cannot be commoditized: where intent originates, where context accumulates, and where work is directed. Operational context. Workflow ownership. The relationship between the human and the work itself, not any tool inside it.

The companies that win the Agentic Era will not be the ones that built the best model, nor the ones that built the best Harness. They will be the ones who reorganized with intelligence at the core of their workflows and applied irreplaceable operational context to the work agents now perform.

The Manifesto I am releasing this week builds the investment framework for that migration. Orchestration Economics identifies the structural position that captures the surplus when intelligence commoditizes: AGNT.

The substrate just went open. Thursday, we map who builds on it, who gets built around, and where the surplus actually accrues.

Speaker on AI, business model resilience, and the economics of the Agentic Era

Raphaëlle d’Ornano is the founder of Decoding Discontinuity and the architect of analytical frameworks, including Durable Growth Moat and ARAF, the Architectural Resilience Assessment Framework. Her work helps investors, executives, and leadership teams understand AI not as another wave of disruption, but as a structural discontinuity that rewrites how value is created, defended, and captured.

Her research focuses on what happens when the marginal cost of cognition collapses, software shifts from tool to actor, and competitive advantage migrates from static moats to orchestration, workflow control, and operational context. She brings a rare combination of market lens, strategic clarity, and practical business relevance to conversations that are often either too technical or too abstract.

Raphaëlle speaks to audiences who need to understand not just what AI can do, but what it changes in the underlying economics of companies, sectors, and capital allocation.

Why audiences book Raphaëlle

Raphaëlle’s talks help leaders and investors answer the questions that matter most now:

• Which business models become stronger as AI agents move closer to execution?

• Which moats are quietly weakening, even when traditional metrics still look healthy?

• Where will value concentrate as orchestration layers emerge across industries?

• How should companies rethink growth, defensibility, and strategic positioning in the agentic era?

She is especially effective with audiences looking for substance over hype: boardrooms, investor gatherings, executive offsites, and high-level conferences where the conversation needs to move beyond generic AI optimism into structural analysis and decision-making.

Speaking formats

• Keynotes for conferences focused on AI, technology, investing, business strategy, and the future of enterprise

• Executive workshops for leadership teams assessing AI’s impact on their sector, operating model, or moat

• Investor briefings for allocators, GPs, and public-market audiences evaluating the long-term repricing effects of AI discontinuity

• Panels and moderated conversations where strategic clarity and original frameworks elevate the discussion

Signature topics

• AI as discontinuity, not disruption

• The collapse in the marginal cost of cognition and what it means for business architecture

• Durable Growth Moat: re-evaluating defensibility in the AI era

• Orchestration Economics: where value accrues when agents become economic actors

• Why the market is still mispricing winners, losers, and transition risk in the Agentic Era

• How workflow intelligence, context depth, and proximity to intent are redefining competitive advantage

Selected appearances

• Milken Institute Middle East & Africa Summit

• Milken Global Conference

• SuperReturn North America

Media and Research

Raphaëlle’s ideas have been featured in leading business and technology publications, including The Information, Fortune, and Forbes. Her ongoing research is published through Decoding Discontinuity, where she analyzes the structural consequences of AI for markets, industries, and business models.

📌 The Information – How SoftBank’s Son Made It Back to the White House

📌 Fortune – Zero-interest-rate babies are facing their day of reckoning. It’s time this generation of startups learns how to fly

📌 Forbes – Is There A Tech Extinction Event Coming?

Contact Information

For speaking inquiries, interviews, or to discuss potential engagements:

📩 For inquiries, availability, and speaker fees, contact:
📧 chris@decodingdiscontinuity.com
📞 WhatsApp: +33 7 62 40 58 19

“The GenAI Investors Playbook offers an insightful exploration into the Advanced Growth Intelligence framework, a vital tool for understanding the nuanced dynamics of disruptive technologies. Investors will find its in-depth analysis indispensable for managing risks and leveraging opportunities in this new, complex environment.” — Gary Survis, Operating Partner, Insight Partners

“[Raphaëlle] is trying to add a dose of sober financial analysis to the hyped-up world of AI, without pouring much cold water on the technology itself.” — Cory Weinberg, Journalist, The Information.

Decoding Discontinuity is a research and investment platform focused on the structural impact of generative and agentic AI.

Through our proprietary frameworks, we analyze how companies are positioned to adapt to generative and agentic AI. Our Durable Growth Moat methodology assesses whether competitive advantages endure through generative and agentic AI discontinuities by examining structural fragility, control layer integrity, and economic resilience.

We apply these frameworks and methodologies to produce exclusive, in-depth teardowns of some of the most high-profile AI-related IPOs for institutional clients. Over the past year, our clients had early intelligence about the durability of S-1s for such companies as:

• Klarna

• CoreWeave

• Chime

• Fermi

• Figma

For investors and corporate leaders seeking to understand the transformative nature of the discontinuity being caused by generative and agentic AI, these report and methodologies offer a playbook for evaluating high-risk AI investments, separati…

Read more

The market is miscategorizing Salesforce. Investors are pricing it as a victim of the agentic transition, sending its stock down ~30% year-to-date. That bundles it with the workflow wrappers being commoditized by the AI labs, transforming into platform plays. And yet at Salesforce’s TDX 2026, the company presented production-scale evidence of something structurally different: an incumbent destroying its own application layer, open-sourcing the governance grammar that sits above MCP’s connectivity layer, and building the orchestration infrastructure the Agentic Era demands. And doing it all at a velocity on a context substrate that the labs cannot replicate. Anthropic escapes model commoditization by orchestrating on frontier capability. Salesforce is mitigating application-bypass risk by orchestrating in the enterprise context. The market recognizes the value of orchestration when labs build it, but applies a blanket SaaS discount when an incumbent …

Read more

Anthropic has inserted a new layer into the AI stack with Claude Managed Agents. We term this managed runtime for autonomous agents the “Metalayer.” It is priced by the session-hour, bundles compute, intelligence, and execution into a single taxable unit. By capturing the Agent Unit, Anthropic positions itself as the essential utility for digital labor: every synthetic colleague, regardless of who builds the last mile, pays the Anthropic “tax”. The Metalayer has the potential to decouple revenue from the linear compute-to-megawatt equation, create software-like stickiness, and accelerate the sorting of enterprise value from technical ability to strategic position. But the product designed to accelerate agentic AI adoption intensifies the very compute bottleneck on which it depends - and the ground beneath it is fragile.

Subscribe now

Read more

Anthropic’s accidental leak wasn’t a security story so much as a strategy document written in TypeScript. What those 512,000 lines reveal is an AI coding tool whose underlying architecture is far more general than the product it currently serves - a domain-agnostic orchestration harness equipped with self-rewriting memory, an unshipped autonomous daemon called KAIROS, and multi-agent delegation models that are already shipping alongside more ambitious coordination capabilities still sitting behind feature flags. The architecture’s most consequential implication is that orchestration and synthetic colleagues are not separate phenomena but two consequences of the same engineering, linked by a causal chain that runs through the memory system now visible in the source - orchestration produces operational intelligence, operational intelligence absorbs the domain expertise moat, and the absorbed moat is precisely what qualifies the agent to perform th…

Read more

In this interview from the AGNT Podcast, recorded at the MCP Dev Summit in Times Square, Raphaelle d'Ornano joins theCUBE + NYSE Wired's Gemma Allen to break down the week's most market-moving developments in AI — from the accidental leak of Anthropic's 512,000-line Claude Code repository to the misread implications of Google's TurboQuant research.

D'Ornano unpacks what the leaked codebase actually reveals: a sophisticated orchestration graph with Claude Code as the entry point for capturing user intent, and an indexing mechanism called the Pointer that solves context entropy by fetching information directly from external sources without inflating the context window.

The conversation also explores the widening gap between Wall Street's interpretation of AI breakthroughs and their technical reality. D'Ornano explains how Google's TurboQuant research — an inference efficiency improvement rooted in papers from a year prior — triggered a sharp selloff in memory stocks despite HBM supply constraints remaining fully intact.

She outlines her investment philosophy around "architectural resilience," evaluating how companies across the AI stack are positioned against rapid innovation cycles. The discussion turns to OpenAI's 35X revenue valuation and the unexpected difficulty of its latest funding round, raising sharper questions about retail investors gaining index-fund exposure to money-losing AI companies as Anthropic, OpenAI, and xAI prepare for landmark IPOs.

From the mismatch between market perception and technical nuance to the democratization risks of AI investing, d'Ornano provides a frank assessment of the industry's most consequential fault lines.

00:00 - Intro

00:01 - Journey Through Innovation: Unveiling Authenticity, Challenges, and Surprises

02:36 - Anthropic's Orchestration Graph: Innovation Unveiled

05:24 - Industry Implications: Shaking the Tech Giants

07:51 - TurboQuant and Market Reactions: A New Memory Paradigm

12:57 - The Mismatch of Market Perception: Understanding Technicalities

16:14 - Changing Investment Paradigms: Messaging in the AI Era

18:34 - OpenAI's Valuation: 35X Revenue and Beyond

22:30 - Transforming Investment and Tech: A New Era in Valuation and Insight

A Google blog post about compressing AI memory went viral. Within 48 hours, the market capitalization of memory semiconductors evaporated by over $100 billion. TurboQuant helps solve a core problem of the Agentic era: making long-context LLM inference efficient. But the algorithm compresses only the inference-time cache, not the model weights, training data, or storage.

Subscribe now

Read more

In this episode of the AGNT Podcast, Raphaelle d'Ornano, analyst and AGNT Podcast co-host, joins theCUBE + NYSE Wired's Gemma Allen to discuss how NVIDIA's GTC announcements are redrawing the battle lines for enterprise software in the agentic AI era.

D'Ornano frames NVIDIA's debut of its agent toolkit as a deliberate echo of the original CUDA strategy — a move designed to make the company the foundational layer for agentic AI and the indispensable infrastructure beneath the entire ecosystem.

She explains why Jensen Huang's approach is intentionally agnostic: whether LLMs or incumbent SaaS platforms win the fight to control the AI orchestration layer, NVIDIA sells more chips either way.

The one scenario that threatens that dominance, d'Ornano notes, is full vertical integration by a hyperscaler supplying its own silicon. The conversation also explores the "SaaSpocalypse" — the argument that AI's disruption of traditional SaaS business models remains unfinished, regardless of macro distractions that have temporarily redirected investor attention.

D'Ornano pushes back on executives who draw historical analogies to Blockbuster or the dot-com era, arguing that the speed of AI progress makes those comparisons dangerously misleading, with change arriving 10 to 100 times faster than any prior technology cycle.

She outlines two non-negotiable moats for software companies under pressure: proximity to human intent and unique, architecturally sound contextual data. Allen adds that what these companies urgently need is an "agentic through line" — a clear, offensive articulation of how they will create and defend value in the paradigm ahead. From the collapse of per-seat pricing to the rise of outcome-based consumption models, the episode provides a sharp framework for how enterprises and software providers can navigate the shift before the window closes. Find more SiliconANGLE news and analysis

00:00 - Intro

00:01 - Exploring AI: From AGNT Podcast Insights to Nvidia GTC Highlights

04:08 - Nvidia's Strategy and Software Impact

09:24 - Job Market Shifts in Tech

13:24 - Adapting to the SaaS and AI Revolution: A Vision for the Future

NVIDIA’s Agent Toolkit, unveiled at GTC 2026, is the company’s bid to become to the agentic era what AWS became to the cloud. The stack is open-source, and its ambition is infrastructural: rather than compete with Anthropic or OpenAI for ownership of the agent, NVIDIA is building the substrate beneath it. When Salesforce ships Agentforce, when SAP deploys Joule, when ServiceNow fields an agentic workflow, each can run on NVIDIA’s runtime, security, and data layer regardless of which model powers the intelligence above. The move arms SaaS incumbents against the model providers, threatening to absorb them, while creating two competing architectures that NVIDIA profits from equally. If model providers win, NVIDIA sells the chips. If incumbents orchestrate on NVIDIA’s stack, NVIDIA owns the hardware beneath and the plumbing between. The Agent Toolkit extends NVIDIA’s lock-in to the agent runtime layer, a new software layer for the agentic era. It is not …

Read more

The consensus that LLMs are a dead end toward AGI is hardening fast. Former Meta Chief AI Scientist Yann LeCun just raised a $1B seed round to replace them, and even OpenAI CEO Sam Altman now admits scaling won’t reach AGI. But two very recent developments suggest the obituary is premature. Axiom Math is producing verified proofs of unsolved mathematical conjectures by co-designing AI with formal verification. It just raised a $200m Series A at a $1.6 billion valuation. A new paper shows that the vision pathway of frozen vision language models (VLMs) can be repurposed as a communication channel between heterogeneous agents, delivering faster and often more accurate coordination than text. Neither result was supposed to be possible with the generative paradigm. Both were achieved by engineering a “layer” above it. The architecture may have a ceiling. But we are nowhere near it yet.

Subscribe now

Read more

TLDR: Anthropic’s new study shows AI is doing a fraction of what it theoretically could. But that gap is closing, and the disruption is already underway through three distinct mechanisms that are currently conflated. I propose a framework to more clearly distinguish them: task substitution (what the Anthropic study measures) is compressing white-collar work from within; middleman compression is quietly dismantling the coordination layer of modern organizations, the roles built to manage human-to-human handoffs, now structurally redundant as agents handle the work beneath them; and process optimization is reducing labor demand in white and blue-collar sectors without touching a single job description. The workpocalypse isn’t a moment. It’s a condition. The jobs data doesn’t properly capture all three mechanisms, but it will eventually catch up. By then, the restructuring will already be substantially complete.

Subscribe now

Read more

OpenAI’s record-setting $110 billion funding round is packed full of revelations about the investors who hedged, the losses that necessitated it, and the competitive shift that threatens it. These looming risks are about more than one company. It’s about the fragility now embedded in the financial architecture of the entire AI economy.

Subscribe now

Read more

In this episode of AGNT, recorded at the New York Stock Exchange, Raphaelle d’Ornano, analyst and co-host of the AGNT podcast, joins theCUBE + NYSE Wired’s Gemma Allen to unpack how the rise of the inference economy is redrawing the competitive map for enterprise software.

She revisits her earlier bearish stance on Figma, noting that after falling nearly 71% from its IPO highs, the company’s Anthropic partnership represents a credible first-mover entry into the agentic orchestration layer.

D’Ornano challenges the traditional vertical-versus-horizontal SaaS defensibility playbook, proposing that the more meaningful distinction is between systems of record evolving into systems of action and tools that must find their place in the emerging orchestration graph.

The conversation also covers Salesforce’s record quarter, where Agentforce, combined with the Informatica acquisition, delivered 169% growth — evidence that the system-of-record-to-system-of-action transition can produce real numbers. She frames Slack as Salesforce’s most defensible agentic asset, citing “proximity to user intent” as a foundational principle of value in the agentic era.

On Nvidia, d’Ornano details the inference economy thesis: as enterprise workloads turn agentic, compute demand for inference will dwarf historical training spend, even as market nervousness lingers around circular financing dependencies tied to OpenAI. She remains bullish on AMD as a strong inference-era alternative, pointing to its Meta partnership and the supply-chain complexity of deploying Blackwell at scale. From Anthropic’s methodical product cadence — Claude Code to vertical plugins to security tooling — to the market’s tendency to confuse narrative with product reality, d’Ornano offers a framework for separating signal from panic.

00:00 - Intro

00:01 - Evolving Markets: Transformations in Media and Industry

03:58 - The SaaS-Pocalypse and the Role of Dev Tools

06:20 - Salesforce’s Journey through the Agentic World

11:01 - Nvidia’s Earnings & The Inference Economy

15:49 - Technological Rivalries: Navigating the GPU and AI Landscapes

23:22 - Conclusion and Looking Forward

This is the third and final installment in a series examining the structural repricing of enterprise software. Part I argued the $285 billion “SaaSpocalypse” was mispriced panic when it came to systems of records that could mutate into orchestrators of agentic workflows. Part II decoded Anthropic’s $380 billion valuation through the lens of orchestration economics. This final piece examines who benefits from orchestration and why the answer is not what the market expects.

Figma’s stock has plummeted ~70% from its $70B IPO peak, but it’s now undervalued as it evolves from a design tool to an essential orchestration node in agentic AI workflows. Via MCP integration and the bidirectional “Code to Canvas” feature, Figma captures semantic density and creates compounding agent-to-tool-to-agent network effects, rewriting software defensibility from seat-based metrics to flow control - separating thriving “throughput nodes” (like Figma) from commoditized tools, while …

Read more