Beyond Coatue's Fantastic 40: Why Architectural Resilience Trumps Growth Projections in the Agentic Era
Introducing the Agentic Resilience Assessment Framework: In Defense of Google, Apple, Visa and Disney
The next 18 months will determine which trillion-dollar companies survive the agentic revolution. While sophisticated analyses like Coatue's Fantastic 40 provide valuable insights into AI's potential winners by 2030, they may be solving for the wrong variables.
Traditional valuation frameworks that were built for human-operated software are about to become obsolete.
The emergence of agentic AI systems represents a paradigm shift that demands an entirely new analytical framework. These autonomous agents can reason, plan, and execute complex workflows. We are entering what we call "Act II of AI." Act I gave us predictive and generative systems that enhanced human decision-making. Act II introduces truly autonomous agents capable of pursuing complex objectives with minimal oversight.
This discontinuity transcends technological enhancement; It represents an architectural transformation that will reshape how value is created across every sector. Agentic AI resets the foundational calculus of human involvement that underpins most traditional valuation models, operating assumptions, and competitive moats. Tasks that previously depended on people, such as workflow routing, decision logic, and coordination costs, are now increasingly abstracted by software agents. This changes where value accrues, how control surfaces form, and what ‘moat’ means.
After having explored this paradigm in our Agentic Era mini-series, Coatue’s list provides us with a unique opportunity to explore what this means for well-known technology names.
Disclaimer: This independent analysis presents our proprietary research methodology. References to third-party research are based on publicly available information and used for analytical comparison. This content is for informational purposes only and does not constitute investment advice.
Coatue's Vision: AI separates winners from losers
Coatue's Fantastic 40 represents a rigorous attempt to identify AI's ultimate beneficiaries through 2030. The methodology appears to center on enterprise value projections assuming significant addressable market expansion, with some companies projected to reach trillion-dollar valuations based on AI-driven growth scenarios. The results reveal interesting insights about perceived winners and losers.
The most striking observation: the Magnificent Seven's dramatic reshuffling.
Microsoft emerges as the clear victor, a projection we agree with, given CEO Satya Nadella's prescient positioning around agentic AI systems. Meta, Amazon, Tesla, and Nvidia maintain their presence, while Apple and Google conspicuously disappear from the top 40. Coatue likely believes their 5-year growth CAGR will be sub-par.
Within software, Palantir and ServiceNow claim the highest valuations, followed by Salesforce, Intuit, and Shopify. While we don't disagree with some names, a critical question remains: how will these companies avoid unbundling by AI-native applications? And what does a software company even look like in an agentic paradigm? Many questions with only nascent answers to date.
Within foundational models, we understand that Coatue has conviction on three closed-source players: OpenAI, Anthropic, and xAI, with OpenAI positioned as the dominant force. While Coatue also includes Meta and Alibaba, we do not know if that is because of their open-source plays in LLMs through Llama and Qwen, or for other non-related reasons.
Data platforms command attention, with Databricks projected to achieve valuations exceeding $300 billion compared to Snowflake's current $70 billion enterprise value. This reflects Coatue's conviction that Databricks is in a stronger position than Snowflake. However, we must acknowledge that we remain early in understanding how data platforms will evolve in the agentic era, and what “modern” data platforms will resemble.
Having just attended the Snowflake developer conference, I was impressed by the company’s recent moves around unstructured data combined with its already strong enterprise presence. So I would not declare victory just yet! Also, enterprises are still very early in AI deployments, and spending on these data platforms will be a major question in the coming months. Will DBU growth hold up as enterprises start realizing the extent of the spend at stake? We will likely see changes here with implications for durable growth.
The AI infrastructure stack receives overwhelming support from TSMC, Nvidia, and Broadcom. We were surprised by CoreWeave's absence, particularly notable given Coatue's stakes in the company and the stock's extraordinary appreciation since its IPO date. While CoreWeave has positioned itself as AI-optimized infrastructure, I have maintained reservations about the sustainability of its premium valuations in an increasingly competitive cloud infrastructure landscape.
Perhaps most tellingly, traditional stalwarts face exile. Where are Visa, Mastercard, Disney, IBM, and SAP? Companies that weathered multiple technological transitions find no place in this AI-centric future.
This conclusion may prove premature.
Why Current Frameworks Fall Short
The fundamental limitation of growth-based projections becomes apparent when we consider what agentic AI systems represent.
Unlike previous technological shifts, autonomous agents don't simply enhance existing workflows. They have the potential to completely redefine every workflow across every industry. This reshuffling of value creation represents the most profound discontinuity since the internet's emergence.
Consider the implications. Every business process currently managed by software applications could potentially be orchestrated by autonomous agents that understand context, make decisions, and execute actions across multiple systems:
Customer service workflows that today require CRM platforms, ticketing systems, and knowledge bases could be managed by agents that directly interface with customers, access relevant information, and resolve issues autonomously.
Financial planning processes that rely on spreadsheet applications and analysis tools could be handled by agents that continuously monitor market conditions, adjust strategies, and execute transactions.
This workflow redefinition creates a value reshuffling that current valuation methodologies fail to capture. Traditional software companies derive value from human dependence on their interfaces and processes. Agentic systems eliminate this dependence by directly accessing underlying data and capabilities, potentially rendering entire categories of human-facing software obsolete.
We are witnessing an evolution "from reactive generative models to autonomous, goal-directed agents" with capabilities including "multi-agent collaboration, dynamic task decomposition, persistent memory, and orchestrated autonomy." This is not incremental improvement; it represents architectural transformation where value migrates from human-operated tools to autonomous orchestration platforms.
Current valuation methodologies, however sophisticated, cannot adequately assess this workflow disruption risk. When autonomous agents can directly manage customer relationships, automate accounting processes, or orchestrate e-commerce operations, traditional software valuations based on user engagement and subscription revenue become irrelevant.
Value instead accrues to whoever controls the orchestration layer: the infrastructure that enables, coordinates, and constrains these autonomous systems.
Introducing the Agentic Resilience Assessment Framework
To evaluate companies' moats in the agentic era, we propose the Agentic Resilience Assessment Framework (ARAF). This methodology measures architectural positioning as a forward-looking indicator of growth.
ARAF evaluates companies across two primary dimensions: Risk Components and Orchestration Opportunity.
Risk Components (Lower is Better):Companies score poorly when:
Agentic protocols (such as MCP or A2A) can abstract away their interfaces,
Their decision-making becomes automatable, or
Their technical advantages become replicable.
Orchestration Opportunity (Higher is Better):The Orchestration Opportunity dimension captures companies' ability to orchestrate autonomous agents rather than merely participate in AI-enhanced workflows. This involves two critical moats:
Context Moat (controlling the semantic layer agents need to operate effectively)
Workflow Moat (integration into value-creating processes where agents coordinate).
The framework's foundation recognizes that "context-rich agents fundamentally outperform generic alternatives because they can make decisions based on accumulated understanding rather than isolated inputs." This accumulated understanding includes the behavioral data, interaction history, and domain knowledge that agents require.
Companies that control this layer possess sustainable competitive advantages.
The Orchestration Imperative: Where Ultimate Value Lies
The agentic era's defining characteristic is the shift from software tools to autonomous orchestrators. Companies that orchestrate agents through such tasks as determining their objectives, mediating their interactions, and capturing their outputs will command premium positions. Companies that merely participate in agent-enhanced workflows face commodity pricing pressure.
Orchestration is about controlling:
What gets triggered?
In what order?
With what context?
Toward what economic outcome?
This turns into economic power when orchestration owners set the context, route the call stack, and capture margin on downstream tools. I explored this applied to software recently, proposing the Agentic Value Framework as a new rule.
Microsoft's inclusion in Coatue's winners reflects this dynamic. Through its Copilot ecosystem, Microsoft positions itself as the orchestrator of business workflows rather than a provider of productivity tools. The platform coordinates autonomous agents across the entire Microsoft ecosystem, creating network effects that strengthen with usage.
Reassessing the Fantastic 40 Through an Agentic Lens
Applying ARAF to Coatue's selections reveals compelling inclusions, intriguing misalignments, and questions:
Coatue's infrastructure bets appear well-positioned. TSMC, Nvidia, and Broadcom control essential coordination points in the agentic stack, from chip fabrication to AI acceleration. Their inclusion reflects solid architectural reasoning about where value accumulates in autonomous systems. Arm is more of a wildcard, and their licensing shifts could raise some concerns.
Palantir and ServiceNow, positioned as the highest-valued software companies, both score exceptionally well in our framework (both above 1,5/2 per ARAF rankings). Palantir needs to continue to scale commercially and manage valuation risks, while ServiceNow needs to accelerate in AI and further tap into its huge potential.
Databricks' projected valuation exceeding $300 billion reflects its strategic pivot toward agentic infrastructure with Agent Bricks and LakeBase positioning them as orchestrators rather than participants. However, execution remains critical as the company transitions from data platform provider to agentic workflow coordinator.
Google's absence appears particularly questionable given Gemini's architectural advantages for real-world agentic applications. While Google faces legitimate execution challenges, the company's integration across search, maps, email, and enterprise applications creates unparalleled context depth that agents require for effective orchestration opportunity. I had recently underlined, when looking at the defensibility of each frontier AI player, that Gemini 2.5 Pro, integrated across Search, Android (2B+ users), and Workspace, lead in multimodal tasks like image reasoning and coding, per Alibaba’s Qwen3 tests (April 2025). The regulatory uncertainty, however, adds significant complexity to any long-term investment thesis.
Apple's absence appears premature when viewed through an agentic lens. Apple controls the most intimate orchestration layer through personal computing devices that users interact with continuously. The company's privacy-first approach, often viewed as a competitive disadvantage, could prove essential for agentic systems that require user trust for an autonomous operation opportunity. And as noted by Knesia Tse in her recent post: “What if Apple Intelligence isn’t about using third-party large models at all?” She adds: “By allowing its on-device model to developers, Apple is inviting a new generation of apps that don’t run in the cloud…The model lives in the OS.”
Contrary to Coatue's exclusion, we remain bullish on SAP's prospects. In our view, SAP is well-positioned to thrive in an agentic world, with a strong orchestration opportunity driven by its deep integration into enterprise workflows and robust context moat in ERP processes. Certainly, to maintain resilience, SAP should further invest in agent orchestration within its ecosystem.
The complete exclusion of companies like Disney, Visa, and Mastercard from the Fantastic 40 may prove premature. Viewed through agentic lenses, all three benefit from strong positions. Disney controls vast intellectual property libraries that could transform into orchestration platforms for personalized entertainment experiences. Financial services incumbents possess transaction networks and regulatory relationships that create natural coordination points for autonomous financial agents, with physical infrastructure potentially serving as the essential rails for agentic systems.
Financial Adaptability and Durable Growth Moats
While architectural resilience determines long-term survival, financial adaptability enables the transition. Our framework combines ARAF scores with financial adaptability assessment to identify companies with "durable growth moats", which we define as the ability to maintain competitive advantages while adapting to agentic disruption.
Financial adaptability encompasses balance sheet strength, revenue diversification, and capital allocation discipline. Companies with strong cash positions can weather inevitable disruption while building orchestration capabilities. Revenue diversification reduces dependence on potentially vulnerable business lines. Management teams with successful pivot experience possess the institutional knowledge crucial for agentic transformation.
The combination of high ARAF scores and strong financial adaptability creates our highest-conviction investment thesis: companies positioned to orchestrate autonomous systems while possessing the resources to execute that transition. A small dose of Darwinian theory applied to the Tech world.
The New Order: Beyond Tech vs. Non-Tech
The agentic era transcends traditional technology sector boundaries. The relevant distinction is not between tech and non-tech companies, but between orchestrators and participants, or what we term "agentic companies" versus legacy enterprises.
Agentic companies share common characteristics. They control critical coordination points, possess rich contextual data, and integrate deeply into value-creating workflows.
Physical infrastructure, often viewed skeptically in digital transformations, gains renewed importance. Autonomous systems require real-world interfaces, including manufacturing robots, delivery networks, and sensor arrays. Companies controlling these physical coordination points may discover unexpected leverage in orchestrating digital agents.
Implications for Market Structure
The agentic transition will likely reshape public market indices in unprecedented ways, more than the internet in our view. The S&P 500 and Nasdaq's current compositions reflect industrial and digital era value creation. Agentic value creation follows different principles, potentially elevating companies based on orchestration capabilities rather than traditional metrics.
This transformation timeline appears compressed relative to historical precedents. Unlike previous technology shifts that unfolded over decades, agentic capabilities are advancing exponentially.
Consider the progression: OpenAI's function calling capabilities emerged in 2023, multi-agent frameworks like CrewAI achieved mainstream adoption within months, and companies like Salesforce are already deploying autonomous agents in production environments. Recent research shows agent endurance—the ability to maintain coherent goal pursuit over extended periods—has doubled in just seven months. Google DeepMind's demonstrations of multi-agent coordination in complex environments suggest we may see enterprise-grade agentic deployment within the next 4 years rather than the typical 5-10 year technology adoption cycles.
Companies face increasingly binary outcomes: orchestrate or be orchestrated. The speed of this transition creates unique investment opportunities for those who can identify orchestration capabilities before they become obvious to broader markets.
Conclusion: Building Orchestration Positions as the Unlock for Value
Coatue's Fantastic 40 represents sophisticated analysis within existing paradigms, yet the agentic era demands frameworks that anticipate architectural disruption rather than extrapolate current trends. The Agentic Resilience Assessment Framework provides such an approach, emphasizing orchestration opportunity over growth projections while maintaining analytical rigor.
The winners of the agentic era will not necessarily be today's fastest-growing companies or largest AI spenders. They will be organizations that position themselves as essential coordination points in autonomous systems networks, regardless of their current sector classifications. Companies that recognize this shift earliest and position themselves as orchestrators rather than participants will capture disproportionate value as autonomous systems reshape global markets.
This analysis represents the first in our detailed examination of individual companies through the Architectural Resilience Assessment Framework (ARAF). Upgrade to receive our deep dives in full.