The quiet hum of Anthropic’s IPO filing on June 1 has been drowned out by the roar of a $2 trillion valuation narrative. Forbes columnist Jim Osman correctly identifies this as a critical juncture for the AI boom’s investment logic, but as a macro watcher who has spent years dissecting the structural fragility of DeFi’s liquidity cycles, I see something far more unsettling. The numbers are staggering: a revenue run rate that vaulted from $14 billion in February to $47 billion in May, a private valuation that climbed from $380 billion to $965 billion in the same span. Yet beneath this exponential curve lies a familiar pattern—one that echoes the way crypto markets once priced in future perfection before the music stopped. This is not just an AI story. It is a global liquidity test, and the signals it sends will ripple through every asset class, including the cryptocurrencies I track.
Context: The Architecture of an AI Giant Anthropic’s ascent is a case study in capital-intensive growth. The company secretly submitted its listing application on June 1, and the market is already speculating on an IPO valuation exceeding $2 trillion. To put that in perspective, that would place Anthropic above the combined market caps of most major crypto projects—including Bitcoin and Ethereum at their peaks. The company raised $65 billion in May alone, earmarked largely for expanding computing power. Its infrastructure deals read like a who’s who of the tech-industrial complex: a 5GW agreement with Amazon for new computing capacity, another 5GW of next-generation TPU computing power via Google and Broadcom, and a utilization of SpaceX’s GPU capacity. The commitment to invest over $100 billion in Amazon Web Services over the next decade further cements Anthropic’s reliance on external cloud providers.
This is the core of the story: Anthropic is not a software company in the traditional sense. It is a hardware-dependent, capital-hungry behemoth that must reinvest heavily to maintain its edge in frontier models. As Osman notes, the key question is how future profits will be distributed among model developers, chip manufacturers, cloud service providers, data centers, and software companies. For crypto observers, this distribution dilemma is painfully familiar. I recall auditing the tokenomics of a dozen DeFi lending protocols during the 2020 summer—each promised sustainable yields, yet almost all failed because the value captured by the protocol layer was siphoned away by gas costs, miner extractable value, and liquidity providers with no loyalty. The same fragmentation of value is now playing out in AI, but at a scale that dwarfs anything crypto has seen.
Core: The Macroeconomic Mirror — How AI’s Capital Intensity Echoes DeFi’s Liquidity Fragmentation Let me be direct: the Anthropic IPO is a stress test for the entire macro narrative that “AI is the new liquidity sink.” Since 2023, global capital has flooded into AI startups, mirroring the 2020-2021 crypto bull run where DeFi protocols absorbed billions in search of yield. But much like those early DeFi protocols, Anthropic’s growth is predicated on a narrative of infinite demand. The revenue run rate jump from $14 billion to $47 billion in three months is impressive, but it implies that the market is already pricing in a future where Anthropic captures a dominant share of an exploding market. This is the same trap that ensnared Terraform Labs and countless other crypto projects: the assumption that exponential growth can continue without structural constraints.
From my perspective as a cross-border payment researcher, I see the liquidity flows as a ghost. The $65 billion raised in May is not profit—it is fuel. Anthropic must spend aggressively to acquire the computing power needed to train and deploy models. The 5GW agreements with Amazon and Google represent long-term commitments that lock in massive operating expenses. Meanwhile, the company’s partnership with SpaceX for GPU capacity suggests a scramble for scarce hardware. In the crypto world, we saw the same scramble for Layer2 scaling solutions—projects like Arbitrum and Optimism promised to scale Ethereum, but instead they fragmented liquidity across dozens of chains, reducing the effective depth of any single market. Anthropic’s computing power diversification may be a strength, but it also fragments its own operational focus and dilutes margins.
The distribution of profits is where the real insight lies. Osman rightly points out that investors need to consider how revenue is split among model developers, chip makers, cloud providers, and data centers. In crypto, we learned the hard way that value often accrues to the infrastructure layer, not the application layer. Miners and stakers capture a disproportionate share of revenue in proof-of-work and proof-of-stake networks, leaving little for developers. Similarly, in AI, the chip manufacturers (NVIDIA, AMD, Broadcom) and cloud providers (Amazon, Google, Microsoft) are likely to capture the lion’s share of economic rents. Anthropic is both a model developer and a customer of those providers—it pays for the picks and shovels. The question is whether its model can generate enough cash flow to service these capital costs and still deliver returns to shareholders.
Let me ground this with a data point: over the past year, I have modeled the economic sustainability of several AI tokens on the crypto side—projects like Render Network and Akash Network that aim to provide decentralized computing for AI. The results were sobering. Even with optimistic assumptions about demand, the cost of GPU compute remains so high that margins are razor-thin. Anthropic’s $100 billion commitment to AWS over the next decade is a classic example of vendor lock-in, and it will be a drag on free cash flow for years. Liquidity is a ghost, but the debt is real.
Contrarian: The Decoupling Thesis — Why This IPO Could Shatter the AI Narrative Here is the contrarian angle that most analysts miss: the Anthropic IPO may not be the validation of the AI boom; it could be the catalyst for its first major correction. The market is currently pricing in a $2 trillion valuation based on a revenue run rate of $47 billion annualized, which implies a price-to-sales ratio of over 42x. Even for a high-growth company, that is extreme. For comparison, NVIDIA—the chip giant that has ridden the AI wave to a $3.5 trillion market cap—trades at around 30x sales. Anthropic would need to maintain its current growth trajectory for several years just to justify that multiple. Yet the competitive landscape is intensifying. Google, Meta, and OpenAI are all investing heavily in frontier models, and the commoditization of large language models is already underway. Pricing power erodes when there are multiple high-quality alternatives.
From a crypto perspective, I see a parallel to the 2021 Bitcoin ETF hype. The approval of spot ETFs was supposed to usher in a new era of institutional adoption, and it did—but it also turned Bitcoin into a Wall Street toy, stripping it of its original peer-to-peer cash vision. The price surged, but the network’s utility diminished. In the quiet aftermath, only the resilient remain. Anthropic’s IPO could similarly transform the AI industry from a frontier of innovation into a playground for institutional capital, where the focus shifts from breakthrough research to quarterly earnings. The company’s secret filing on June 1 suggests a desire to avoid scrutiny, but the market will eventually demand transparency.
Furthermore, the capital intensity of AI is a double-edged sword. Anthropic must keep reinvesting to maintain its edge, which means that even if revenue grows, free cash flow may remain negative for years. This is reminiscent of the DeFi protocols that burned through treasuries to pay high yields. DeFi’s glass house shatters under its own weight—the same could happen to AI if the promised profits never materialize. The market is betting on a future where Anthropic becomes a quasi-monopoly in foundational models, but history shows that technology stacks tend to fragment. Just as Ethereum’s Layer2 ecosystem fragmented liquidity, the AI model market may fragment into specialized niches, diluting Anthropic’s dominance.
Takeaway: Positioning for the Cycle The takeaway for macro watchers is clear: the Anthropic IPO is a pivotal moment for global liquidity allocation. If the IPO succeeds at a $2 trillion valuation, it will signal that the market is willing to absorb even more capital into AI, potentially crowding out other asset classes, including crypto. But if it struggles—if the valuation is cut or the IPO is delayed—we could see a rotation back into verifiable stores of value like Bitcoin. Beyond the illusion, the current never truly stops.
My advice to readers is to watch the cash flow data. Anthropic’s ability to convert revenue into free cash flow will be the ultimate test. In crypto, we learned to look past TVL (total value locked) and focus on protocol revenue versus token incentives. The same principle applies here: ignore the revenue run rate hype and examine the operating expenses, capital commitments, and competitive moats. Fragility is the price of unsecured innovation.
As I wrote in my 2024 whitepaper on ETF liquidity flows, the integration of a new asset class into traditional finance always creates a period of mispricing. The Anthropic IPO is that moment for AI. Whether it becomes a new anchor or a sinkhole depends on whether the underlying economics can support the valuation. From my years of observing DeFi’s rise and fall, I have learned that when the flow stops, we see what truly holds. The flow is still strong, but the cracks are visible. Stay skeptical, and keep your eyes on the cash conversion.