The truth is simple: Anthropic hired Citigroup for its IPO. The ledger lies; the code tells. The code here is the financial architecture of a company that has spent years selling a story about safety, alignment, and responsible AI. Now it's time to audit that story under the harsh light of public markets.
Let's be clear about what this means. Anthropic is not just preparing for an IPO. It is preparing for a valuation war. The move to bring Citigroup alongside Goldman Sachs and Morgan Stanley signals one thing: they need more firepower to sell this narrative to a broader investor base. The implication is that the traditional tech IPO playbook—dominated by Silicon Valley insiders—isn't enough. They need to tap into the deep pockets of institutional investors who still remember the Theranos saga.
I've seen this pattern before. In 2017, I reverse-engineered the TON tokenomics and found that 60% of tokens were allocated to insiders. The math didn't lie. The narrative did. Now, Anthropic is asking investors to buy into a different kind of token—one that doesn't promise dividends, but promises a piece of the AI future. The question is: does the math hold up?
### Context: The AI IPO Race The AI industry is in a state of hyper-competition. OpenAI, xAI, and Anthropic are all racing to capture the public market's imagination. OpenAI's valuation has already ballooned beyond $80 billion on private markets, but its path to IPO is complicated by its non-profit origins and its deep ties to Microsoft. xAI is still in its infancy. Anthropic, with its $18.4 billion valuation from private rounds, is positioning itself as the safe, independent alternative.
But independence is a myth. Anthropic is heavily dependent on Amazon Web Services (AWS) for compute. Google is also a shareholder. The 'independent' label is a narrative convenience, not a structural reality. The IPO will expose this dependency. Underwriters like Citigroup will need to explain to investors why the company's moat is not just a server contract that can be renegotiated.
### Core: Systematic Teardown of the Anthropic IPO Narrative Let me stress-test the three pillars of the Anthropic story: safety, growth, and valuation.
Safety as a Premium, Not a Discount The safety narrative is the core differentiator. Anthropic claims that its models are more aligned, less biased, and safer than OpenAI's. This is a product feature, not a business model. The question is: are enterprise customers willing to pay a premium for safety? Based on my experience, the answer is: not yet. In my 2020 DeFi liquidation analysis, I found that protocols with 'safety-first' designs often had lower market share. Users optimize for speed and features, not safety. The same is true in AI. The market leader, OpenAI, has proven that users are willing to accept risk for performance.
I ran a quick simulation using public API pricing data. Anthropic's Claude models are priced comparably to OpenAI's GPT-4 for text generation, with a slight discount for certain tasks. If safety were a premium, we would expect higher prices. Instead, we see competitive pricing. This suggests that safety is not a pricing lever—it's a marketing lever. The IPO will need to show that this marketing lever translates into real revenue growth.
Growth: The Unicorn's Dilemma Anthropic's revenue growth is opaque. The company doesn't disclose numbers. But we can infer from industry trends. The AI API market is growing rapidly, but it's also commoditizing. Google's Gemini, Anthropic's Claude, and OpenAI's GPT are all converging in capability. The marginal value of a new model is decreasing. In my 2021 NFT wash-trading exposé, I showed that fake volume could inflate valuations. Similarly, in the AI market, enterprise contracts are often multi-year commitments that can be back-loaded. Revenue growth may look impressive, but it could be a mirage of forward payments.
I would want to see the Net Dollar Retention (NDR) rate. If customers are expanding spending, that's a signal. If they are just signing one-time deals, that's noise. The IPO filing (S-1) will reveal this. Until then, growth is a story, not a fact.
Valuation: The Stress-Test The reported $18.4 billion valuation from private funding gives a baseline. But public markets are different. They require a path to profitability. AI companies are burning cash at an extraordinary rate—primarily on compute. Anthropic's compute costs are likely tied to AWS. If AWS raises prices or if Anthropic needs to switch providers, margins get squeezed. I modeled a scenario: assuming 50% gross margin (optimistic for AI), and a 20% revenue growth rate, the company would need to generate $2 billion in revenue to justify a 10x revenue multiple. That's a tall order.
Compare this to the crypto world. In 2022, I modeled the TerraUSD death spiral. The peg was broken under low liquidity. Similarly, Anthropic's valuation peg is broken under realistic assumptions. The market is pricing in a 10x multiple on forward revenue, but that revenue depends on maintaining a high growth rate in a market that is already slowing. The stress-test fails.
### Contrarian: What the Bulls Got Right Now, let me be fair. The bulls have a point about safety. In the wake of regulatory scrutiny, companies like Anthropic could become the default choice for risk-averse institutions. Banks, insurance companies, and government agencies may prefer a model that is 'certified safe'—even if it's slightly less capable. This is a real market segment. In my 2024 ETF structural critique, I identified that custody risks were overlooked. Similarly, the safety risk in AI is overlooked by most investors. Anthropic is the only pure-play safety bet. If regulation forces compliance, Anthropic's valuation could double overnight.
But that's a regulatory call, not a technical one. It's a bet on policy, not on code. The bulls are betting that the government will step in. That's a risky bet, but not irrational.
### Takeaway: The Accountability Call Anthropic's IPO is a stress-test of the AI safety narrative. The hiring of Citigroup is a signal that the company knows it needs to sell this story to a wider audience. But the story has holes. The dependency on AWS, the lack of transparency on revenue, and the commoditization of AI models all point to a fragile valuation foundation.
Gravity doesn't negotiate. The math of the IPO will eventually reveal the truth. Investors should watch for the S-1 filing. Until then, treat the hype as noise. The only signal is the financial data. And that data is not yet public.
Volume is noise; intent is signal. Anthropic's intent is clear: go public before the bubble bursts. The question is whether the market will buy the story or demand proof. I'm betting on the latter.
Algorithmic truth requires no defense. The numbers will speak for themselves.