Gas is the toll for chaos. And right now, the chaos around Anthropic's IPO is burning capital faster than a mispriced perpetual swap.
Polymarket says there's a 70% chance the AI lab goes public in October. By year-end, that probability jumps to 83%. The target valuation? Two trillion dollars. That's not a typo — it's a narrative priced to perfection, and I'm here to stress-test the assumptions before the market takes its toll.
Let's strip away the adjectives. Anthropic is the developer of Claude, a frontier model that competes neck-and-neck with OpenAI's GPT and Google's Gemini. It's backed by Amazon and Google — two giants who see it as both a strategic hedge and a profit center. Its current annualized revenue sits at roughly $47 billion, according to Reuters. By 2028, the company projects $190–200 billion in revenue. The private market last valued it at around $965 billion in May, just five months ago.
Now, the IPO targets $2 trillion. That's a 107% leap in valuation in less than half a year. Even in the 2021 SPAC frenzy, that kind of velocity was reserved for companies with hockey-stick growth and a clear path to profitability. Anthropic has one, but not the other.
The Core: Deconstructing the Multiples
A $2 trillion market cap against a $47 billion revenue run rate gives a current P/S multiple of 42.6x. That's not just high — it's speculative. Traditional SaaS companies trade at 8–12x forward revenue. High-growth AI names like Nvidia or Palantir command premiums, but they have proven margins and visible earnings. Anthropic's multiple implies the market is discounting future revenue at a rate that assumes near-perfect execution.
Let's look at the forward P/S instead. If the 2028 revenue target of $200 billion is achieved, the forward P/S drops to 10x. That's within the range of high-growth enterprise software, but it requires three years of 60% CAGR. That's aggressive but not impossible — if the AI market expands as predicted. The problem is the underlying cost structure.
Inference costs are the silent killer. Every Claude API call eats into margins. If Anthropic's gross margin is 40–50% — a reasonable estimate for an AI company that rents compute from Amazon and Google — then a 10x forward P/S on revenue translates to a 20–25x P/S on gross profit. That's still rich, but manageable. However, if margins are lower, the valuation becomes a stretch. If they're higher, the company is either lying about its costs or has a secret compute advantage.
Based on my experience during the DeFi Summer of 2020, I learned that the difference between a 40% APY yield strategy and a 20% one often came down to hidden gas costs. Similarly, Anthropic's true cost of compute is the unspoken variable. The company hasn't disclosed its gross margin, and the IPO filing will be the first time we see the numbers. Until then, the $2 trillion valuation is a bet on a black box.
The Prediction Market Trap
Polymarket's 70% and 83% probabilities are cited as evidence of market conviction. But Polymarket is a crypto-native prediction market with limited liquidity and a participant base skewed toward retail degens. It's not a proxy for institutional sentiment. In 2022, I watched Polymarket's odds on the Celsius bankruptcy flip from 10% to 90% in 48 hours — not because of new information, but because a whale dumped a position. The same could happen here.
A more reliable signal would be the pricing of credit default swaps on Anthropic's debt, but there are none. So we're left with a prediction market that has a known bias toward the optimistic. The deep analysis of the original article flagged this as a "Moderate" source reliability issue. I'd call it a red flag.
The Contrarian Angle: What If the Market Is Wrong?
The bull case for $2 trillion is that AI will eat the world, and Anthropic will be one of the two or three dominant providers. The bear case is that the market is pricing a monopoly before the competitive landscape is settled.
Consider the positioning. Anthropic is uniquely tied to Amazon and Google. Amazon is its primary cloud provider and investor. Google is also an investor but competes directly with Gemini. This dual allegiance creates a strategic tension. If Amazon decides to prioritize its own AI models, Anthropic's revenue could be squeezed. If Google pulls its funding, the company loses a key backer. The IPO is partly a way to dilute these dependencies, but it also exposes the company to the whims of public markets.
Liquidity dries up when fear sets in. In a bear market, AI stocks are the first to be sold — they have no earnings, no dividends, and no moat that can't be crossed by a better model. The $2 trillion valuation assumes that fear will never materialize. That's a dangerous assumption.
The Takeaway: A Bet on Conviction, Not Fundamentals
Anthropic's IPO is a liquidity event for the AI narrative, not a reflection of the company's current cash flows. The $2 trillion target is a round number designed to capture headlines, not a precise calculation of discounted future earnings.
For traders, the real opportunity lies in the volatility of the related assets. If the IPO is priced at $1.5 trillion instead of $2 trillion, the reaction will be a sell-the-news event. If it hits $2 trillion, the euphoria could spill over into AI infrastructure tokens like Render or Akash, or into the stocks of Amazon and Google.
But don't confuse a narrative with a trend. Code is law, but bugs are fatal. The bug here is that the market is pricing a future that may not arrive on schedule. When the IPO window opens, ask yourself: are you buying a company, or are you buying a conviction? The answer will determine whether you profit or get liquidated.
Bots don't blink. But humans do. And when the first revenue miss comes, the liquidity will vanish faster than a bad oracle update.