The market is pricing Anthropic's IPO at the edge of a trillion dollars. The public is pricing AI at the edge of a pitchfork. One of these numbers is wrong, and it is not the public's.
Over the past 12 months, the percentage of Americans expressing opposition to AI data centers has surged from 42% to 75%. That is a 33-point swing in a single calendar year. For context, that is the same velocity of sentiment shift we saw in the collapse of Terra's UST peg in May 2022. It took nine days for that sentiment to destroy $40 billion in market cap. The market is about to test whether a similar sentiment shift can destroy a narrative, not just a token.
Anthropic is preparing for a blockbuster IPO with a valuation that reportedly approaches the trillion-dollar mark. Its annualized revenue run rate sits above $65 billion. The numbers are spectacular. The timing is not. Because while the company has been building out its compute infrastructure and refining its Constitutional AI approach, a separate infrastructure has been building in parallel: a political and social resistance movement with its own energy, its own funding, and its own momentum.
We do not predict the storm; we short the rain.
The Data That Changes the Risk Equation
Let me be precise about the data points, because this is where the market's pricing mechanism begins to break down.
Gallup polling shows 71% of American adults expect AI to reduce jobs. The Heatmap Pro survey shows 75% opposition to new data centers in their communities. Pennsylvania and New York governors have issued executive orders targeting data center development. Investors are asking pointed questions about compute infrastructure slowdowns.
Here is what the market is not connecting: these are not separate data points. They are one continuous signal that flows from public sentiment into political action, from political action into permitting delays, from permitting delays into compute scarcity, and from compute scarcity directly into revenue projections.
The signal chain is direct. Compute capacity is the only real constraint on AI revenue growth. The article is clear on this: compute and revenue are directly correlated. When you restrict compute, you restrict revenue. When you restrict revenue, you restrict the multiple the market is willing to pay. This is not complicated mathematics. This is a linear equation.
The market is treating public sentiment as a PR problem. It is not. It is a supply chain problem wearing a PR costume.
The Balance Sheet Blind Spot
I spent three months auditing 0x Protocol v2 contracts in 2018. I learned that what the balance sheet does not show is often more important than what it does. The same principle applies here.
Anthropic's balance sheet, if it were a traditional company, would show revenue growth, operating costs, and capital expenditure. But it would not show the cost of community opposition. It would not show the price of a governor's executive order. It would not show the opportunity cost of a two-year permitting delay on a 500-megawatt data center.
This is the hidden tax that no spreadsheet captures: the sentiment tax.
The sentiment tax operates through three distinct mechanisms. The first is the permitting tax, where each new data center requires more environmental review, more community consultation, more energy efficiency guarantees. The second is the energy tax, where AI companies must secure power purchase agreements in an increasingly competitive and politically charged energy market. The third is the retention tax, where the best AI talent—people who joined Anthropic because they believed in the safety mission—begin to question whether the company's growth trajectory is compatible with their values.
Each of these taxes compounds. They are not one-time costs. They are recurring, escalating, and fundamentally unpredictable.
The Competitive Asymmetry
Now let me talk about how this risk is distributed across the competitive landscape, because this is where the real trade emerges.
Anthropic is a pure-play AI company. It does not own its own cloud infrastructure. It relies on third-party providers for compute. This means its entire supply chain is exposed to the sentiment tax. Every data center that gets delayed is a direct constraint on its ability to train models, serve customers, and generate revenue.
Compare this to the competitive set. Microsoft has Azure. Google has its own TPUs and data centers. Meta has an open-source strategy that allows models to be deployed locally, partially decoupling from the data center debate. Even OpenAI has a deep strategic partnership with Microsoft that provides a degree of supply chain certainty.
Anthropic has none of these buffers. It is the most exposed player in the AI ecosystem to the exact risk that the market is least equipped to price.
This is a structural asymmetry. It is not a matter of opinion. It is a matter of balance sheet composition. When I built cross-exchange statistical arbitrage strategies in 2025, I looked for persistent pricing discrepancies driven by fragmented regulatory reporting. This is the same pattern: a structural inefficiency in how the market prices risk across comparable assets.
The market is pricing Anthropic and OpenAI as comparable assets. They are not. One has a supply chain moat. The other does not.
The Narrative Trap
There is a deeper problem here, and it is one that I recognize from the DeFi summer of 2020.
Back then, every protocol was promising "sustainable yield." The narrative was that DeFi was the future of finance. The reality was that most yields were subsidized by token inflation and would collapse the moment the incentives stopped. The protocols that survived were the ones that built real revenue, not real narratives.
Anthropic's "safety" positioning is a narrative, and a powerful one. But the market is about to discover that the safety narrative has a dark side. When public sentiment turns against AI, the safety narrative becomes a liability. It says: "We are the ones who know AI is dangerous." The public hears: "AI is dangerous."
The safety narrative validates the fear. It does not mitigate it.
This is the narrative trap. Anthropic has positioned itself as the responsible AI company. But responsibility, in the public's view, is not the same as safety. Responsibility means not building dangerous things in the first place. Safety means building dangerous things carefully. The public is increasingly questioning the first premise, not just the second.
The Valuation Question
Let me put some numbers on this, because the market loves numbers and I love markets.
Anthropic's reported annualized revenue run rate is approximately $65 billion. The reported IPO valuation is approaching $1 trillion. That implies a price-to-sales multiple of roughly 15 times.
In a vacuum, a 15 times P/S ratio for a company growing at 200% year over year is not unreasonable. The market has paid far more for far less. But the market has never priced a company whose core infrastructure expansion is directly opposed by 75% of the public in the country where it operates.
This is not a standard growth multiple. This is a growth multiple with an embedded short option on public sentiment. And the market does not know how to price that option because it has never been exercised before.
Let me be direct: the risk is not that the market is wrong about AI's potential. The risk is that the market is wrong about the timeline.
AI is the most significant technological shift since the internet. I am not arguing otherwise. The question is whether the market is pricing the transition to an AI-driven economy at the exact moment when the public is demanding a slowdown. That is a timing mismatch. And timing is everything in this business.
The Contrarian Angle
Now let me give you the other side of this trade, because a good trader always knows the other side.
The contrarian view is that public sentiment is a lagging indicator. It reflects the fear of a technology that is poorly understood. As AI becomes more integrated into daily life, as people see the benefits, the fear will subside. The data centers will get built. The jobs will shift. The public will adapt. This has happened with every major technology, from the railroad to the internet.
There is a version of this argument that I find compelling. The anti-AI sentiment is real, but it is also concentrated in specific communities—the ones that are being asked to host the data centers. The "not in my backyard" phenomenon is powerful, but it is also localized. The 75% opposition number is a national average. The actual opposition in communities that understand the economic benefits of a data center might be lower.
And there is a second part of this argument: the sentiment creates a barrier to entry. If it becomes harder to build data centers, then the companies that already have compute capacity become more valuable. The scarcity premium increases. Anthropic's existing compute agreements become more valuable, not less.
This is a real argument. It is the same logic that made NVIDIA the most valuable company in the world. Scarcity creates value. I respect this argument. I would even trade it.
But I would not trade it at a $1 trillion valuation.
The Structural Reality
Here is what the market is missing. The anti-AI sentiment is not a temporary phenomenon. It is a structural shift in how the public perceives technology companies.
The past decade has eroded public trust in big tech. Data privacy scandals, algorithmic manipulation, social media addiction, and the general sense that these companies have too much power—all of this has created a foundation of skepticism. AI is landing on top of that foundation. The public is not reacting to AI in isolation. They are reacting to AI as the latest manifestation of a technology industry they no longer trust.
This is why the sentiment data is so sharp. It is not a rational assessment of AI's risks and benefits. It is a cumulative judgment on an industry that has lost the benefit of the doubt. And that judgment is unlikely to change quickly.
The data supports this. The one-year sentiment swing from 42% to 75% is not a response to a single event. It is a response to a pattern. It is a response to the sense that AI is being built without public consent, that it is being built for the benefit of a few, and that its costs will be borne by everyone else.
This is not a technical problem. It is a political problem. And political problems do not have technical solutions.
The Trade
So let me give you the trade, because that is what you are here for.
The trade is not short AI. The trade is not long AI. The trade is being aware of the asymmetry between what the market is pricing and what the public is signaling.
If you are an investor considering participating in the Anthropic IPO, the question is not whether AI will change the world. It will. The question is whether the market is pricing the path to that change correctly. The path includes data center delays, permitting battles, community opposition, and regulatory uncertainty. The path includes a sentiment tax that no spreadsheet captures.
I have been through this before. I watched DeFi protocols with strong narratives and weak fundamentals collapse when the incentives ended. I watched NFT collections with high floor prices and thin liquidity vanish when the whales exited. I watched the 2022 bear market punish every company that had priced in permanent growth without accounting for structural risk.
Leverage doesn't care about feelings. But markets care about sentiment. And sentiment, right now, is the most mispriced asset in the AI trade.
The smart play is not to bet against AI. The smart play is to respect the gap between the market's pricing of AI and the public's acceptance of AI. That gap will close. It always does. The only question is whether it closes through price appreciation, as the public warms to AI, or through price correction, as the market wakes up to the sentiment tax.
Based on the data, I know which side of that gap I am positioning on.
The Forward Question
When the Anthropic S-1 is filed, read the risk factors section carefully. You will see the standard language about competition, regulatory changes, and technological disruption. The question is whether you will see language about public sentiment. Because if the company does not list it as a risk factor, the market will not price it. And if the market does not price it, there is an arbitrage opportunity for the investors who see it.
I have spent my career finding inefficiencies in the gap between what the market prices and what the code says. The code of this market is written in public opinion surveys, executive orders, and community resistance. The market is not reading that code.
I am.
We do not predict the storm; we short the rain.