ANTHs 965B Upset|The Safety Bet That Beat OpenAI

The $965 Billion Reversal

For the past two years, the working assumption on every institutional desk was simple. OpenAI leads, everyone else follows. That assumption died on Thursday.

Anthropic, the AI company behind the Claude chatbot, announced a sixty-five billion dollar funding round that pushed its post-money valuation to nine hundred and sixty-five billion dollars. That number matters for one reason above all others: it puts Anthropic ahead of OpenAI, which raised a hundred and twenty-two billion dollars in March but sits at an eight hundred and fifty-two billion dollar valuation. Anthropic raised less money and is worth more. That is not a rounding error. That is a fundamental repricing of who wins the AI arms race.

The round was led by Altimeter Capital, Dragoneer, Greenoaks, and Sequoia Capital. But the lead investors are almost secondary to what the number itself signals. Anthropic's revenue has gone from roughly one billion dollars in annualized recurring revenue in early 2025 to forty-seven billion dollars in May 2026 — a forty-seven-fold increase in just over a year. The company says it expects to reach profitability in the current quarter if it hits its ten-point-nine billion dollar revenue target for the period.

The engine behind that growth is Claude Code, the AI-powered coding assistant that Anthropic launched in May 2025. Within three months of launch, Claude Code crossed five hundred million dollars in annualized revenue. By February 2026, that figure had quintupled to two-point-five billion dollars. According to Menlo Ventures, Anthropic's share of enterprise spending on large language models climbed from twenty-four percent in 2024 to an estimated forty percent today. In the AI coding segment specifically, Anthropic's share rose from forty-two percent in June 2025 to fifty-four percent by year end. Anthropic is not catching up to OpenAI in enterprise. It has already lapped it.

Here is where the contradiction becomes almost uncomfortable to state plainly. Anthropic is the company that refused to remove safeguards that would have allowed Claude to be used for mass domestic surveillance or lethal autonomous weapons systems. It is currently in a legal battle with the Pentagon over exactly that refusal. It is the company whose co-founder stood beside Pope Leo at the release of a forty-three-thousand-word encyclical warning against the dangers of AI. It withheld its latest Mythos model over cybersecurity concerns serious enough to spark a diplomatic incident and a phone call from Vice President JD Vance to AI company heads. Anthropic is, by every conventional measure, the AI company that keeps saying no to the fastest paths to revenue.

And it is now worth nearly a trillion dollars.

The market is not rewarding Anthropic despite its safety posture. The market may be rewarding it because of it. Enterprise buyers — the customers generating forty-seven billion in ARR — are choosing Claude not because it is the most permissive model available but because it is the one their legal and compliance teams can defend in board meetings. The commercial logic of safety, which looked like a liability in 2023, is looking like a moat in 2026.

One data point cuts against the clean narrative. An unnamed company spent five hundred million dollars in a single month on Claude AI because no one had put usage caps on employee licenses. That is not a success story. That is a procurement failure of historic proportions, and it points to a risk embedded in Anthropic's growth numbers: some portion of that forty-seven-billion-dollar ARR is not disciplined enterprise adoption. It is uncapped consumption by organizations that have not yet built the governance structures to manage AI spend. When those caps go in — and they will — the revenue curve faces its first real test.

The Infrastructure Tax

While Anthropic rewrote the AI valuation hierarchy, Dell Technologies printed one of the more striking single-session moves in recent large-cap history. DELL closed up thirty-two-point-seven-six percent on the session — and on the same day, UBS downgraded the stock.

The numbers underneath that move are real. Dell closed fiscal 2026 with total revenue of a hundred and thirteen-point-five-four billion dollars, up nineteen percent year over year. In the fourth quarter alone, AI-optimized server revenue hit eight-point-nine-five billion dollars, a three-hundred-and-forty-two percent increase year over year. Management guided fiscal 2027 revenue to between a hundred and thirty-eight and a hundred and forty-two billion dollars, with AI-optimized server revenue projected to approach fifty billion dollars for the year. Dell entered the current fiscal year with a forty-three billion dollar order backlog and booked thirty-four-point-one billion dollars in new AI orders in the fourth quarter alone.

Goldman Sachs estimates total global spending on AI infrastructure could grow from seven hundred and sixty-five billion dollars in 2026 to one-point-six trillion dollars by 2031. Dell, which held roughly a fifth of the AI server market in 2024, is positioned to absorb a meaningful share of that spending. The stock has already jumped sixty-eight percent in 2026 alone.

Here is what the thirty-two percent single-day gain does not tell you. UBS raised its price target for DELL to two hundred and forty-three dollars from a hundred and sixty-seven — a forty-five percent increase — and simultaneously downgraded the stock to Neutral from Buy. Read that again. The analyst who most aggressively revised earnings expectations upward is the same analyst who stopped recommending the stock. The reason: after a hundred-and-fifty-seven percent run over twelve months, Dell now trades at thirty times trailing earnings. The valuation has caught up to the story before the story is finished.

More structurally, Dell's GAAP gross margin compressed to twenty percent from twenty-four percent as the AI server mix grew. AI servers run on Nvidia and AMD silicon, which is expensive, carries thin margins, and is sold to hyperscalers with enormous negotiating leverage. Dell is winning volume by accepting thinner margins. That trade works in a growth environment. It faces pressure the moment AI infrastructure spending plateaus or hyperscalers bring more server assembly in-house. The S&P 500 closed at a record seven thousand four hundred and forty-four, the Nasdaq crossed twenty-six thousand four hundred, and six of the Magnificent Seven gained between one-point-four and three-point-nine percent. Producer prices jumped one-point-four percent last month, the largest monthly increase in four years. Boston Fed President Susan Collins put a rate hike back on the table. Kevin Warsh, Trump's nominee to succeed Powell, was confirmed by the Senate. The AI rally and the inflation data are now running on parallel tracks.

The same capital that drove DELL up thirty-two percent has not yet fully priced the margin compression that comes with being the dominant AI server provider. That gap between price reaction and margin trajectory is the variable worth watching into the next earnings cycle.

The IPO Gauntlet

The unresolved question from Thursday's valuation repricing is this: if Anthropic is worth nine hundred and sixty-five billion dollars in a private round, what does that number do to OpenAI's IPO math?

OpenAI is preparing to confidentially file its IPO prospectus as soon as Friday, working with Goldman Sachs and Morgan Stanley on what could be one of the largest public market debuts in history. The company is valued at more than eight hundred and fifty billion dollars by private investors. SpaceX, most recently valued at one-point-two-five trillion dollars in February, is also moving toward a public offering with Goldman Sachs in the lead left position. Both companies are targeting 2026 debuts.

The timing is not coincidental. Anthropic's nine-sixty-five valuation has made OpenAI's eight-fifty-two private mark look like it is falling behind. Sam Altman needs the public markets to reset that comparison. The IPO is not just a liquidity event. It is a valuation argument being made in real time, against a competitor that just moved first.

SpaceX received an additional piece of news on Thursday. The company won a contract worth more than four-point-one-six billion dollars to build satellites for President Trump's Golden Dome missile defense shield. That contract arrives the same week Blue Origin's New Glenn rocket exploded on the pad at Cape Canaveral, destroying what Bezos spent more than a billion dollars building at Launch Complex 36 and raising serious questions about the Artemis program's backup launch architecture. New Glenn has flown three times. Its only pad is currently not operational. Artemis III is targeted for mid-2027.

For investors tracking the AI infrastructure capital frame, the week draws a line between compounders and rebuilders. Anthropic at nine-sixty-five, Dell with a fifty-billion-dollar AI server pipeline, SpaceX with a Golden Dome contract and a trillion-dollar IPO in preparation — these are the compounders. The Blue Origin situation is a one-sentence transmission: Bezos's space program is rebuilding, and the Artemis timeline will slip.

The deeper question Thursday leaves open is about the IPO gauntlet itself. Anthropic raised at nine-sixty-five without public market pressure. Once OpenAI files and begins trading, it will face quarterly disclosure requirements, analyst coverage, and public shareholders who can act on the valuation gap between OpenAI and Anthropic in real time. The unstated premise embedded in OpenAI's IPO rationale is that public market investors will assign a higher multiple to OpenAI's consumer reach — ChatGPT's brand recognition and user base — than to Anthropic's enterprise depth. That premise requires that enterprise AI coding market share, the metric where Anthropic now leads, does not become the primary valuation driver for AI companies. If it does, the IPO that was supposed to reset the valuation narrative may accelerate it in the wrong direction.

The company that said no to the Pentagon is worth nearly a trillion dollars. The company that said yes to everything is now racing to Wall Street. The verification benchmark is simple: watch whether Anthropic's IPO valuation, when it comes, clears a trillion dollars or falls back below OpenAI's public price. That spread will tell you whether the market was pricing safety or just pricing speed.