Anthropic keeps the AI party swinging, but the atmosphere is bubbly

Can the newly listed AI company really be worth $2tn? Anthropic flew out of the blocks on its first day as a listed company.
Can the newly listed AI company really be worth $2tn? Anyone who remembers the late 1990s has good reason to ask
Roll up, roll up, who wants to buy shares in a company whose technology “could pose catastrophic or existential risks to humanity”, as the prospectus had it? Virtually everybody, as it turned out. Anthropic flew out of the blocks on its first day as a listed company.
From the point of view of the promoters, the result was as good as it gets. The final pricing for the shares at $83 apiece was meaningfully higher than the indicated range of $70 to $78, but still left room for a big initial “pop”. There were positive vibes all round – especially for the battalion of investment banks collecting $550m of underwriting fees.
And, to be fair, it wasn’t nailed-on that the show would go so swimmingly. There was a late risk that investor sentiment towards the AI universe could be jinxed – or Jensened – by the soap opera at Nvidia. More generally, there was a worry that the AI bubble, if that’s what it is, could burst when somebody tried to unload too much stock on the stock market in one go.
Anthropic, though, has set records by raising $107bn. It is an extraordinary sum, especially for a business founded five years ago. At the listing price, Anthropic was valued at $1.5tn. When shares began trading at £111.50, its market capitalisation was up to $2tn. Only 14 months ago, it raised money at a valuation of a “mere” $183bn.
The feat, though, is not wholly inexplicable. Some $26bn of the listing proceeds were virtually in the bag, thanks to the cornerstone placement to AMD, Broadcom, Google, Micron, Microsoft, Nvidia and TSMC. That assembly of AI and AI-adjacent names fed worries about the industry’s circular financing.
But, on the other hand, every fund manager knew he or she would be betting against the crowd by saying no. Momentum can carry a stock a long way, even before the index funds and index-hugging funds are obliged to buy (Nasdaq entry will be quick; the S&P 500 will take at least a year).
Take a step back, though, and ask whether Anthropic – loss-making and with projected revenues in 2026 of $50bn-ish – can possibly be worth $1.5tn-plus.
The short, and unhelpful, answer is that virtually any valuation could be made to sound plausible if you believe, first, that AI is a generational technology that will be adopted at never-seen-before speed and, second, that “frontier” pioneering firms, of which Anthropic is the leader, will capture an outsized slice of the commercial value for years and years.
Consider the assumptions behind that thesis, however. Can Anthropic really improve from $50bn-ish of revenue this year to $600bn in 2030, which is roughly what tech analysts have pencilled in? Given the current rate of growth, perhaps it can. But sustaining the pace, or something like it, for another 16 quarters simply won’t be straightforward or mechanical.
What is the real level of demand from corporate users for frontier models? Nobody doubts that AI adoption is still in the early stage, but does everybody need the finest product from Anthropic’s Claude suite? Already executives grumble that their IT budgets are finite and that nine-month-old versions, or cheaper “open weight” models from rivals, do the job for basic work. Lasting pricing power in AI is simply unclear. “Commoditisation risk”, in the jargon, could be the next year’s valuation debate.
Then there is the investment bill to make it all happen – $900bn of spending commitments were detailed in the prospectus for computing power and infrastructure. The listing proceeds help, as does the visibility and credibility that comes with being on a public market. But debt, not just equity, will be a crucial part of this story.
“Obviously, at some point in the future we’re hoping to be investment grade,” said co-founder Dario Amodei as trading in the shares started. You bet: the price at which Anthropic borrows matters fundamentally.
Away from the party atmosphere, the European Central Bank has its eye on financial stability in the eurozone. This week, it issued one of those fuddy-duddy reminders about how much capital is being bet on the AI revolution, and how little transparency there is around the funding of hyperscalers and other suppliers to the boom.
“Big tech firms are estimated to have total off-balance-sheet commitments of more than $4tn as of the second quarter of 2026,” its report said, warning of “historical parallels with the dotcom and sub-prime bubbles”.
It is not the only one to see warning signs. Anthropic is an astonishingly successful company already and is plainly at the forefront of a significant technological shift. But it has been listed in an environment that looks – to most of us who remember the late 1990s and the subsequent bust – to have bubbly characteristics. AI technology, like the internet, is gamechanging. But it doesn’t mean today’s share prices are right.
- anthropic
- company
- party
- swinging
- atmosphere
- bubbly
- share
- reason





