Anthropic shares surge 30% on debut as largest-ever IPO puts AI appetite to the test
Neil Wilson
Investor Content Strategist
The biggest IPO in history - Anthropic started trading on the Nasdaq with a bang. Shares opened at $111.50, about 34% above the IPO price of $83, and handing the Claude owner a $2tn valuation. The $83 IPO price was well above the expressed range and raised $107bn including a private placement led by Nvidia and Google, at a $1.5tn valuation on a fully diluted basis. The offering consisted of 980,000,000 shares of Class A common stock, and there's a greenshoe of 147mn shares. The scale of the customary pop at the open was significant and it underscored exceptional demand for shares and handed Anthropic a $2tn valuation on debut.
The higher-than-expected pricing of the IPO gave the broader tech sector a boost with the Nasdaq-100 rallying ~2% before it reversed a bit. AI stocks really need some tailwinds after choppy action over recent weeks and the drama around Nvidia.
This IPO is the Big One – the key test of whether it's the beginning of the end of the AI bubble. It's bigger than SpaceX and means 2026 has been a year in which the market has absorbed a lot of new issuances. The volume of equity and vast deluge of new debt by companies financing the AI buildout has been an obstacle for markets this year – combined with a surge in real rates and you can see why multiples have compressed so much despite monster earnings growth.
The stock has the hallmarks of a retail favourite and has been massively oversubscribed. It feels like Anthropic has picked up the baton from Nvidia as the point man for the AI trade. Its success or failure will have huge ramifications for the wider AI ecosystem and in particular the timing for OpenAI's own delayed public listing.
The question we ask is simple: is it worth it? Can Anthropic justify at $1.5tn valuation, or $2tn for that matter? There's no obvious peer to compare it to. I don't think anyone really knows but we'll have a stab at it and look at its numbers and look at it in the broader context of what's going on with the stock market, new issues, macro forces and demand.
IPO performance this year has been terrible
First, it's not been a good year to ride an IPO. The largest IPOs on Wall Street this year have not done well. Hat-tip to the Financial Times, which looked at this earlier this month and I've simply updated the numbers.
Eight of the ten listings were below their offer price at the close on 18 November 2026. Their simple average share-price return was −23.1%. The median return was −24.0%. Only Bending Spoons and SpaceX remained above their offer prices.
Company | US ticker | Offer price | Close, 18 Nov | Return from offer |
Bending Spoons | BSP | $29.00 | $42.79 | +47.6% |
SpaceX | SPCX | $135.00 | $143.50 | +6.3% |
Madison Air Solutions | MAIR | $27.00 | $22.86 | −15.3% |
SK Hynix — US ADSs | SKHY | $149.00 | $122.47 | −17.8% |
Forgent Power Solutions | FPS | $27.00 | $21.70 | −19.6% |
Pershing Square USA | PSUS | $50.00 | $35.81 | −28.4% |
Quantinuum | QNT | $60.00 | $31.99 | −46.7% |
Cerebras Systems | CBRS | $185.00 | $94.15 | −49.1% |
INNIO | INIO | $27.00 | $13.59 | −49.7% |
Fervo Energy | FRVO | $27.00 | $11.40 | −57.8% |
Anthropic's IPO comes at a volatile moment
A lot of stock market performance is about luck. Anthropic has pressed ahead with its IPO despite the AI trade blowing up in many areas. No one would have predicted that Nvidia's CEO would be forced to resign over an affair with an alleged Chinese spy, but the trouble in the tech sector – particularly areas linked to AI such as semiconductors and memory – has been brewing for a while now.
Chip stocks peaked earlier in the year but selling picked up through October's Q3 earnings season amid fears investment in AI was not translating into revenue. OpenAI's revenue shortfall and big negative free cash reads from the likes of Alphabet spooked the horses. November's US midterm election sparked another wave of pessimism over worries the Democrats would effectively shut down data centre building. This is a tough setup for the IPO.
Global fund managers have dumped stocks at the fastest pace in 20 years, while retail have been willing buyers of the dip. The S&P 500 just closed below its 200-day moving average for the first time since April. Many believe this is it – the AI bubble bursting.
The VIX – Wall Street's fear gauge – has been pretty quiet all year but jumped to 28 on the 5 Nov capitulation event and was last around 22. But S&P 500dealer gamma is at one of the most negative levels ever recorded, which could amplify volatility heading into the year end.
Small float, lockups to consider
Just 6% of shares are in the free float, which means the release of stock from insiders will be important. Here's a rough schedule of what is in the pipeline.
Day 1 (November 19, 2026): Insiders – employees and directors - can sell up to 10% of their locked-up shares as trading begins.
Starting December 15, 2026: Small investors holding less than 0.5% company can begin unlocking shares in incremental steps, with up to 303 million shares potentially unlocking in December.
From Spring 2027: Releases continue gradually through 2027—with roughly 22% of locked shares estimated to be freed by early May 2027 and around 39% near the 180-day mark.
Founders: The seven co-founders are restricted from selling most of their shares for 365 days.
Anthropic has not announced an exact date for its full-year 2026 financial results, which will trigger a fresh tranche of selling as larger investors holding more than 0.5% of the company are locked up until the second full trading day after results for the quarter ending 30 September 2027.
Index and ETF inclusion
Although Anthropic's listing makes it one of the world's most valuable public companies, it is unlikely to carry a significant weighting in major indices when it eventually joins, as most are weighted by float-adjusted market capitalisation.
It may also take a while for some index funds to include Anthropic, with the Nasdaq-100 and Russell 3000 set to take longer than they did for SpaceX.
Nasdaq would normally have a 15-day wait to add Anthropic stock to its indices like the Nasdaq-100 but has said it won't be added until 18 Dec due to the timing of its periodic review of its indices. Nasdaq doesn't use a float-adjusted market cap so ANT will be a larger weight in funds that track Nasdaq indices such as the Invesco EQQQ Nasdaq-100 ETF.
The Morningstar U.S. Total Market Index is set to add Anthropic after a five-day waiting period following its IPO.Funds that track this index such as the Vanguard Morningstar Total Stock Market Index Fund (VI) will probably have the stock in the basket by the end of the month.
MSCI can add a large cap stock after ten days, so funds like the iShares MSCI USA ETF (MLRG) should have Anthropic included by the start of December.
FTSE Russell said it will take longer to add it to its indices like the Russell 3000 because of the timing of Thanksgiving. So, it won't be included in funds like the iShares Russell 3000 ETF (IWV) for a little longer.
Some financials
At $83 a share, Anthropic’s headline valuation is approximately $1.5tn on a fully diluted basis.
Using the latest reported forecasts, that means roughly 29x2026 revenue, 11x 2027 revenue and 5.5x 2028 revenue.
Revenue growth has been pretty staggering. For the first six months of 2026, Anthropic's revenue grew 1305% year over year to $16.3 billion. It then exceeded that total in Q3 alone. In Q2 it became the first genuine frontier AI lab to report profitable quarter, showing investors that it can outrun its compute bill. The company posted an adjusted operating margin of 19% in the second quarter.
Revenue has accelerated sharply, but note the $81bn figure is an annualised run rate.
Period | Revenue | Basis |
2024 | $386m | Reported full year |
2025 | $4.59bn | Reported full year |
Q1 2026 | $4.73bn | Reported quarter |
Q2 2026 | $11.56bn | Reported quarter |
Q3 2026 | $16.96bn | Reported quarter |
First nine months of 2026 | $33.24bn | Reported year to date |
October 2026 | Approximately $81bn | Annualised run rate |
Q3 revenue increased approximately 47% quarter on quarter. The October run rate annualises recent usage and subscription fees; it is neither booked annual revenue nor contracted future revenue.
For the first nine months, usage-based revenue contributed $26.3bn, or 79%, while subscriptions generated $6.9bn, or 21%. That makes continued customer usage central to the business model. Operating performance is improving, although compute costs remain substantial.
In Q3, Anthropic reported:
$567m operating loss, approximately a negative 3% margin.
$2.46bn adjusted operating profit, approximately a 15% margin, excluding stock compensation and non-cash charitable contributions.
$9.86bn compute and infrastructure costs, up from $5.56bn in Q2. Adjusted operating margin consequently fell from approximately 19% to 15%.
The $104bn net loss for the first nine months included approximately $94bn of financing-related accounting adjustments, so it should not be treated as cash burn. Operating cash flow was positive $6.6bn over that period.
The spending commitments are enormous—and extend much further than the revenue forecasts.
Aggregate purchase commitments and leases total approximately $900.9bn, substantially for compute and infrastructure.
Payment horizon | Commitments |
Twelve months following 30 September 2026 | $70.7bn |
Cumulative through end-2030, including the above | $451.6bn |
Remaining balance, substantially during 2031–2045 | Approximately $449.3bn |
Named agreements include Google: at least $121.3bn; Amazon: at least $110bn; Microsoft: $47.2bn–$49.2bn; and Broadcom equipment leases: approximately $161.2bn excluding interest. These sit within the broader commitment disclosures.
Customer commitments run in the opposite direction: remaining performance obligations were $13.8bn, with approximately 75% expected to become revenue within twelve months. The horizons differ, and usage revenue can exceed contracted minima, but this illustrates the underlying exposure: long-term spending obligations supported largely by usage-dependent sales.
The IPO provides a substantial funding buffer.
Anthropic held $131.8bn of cash and short-term investments at September-end. The financing adds:
Source | Gross proceeds |
IPO shares sold by Anthropic | $79.38bn |
Concurrent strategic investor placement | $26.00bn |
Total fresh capital for Anthropic | $105.38bn |
The valuation becomes much easier to justify if the 2027 and 2028 revenue forecasts materialise. At 11x forward earnings (2027) it would be trading well below the ~19x for the Nasdaq-100.
Reuters reported the following bank estimates during the IPO roadshow. These are reported analyst projections, rather than completed results or firm customer commitments.
Revenue basis | Revenue | $1.5tn equity value / revenue |
Latest twelve months to September 2026 | Approximately $35.3bn | 42.5× |
October annualised run rate | Approximately $81bn | 18.5× |
2026 forecast | $52bn–$53bn | 28.3–28.8× |
2027 forecast | $135bn–$137bn | 10.9–11.1× |
2028 forecast | $270bn–$280bn | 5.4–5.6× |
At the forecast midpoints, revenue must grow approximately 159% in 2027 and 102% in 2028—more than fivefold between 2026 and 2028.
There is also a margin hurdle. On an illustrative $275bn of 2028 revenue:
Assumed net margin | Implied earnings | $1.5tn valuation / earnings |
10% | $27.5bn | 54.5× |
20% | $55.0bn | 27.3× |
30% | $82.5bn | 18.2× |
At $83, investors are paying for both exceptional revenue growth and durable profitability. The main vulnerability is that customer demand and pricing can change faster than Anthropic can reduce its infrastructure payments.
The key question is where all the revenue is going to come from. The multiple doesn't look too hard, the test is the earnings momentum.
Mis-anthropic: What about the killer robots?
How do you value a business that might wipe out humanity? Anthropic warned in its SEC filing that it could kills us all.
It noted for instance that AI models could go rogue:
"Our models have in the past behaved, and may in the future behave, in unintended, misaligned, or autonomous ways, including by taking harmful actions that no user or third party directed, which could adversely affect our reputation and brand, business, financial condition, and results of operations."
And AI could kills us all:
"Concerns have been raised both within our organization and by other researchers, regulators, policymakers, and industry participants that sufficiently advanced AI systems could, if not developed, deployed, and operated with appropriate safeguards, oversight, and security measures, pose catastrophic or existential risks to humanity, particularly if such systems were misused, replicated, weaponized, or allowed to operate with a high degree of autonomy."
So far so typical kitchen sinking of risks ahead of the listing. But we do know even Anthropic thinks it should probably slow down and has called for pacing the frontier.
While warnings about AI's potential threat to humanity are nothing new, in September Anthropic CEO Dario Amodei published an essay calling for a coordinated slowdown and safeguards to prevent bad stuff happening.
But some are warning that Amodei's public concern for safety is at odds with what the company is actually doing. On the eve of the IPO a group of AI researchers, including founders of Hugging Face and xAI, pointedly said Anthropic is not safe.
They said: "Anthropic warns that recursively self-improving AI could concentrate power and escape human control, yet races to build it first. The company uses safety branding to court investors, employees, and the public—even as it has quietly weakened its safety pledges and lobbied for self-serving regulation. Anthropic is not a safe AI company, and we must reject its reckless attempt to amass power in the name of the greater good."
Safety concerns are real, but is it overdone.? I noted in September that Anthropic's warning about safety may have actually been a signal that spending is unsustainable and that there is a risk that it misses expectations; the huge amount of debt and circular economics of the investments across the sector means the failure of a single startup could create a cascade effect through financial markets.
Amodei has previously acknowledged that missing growth estimates for just a year could threaten the firm’s survival. Even a slip from 10x revenue growth to 5x could be catastrophic. So, what if the requirement to pace frontier AI development is bred by a worry that it could miss these estimates?
Slowing allows revenues to catch up – frontier development is pure cost, whilst revenues are being generated from the commercial applications of existing models built in the last two years...and these may be slowing, or at least revenue channels may be closing off. Ramp warned that their latest data shows “businesses are hitting their limit on AI spend”. They note that adoption is slowing and therefore more of their growth will have to come from existing businesses already spending on AI, “particularly the advanced spenders, and those businesses are increasingly spending on open source”. This implies a narrower base upon which the likes of Anthropic can grow, raising doubts about scalability.
Adoption is key to revenue growth
Which brings us to adoption, and the killer question: where does the revenue come from? Ramp's AI Index shows continued adoption growth. The total share of businesses using AI in some way rose 1ppt from September to 56.8%, with Anthropic leading the way. In the week ending November 1, 2026, Anthropic captured the largest share of AI token spending among businesses on Ramp, at 49.3%. Total token spending rose 10.5% compared with the same week last month.
Anthropic's S-1 filing shows it has more than 11,900 enterprise customers with $100,000 or more in trailing 12-month revenue, a significant gain from around 950 a year before.The baseline is still offering a decent risk-reward profile. Of publicly traded Fortune 500 companies spending at least $100k a year with the firm, their outlay represented just 0.1% of their total expenses in Q3, according to Pitchbook.
The customer mix is broadening, albeit some very large tech giants dominate the revenue. The top five customers made up~25% of revenue in the first nine months of 2026, though broader enterprise adoption means its largest single customer has fallen from 43% of Anthropic'srevenue in 2023 to 11% in the first half of 2026.
Finally, missed out on the Anthropic IPO in the primary? At least you're not jailed crypto scammer mogul Sam Bankman-Fried. He would have made more than $40bn from his early investment in Anthropic. But if anything, this year's IPO market has taught investors that it might be wise to save some dry powder to deploy when shares hit the secondary market.
This content is marketing material.
None of the information provided on this website constitutes an offer, solicitation, or endorsement to buy or sell any financial instrument, nor is it financial, investment, or trading advice. Saxo Capital Market Ltd. (SCML) provides execution-only services, with all trades and investments based on self-directed decisions. Analysis, research, and educational content is for informational purposes only and should not be considered advice or a recommendation.
SCML content may reflect the personal views of the author, which are subject to change without notice. Mentions of specific financial products are for illustrative purposes only and may serve to clarify financial literacy topics. Content classified as investment research is marketing material and does not meet legal requirements for independent research.
SCML partners with companies that provide compensation for promotional activities conducted on its platform. Some partners also pay retrocessions contingent on clients investing in products from those partners.
While SCML receives compensation from these partnerships, all educational and research content remains focused on providing information to clients.
Before making any investment decisions, you should assess your own financial situation, needs, and objectives, and consider seeking independent professional advice. SCML does not guarantee the accuracy or completeness of any information provided and assumes no liability for any errors, omissions, losses, or damages resulting from the use of this information.
Please refer to our full disclaimer and notification on non-independent investment research for more details.