Why is Anthropic's IPO six times oversubscribed despite the AI selloff?

Anthropic is approaching its Nasdaq debut with investor demand far stronger than the broader AI market would suggest.
The Claude developer's IPO was around six times oversubscribed by the end of last week, according to people familiar with the offering.
That demand is notable because AI stocks have been falling sharply and investors are increasingly questioning high valuations, infrastructure spending and the durability of the current investment cycle.
Anthropic is therefore becoming a useful test of something larger than one IPO.
Investors are deciding whether frontier AI companies still deserve premium valuations even as enthusiasm around the wider sector cools.
Anthropic's IPO demand has accelerated quickly
The offering was already around twice subscribed on the first day of the roadshow.
By Thursday, investor orders had reached roughly six times the number of shares available.
That is strong demand for an offering expected to raise more than $60 billion.
The roadshow is scheduled to continue through Wednesday, when the company is expected to price the IPO. Trading is expected to begin on Nasdaq later in the week under the ticker ANT.
Strong oversubscription does not guarantee that the stock will rise after listing.
Institutional investors can submit larger orders than they expect to receive, particularly when they believe an IPO will be heavily allocated.
But the level of interest still shows that Anthropic has not been dragged down by the broader AI selloff in the same way as many public technology companies.
Investors are buying scarcity as much as growth
Part of the demand comes from Anthropic's position in the AI market.
There are very few opportunities for public market investors to own a leading frontier model developer directly.
Microsoft offers exposure to OpenAI through partnerships.
Alphabet has Gemini.
Amazon and Google have invested in Anthropic.
But those companies are diversified technology businesses.
Anthropic would offer something more direct.
Investors would be buying a company whose value depends primarily on the growth of Claude and enterprise AI.
That scarcity can support a premium valuation.
It also means the market has few obvious public companies to use as direct comparisons.
Revenue forecasts explain some of the enthusiasm
Banks working on the IPO are reportedly forecasting Anthropic revenue of roughly $52 billion to $53 billion in 2026.
That could rise to around $135 billion to $137 billion in 2027 and potentially $270 billion to $280 billion in 2028.
Those forecasts are extraordinary.
If Anthropic comes close to them, the company's current valuation becomes easier to defend.
The problem is that investors do not agree on whether those numbers are realistic.
Some buy side estimates are significantly lower.
One investor cited by Reuters estimated 2028 revenue roughly 30% below the projections used by banks marketing the IPO.
That gap matters.
At Anthropic's valuation, relatively small changes in long term growth assumptions can produce enormous differences in what the company is worth.
Claude's coding strength may not be enough
Anthropic has built a particularly strong position in coding.
Claude is widely used by software developers and enterprise customers.
That has helped the company gain market share from competitors.
But coding is only one part of the potential AI market.
Investors want to know whether Claude can become equally important across research, finance, legal work, customer service and other knowledge worker applications.
If Anthropic remains unusually strong in coding but fails to expand into broader enterprise workflows, future revenue may fall short of the most aggressive forecasts.
This is one of the reasons some institutions remain cautious despite strong demand for the IPO.
Competition is getting harder
Anthropic also needs to defend its technical lead.
Google, OpenAI and other model developers continue releasing new systems.
AI performance gaps can narrow quickly.
A model that leads one benchmark today may lose that position after a competitor releases an update.
That makes frontier AI very different from a traditional software business where customers can remain locked into the same product for many years.
Anthropic needs to maintain model quality while also building deeper enterprise relationships.
The second part may eventually matter more.
A durable business cannot depend on winning every model benchmark.
It needs customers that continue paying even when competitors improve.
The AI selloff makes this IPO more interesting
Anthropic is going public at an unusually difficult time.
Nvidia has lost roughly one third of its value from its October peak.
Other AI infrastructure companies have also fallen as investors question debt, capital expenditure and future returns.
Yet Anthropic's order book remains heavily oversubscribed.
That suggests investors are separating frontier AI software from parts of the infrastructure trade.
They may be less willing to pay extreme valuations for companies funding data centers with debt while still willing to pay a premium for businesses they believe can capture recurring AI revenue.
Whether that distinction survives after Anthropic begins trading is another question.
What should investors watch at pricing?
The first signal will be where Anthropic prices within its range.
Pricing at the upper end would confirm that institutional demand remained strong through the roadshow.
A price above the initial range would be even more aggressive.
The second signal will come from allocation.
Heavy oversubscription means many investors may receive fewer shares than requested.
That can create additional buying after the stock begins trading.
But IPO oversubscription should not be confused with guaranteed upside.
Anthropic still needs to prove that its growth forecasts are achievable.
The company is entering public markets at a valuation that assumes it will become one of the world's largest technology businesses.
My view is that the six times oversubscription tells us investors still want frontier AI exposure.
It does not tell us they have solved the valuation problem.
The real test begins after the shares start trading.
Author

Mindy
MEXC
Mindy is a Digital Asset Market Analyst at MEXC, covering cryptocurrency market trends, macroeconomic developments, derivatives positioning, and digital asset market structure.

















