The private market has been waiting for a major AI company to finally step into the public markets and give investors a real valuation benchmark.
For a while, Anthropic looked like one of the companies that could make it happen. Now, that timeline appears to be moving.
Anthropic has reportedly pushed its planned IPO from October to November, giving investors another quarter of financial results to look at before the roadshow. At first glance, a one-month delay may not sound significant. But when you look at what is happening across the broader IPO market, the timing is worth paying attention to.
Holtec postponed its IPO just one day before pricing. SB Energy reportedly delayed its listing after investors questioned its valuation. Nscale has moved ahead with a filing, but its customer concentration raises questions. And Oura is heading toward the public markets with a deal that is largely a liquidity event for existing shareholders.
Put together, these developments point to a market that is open to new listings, but not at any price.
Anthropic Wants Investors to See More Before They Price It
Anthropic’s reported decision to move its IPO to November is relatively straightforward.
Advisers reportedly wanted investors to have access to third-quarter financial results before the company begins its roadshow. That gives potential investors more recent information to work with when thinking about valuation.
That matters even more because Anthropic is reportedly showing very strong growth.
The company is expected to cross $100 billion in annualized revenue this year, according to reporting cited in the source. That would represent a significant increase from estimates just two months earlier.
Anthropic has also reportedly told shareholders to expect a second consecutive quarter of adjusted operating profit.
Those numbers could make a November roadshow more useful than an October one.
But there is another factor hanging over the offering: how investors value frontier AI companies in the first place.
The market is still trying to figure out what these businesses are ultimately worth when they require enormous amounts of computing power, capital and infrastructure.
A strong revenue number is helpful.
A path toward profitability is helpful.
But investors will also want to understand how sustainable those numbers are and how much capital is required to keep growing.
Anthropic and OpenAI Are Taking Very Different Paths
It is easy to put Anthropic and OpenAI in the same bucket because both are leading AI labs.
Their IPO timelines, however, are telling very different stories.
Anthropic’s reported delay is only about a month. OpenAI’s timeline is much less certain.
OpenAI has said it does not plan to list in 2026, while its reported internal projections show just how capital-intensive its growth plans are.
A company presentation reportedly seen by the Financial Times projects:
- $278 billion in cash burn between 2026 and 2030
- Revenue rising from $36 billion in 2026 to $350 billion in 2030
- Around $856 billion in compute spending over the same period
- Its reported $122 billion funding pool being exhausted by 2028
These are enormous numbers.
And they highlight the central question facing the AI industry:
How much capital will it take to turn AI demand into durable profits?
For private-market investors, that question is particularly important because public comparables are still limited.
A successful Anthropic IPO could provide the market with one of the first major public reference points for a frontier AI company.
That is why the timing matters beyond Anthropic itself.
The IPO Market Is Open, But Investors Are Pushing Back
The broader IPO calendar is sending a similar message.
Holtec Nuclear postponed its Nasdaq IPO the night before it was due to price, citing market conditions.
The company had planned to raise as much as $900 million.
That is not a small adjustment. Pulling a deal at the last moment suggests that pricing expectations and investor demand were not lining up.
SB Energy has faced a similar problem.
The SoftBank-backed data-center and power developer reportedly delayed a listing it had targeted for this month after investors questioned a valuation of more than $50 billion.
Interestingly, Nvidia agreed to purchase another $1.5 billion of SB Energy shares ahead of the IPO at 90% of the eventual IPO price, bringing Nvidia’s total investment to $3 billion.
That gives SB Energy additional private-market validation.
But it also highlights the difference between strategic investment and public-market pricing.
A private investor can have strategic reasons for investing in a company. Public investors ultimately have to decide whether the valuation makes sense based on the opportunity, financials, risks and expected returns.
Nscale Has a Huge Backlog. It Also Has a Huge Concentration Risk
Nscale’s public filing provides another interesting look at the AI infrastructure boom.
The London-based data-center developer reported $103 billion in total contract value.
That sounds enormous.
But there is an important detail:
Microsoft and Anthropic account for 85% of those contracts.
Only around $2.6 billion was active at the end of August, while the company reported a $1.02 billion loss on $140.6 million of revenue in the first half.
This is exactly the kind of situation public investors will have to unpack.
Large contracts can demonstrate strong demand and future visibility.
But investors will also ask:
- How much of the contracted value will actually convert into revenue?
- How long will that take?
- How dependent is the business on a small number of customers?
- How much capital is required to build the infrastructure?
- What happens if one major customer changes its plans?
Nscale is reportedly targeting a valuation of up to $35 billion.
Its roadshow could therefore become another useful test for how public investors are thinking about AI infrastructure.
Oura Shows That Not Every IPO Story Is About Fresh Capital
Then there is Oura.
The consumer health hardware company launched a deal offering 50 million shares at $40 to $44, implying a market value of about $14.1 billion at the top of the range.
But there is a key detail that private-market investors should notice.
Existing shareholders are selling 36.5 million shares.
That means the IPO is largely a liquidity event for existing holders rather than a major capital raise for the company.
Oura expects to generate roughly $533 million in net proceeds, but approximately $526 million is expected to go toward tax obligations related to employee shares vesting at the IPO.
That leaves only around $6.2 million.
So while the headline number is a potential $2.2 billion IPO, the amount of fresh capital actually remaining with the company is very small.
This is an important distinction when looking at IPO headlines.
The size of an IPO does not necessarily tell you how much money the company itself is raising.
For private investors, the bigger question may be whether Oura’s listing establishes a strong public-market valuation and creates liquidity for existing shareholders.
The $278 Billion Question
The OpenAI numbers may be the most eye-catching data point in the entire story.
$278 billion of projected cash burn through 2030.
At the same time, OpenAI reportedly expects revenue to grow from $36 billion to $350 billion.
That is an extraordinary growth trajectory.
But it also means the company is projecting a business that needs to spend heavily to reach that scale.
The reported $856 billion in compute spending over the period puts the economics of frontier AI into perspective.
AI is not simply a software story.
The industry increasingly depends on:
chips + data centers + electricity + networking + capital.
That is why companies such as Nscale, SB Energy and other infrastructure providers are becoming an important part of the AI investment story.
The AI boom is creating an ecosystem around the model providers themselves.
And public markets now have to decide how much of that future growth is already reflected in valuations.
Private-Market Investors Are Watching the Same Thing
For investors holding private shares, this is where the IPO calendar gets particularly interesting.
A public listing can provide something private markets often lack:
a visible, continuously updated market price.
But that price only becomes useful if the IPO trades well.
If major AI and infrastructure companies come public at aggressive valuations and then struggle in the public market, that could put pressure on private-market marks.
On the other hand, if Anthropic, Oura and Nscale perform strongly after listing, it could create more confidence around the broader pipeline.
That is why the quality of the IPO matters just as much as the number of IPOs.
What About the Rest of the Private Market?
The funding market is still active.
Profound raised $180 million in Series D funding at a $1.8 billion valuation, while manufacturing AI company CADDi raised $114 million at a $1.2 billion valuation.
CADDi’s valuation has reportedly more than doubled from $470 million in March 2025.
That tells us something important.
Even while some IPOs are being delayed or repriced, private investors are still willing to put substantial amounts of capital into companies they believe have strong growth potential.
The question is whether those private valuations can eventually translate into public-market valuations.
That bridge is still being tested.
The Bigger Picture: The Market Is Becoming More Selective
There is a temptation to look at IPO delays and conclude that the public market is shutting down.
The picture is more nuanced.
Companies are still filing.
Investors are still funding private businesses.
Oura is moving toward a listing.
Nscale has filed publicly.
Anthropic is reportedly preparing for a roadshow.
But investors appear to be becoming more selective about price, profitability, capital requirements and risk.
That is probably the more important takeaway.
The market is not necessarily saying no to AI.
It is asking a harder question:
At what valuation does the AI opportunity make sense?
That distinction matters.
A company can have exceptional growth and still be too expensive.
A company can have enormous contracts and still carry significant customer concentration risk.
And a company can raise billions in an IPO while receiving very little fresh capital itself.
What We’re Watching Next
A few upcoming events could give private-market investors more clues.
Anthropic’s reported November roadshow
This could become one of the biggest valuation tests for frontier AI. Investors will get a closer look at its financial performance and how the company positions its growth and capital requirements.
Nscale’s roadshow
The key question here is how investors treat its large contracted backlog alongside customer concentration and current losses.
Oura’s trading debut
The company’s performance could provide a useful read on demand for profitable consumer technology companies, although it is not directly comparable with AI infrastructure businesses.
Bamboo Insurance’s pricing
The homeowners insurer is another test of investor appetite following the postponement of Holtec.
The Takeaway for Private Investors
The private market has been waiting for public markets to answer a very simple question:
What are these companies actually worth?
We may finally be getting closer to an answer.
But the current IPO calendar suggests that investors are not going to give every AI or infrastructure company a blank cheque.
Anthropic delaying its IPO by a month is not necessarily a warning sign. In fact, waiting for another quarter of financial results could strengthen the company’s case.
The more interesting signal is what happens when these companies actually meet public investors.
Can high growth justify enormous capital requirements?
Can huge contract backlogs translate into real revenue?
Can private valuations hold up once investors have a liquid alternative?
And perhaps most importantly for private-market investors:
Will the next wave of IPOs create a new benchmark for private AI valuations, or force the market to rethink them?
The next few listings could tell us a lot.