Which private AI companies are closest to an IPO?

Why does the private AI IPO race suddenly feel real?

The private AI IPO race feels real because Anthropic has published its S-1, while Cerebras has already shown that public investors will buy a huge pure-play AI offering.

Cerebras is useful as the starting point because it has already left this ranking. The AI-chip company began trading on Nasdaq under the ticker CBRS after pricing its IPO at $185 a share and ultimately raising about $6.4 billion including the underwriters' option. Nasdaq said the deal was heavily oversubscribed. A company that had spent years being discussed as a difficult AI IPO actually got the offering done. An AI IPO that actually got done is a real shift; we cover how much moved in a year in our report on AI infrastructure.

The frontier-model companies followed. Anthropic and OpenAI both submitted confidential draft S-1s earlier in 2026. On 26 October 2026, Anthropic went further and filed its S-1 publicly, with audited financials anyone can read.

The financial scale has also changed. Anthropic's S-1 shows $16.3 billion of revenue in the first half of 2026, against $1.2 billion a year earlier. Databricks says its annualized revenue run rate exceeds $7 billion. Anduril expects about $4.3 billion of revenue this year, roughly twice its 2025 level. ElevenLabs crossed $500 million of ARR, while Glean reached $300 million only 15 months after passing $100 million.

A few years ago, most private AI companies were simply too young or too small for a serious IPO discussion. Today, several already have the revenue of large public software companies. The harder question is which management teams actually want to list.

What does “closest to an IPO” really mean for a private AI company?

The private AI company closest to an IPO today is the one furthest through the actual listing process, rather than the company with the largest valuation.

That distinction changes the ranking immediately. A public S-1 is the strongest evidence short of pricing, because investors can already read audited financial statements, risk factors and the governance rules. A confidential draft S-1 is weaker but still serious. Hiring a finance executive or talking about being “IPO ready” is much weaker.

Business maturity still counts. A company needs enough revenue, a credible financial story, internal controls and management able to operate under public scrutiny. Those things tell us whether an IPO is possible. They do not tell us whether management intends to do one soon.

Databricks illustrates the difference perfectly. The company has multibillion-dollar revenue and positive cash flow, yet CEO Ali Ghodsi has been comfortable raising privately. Anthropic is economically less conventional, but it has already put its full prospectus in front of the market.

Even a public filing is not a listing. The SEC still has to declare the registration effective, the company has to set a price range, run a roadshow and price the shares. Until then, any reported date or valuation is a plan.

IPO evidence	What it tells us
Price range set, registration effective	The offering is about to price and trade
Public S-1 filed	Investors can now see the company in detail
Confidential S-1 filed	The company has entered the SEC review process
Lead underwriters chosen	The offering is moving toward execution
Management gives a listing window	The company has shown real timing intent
IPO-ready finance and governance	The company could list, although it may still wait

Is Anthropic the private AI company closest to an IPO right now?

Yes. Anthropic is the clear number one because it is the only major private AI company with a public S-1 on file.

The filing, made on 26 October 2026, says Anthropic has applied to list its Class A common stock on the Nasdaq Global Select Market under the symbol ANT. Goldman Sachs and Morgan Stanley are listed first on the cover, ahead of J.P. Morgan, Citigroup and more than 20 other banks.

The deal itself is not set. The cover still leaves the price range and the number of shares blank, and the registration is not effective. The Wall Street Journal reported that Anthropic aims to go public during the week of 16 November and that valuation talk reaches $1.5 trillion to $2 trillion. Those are reported targets, not terms.

What the filing already shows is a business far larger than many investors assumed. Revenue reached $16.3 billion in the first half of 2026. More than 1,900 customers generated at least $1 million each over the twelve months to 30 September.

The process could still slow. Market conditions could turn, investors could push back on price, or the SEC review could take longer. None of that changes where Anthropic stands today.

Among the private AI companies we reviewed, Anthropic is the only one actually in the final stretch of a listing.

Is OpenAI still close to an IPO after ruling out 2026?

OpenAI is still the second-closest private AI company to an IPO, but it is no longer in the same tier as Anthropic, because its own CEO has pushed the timing past 2026.

OpenAI submitted a confidential S-1 earlier in 2026. That already separates it from almost the entire private AI market. Major banks including Goldman Sachs, Morgan Stanley and JPMorgan have been involved around its preparations, according to Axios. OpenAI's confidential filing sets it apart; we compare who is ahead and who is falling behind in our AI infrastructure report.

The difference is intent. When it announced the filing, OpenAI said it had not decided on timing. Then, in an interview with Fortune released on 12 September 2026, Sam Altman said that right now would be an ill-advised moment to go public, citing safety, and answered “not 2026” when asked about this year.

A confidential filing and a later listing do not conflict. The filing keeps the option open. It just means OpenAI has chosen not to use it yet, while Anthropic has.

That gap could still close quickly in 2027. A confidential filer can accelerate once management decides the timing is right, and OpenAI remains far closer to a listing than Databricks, Anduril or anyone else here.

For now, OpenAI has an available strategic choice. Anthropic has a transaction in progress.

What does Anthropic's S-1 reveal about its business?

Anthropic's S-1 shows very fast revenue growth, a much smaller operating loss and a huge headline net loss that comes mostly from accounting, not cash.

Revenue for the six months to 30 June 2026 was $16.3 billion, up from $1.2 billion in the same period of 2025. The operating loss shrank to $693 million from $2.1 billion. Compute and infrastructure costs rose to $8.9 billion from $2.3 billion, which shows how much of each dollar of revenue still goes to chips and cloud capacity.

The GAAP net loss was $80.7 billion. That number is real but misleading on its own: the filing books about $72.4 billion of changes in the fair value of financial instruments, which is a noncash item tied to its financing. It is not cash burn, and investors will look at the operating line instead.

Going first also helps Anthropic tell its own story. OpenAI has a wider business spanning consumer subscriptions, enterprise software, APIs and huge infrastructure commitments, and it has chosen to stay private for now. Anthropic gets to present frontier-model economics before investors can benchmark every number against OpenAI's disclosures.

That advantage only lasts if the pricing goes well.

Has OpenAI fixed the corporate structure that once made an IPO difficult?

OpenAI has removed one of the biggest practical barriers to an IPO by converting its operating business into a public benefit corporation with conventional equity, but it has not published a prospectus that shows how public shareholders would fit in.

The old capped-profit arrangement was awkward for public markets. Investors were buying economic rights through a structure designed to limit returns and preserve nonprofit control. That could work with a small group of sophisticated private backers, but it would have been difficult to explain to millions of public shareholders.

OpenAI Group PBC now issues conventional equity. Microsoft and other investors can hold recognizable ownership stakes, while the OpenAI Foundation retains control of the group.

The governance is still unusual. Public shareholders would own an economic interest in a company whose ultimate controlling body is mission-driven. Without a public S-1, investors cannot yet see exactly how board appointments would work or when the Foundation could override commercial incentives.

Anthropic has now shown its version of the answer. Its S-1 describes the Long-Term Benefit Trust as five financially disinterested trustees who, through a non-economic Class T stock, can generally elect and remove the Class T directors. The filing also gives a separate Class F stock 50.1% of total voting power. Public Class A buyers would therefore own economics with limited control.

Frontier-model IPOs come with a governance question that Databricks or Glean would not have to explain to the same degree. OpenAI has made its structure more investable, but Anthropic is the one public investors will judge first.

Is Databricks actually more IPO-ready than Anthropic and OpenAI?

Databricks may be the most financially ready private AI company in the world, but it is not closer to an IPO than Anthropic, because it has chosen private capital and has not been reported to file.

The numbers are already well beyond what most software companies had when they went public. Databricks said in August 2026 that its annualized revenue run rate had passed $7 billion, growth remains fast at its current scale, and the company is cash-flow positive.

Its latest fundraising says even more about why an IPO can wait. Databricks raised about $5 billion at roughly a $190 billion valuation. Ali Ghodsi told TechCrunch that the company initially wanted around $1 billion, but investor demand reached roughly $15 billion. At a $190 billion valuation, we compare what each company was worth then and now in our report on top AI infrastructure companies.

That tells us something important about the IPO market. Databricks can obtain billions of dollars, offer liquidity to shareholders and set a new valuation while staying private. Going public would add quarterly reporting and disclosure without solving an urgent financing problem.

Management has been clear about this. Morningstar reported in September 2026 that Ghodsi had ruled out a 2026 listing and that Databricks had not confidentially filed, which points to 2027 at the earliest.

If Databricks changed its mind, it could probably move faster than most companies below it. Until then, financial readiness alone cannot put it ahead of a company with a public S-1 or one already in SEC review.

Company	Where it stands now	Main thing holding it back
Anthropic	Public S-1 filed; Nasdaq listing applied for as ANT; price not set	SEC review, pricing and market conditions
OpenAI	Confidential S-1 filed; CEO says not 2026	Management has chosen to wait
Databricks	$7B+ run rate and cash-flow positive; no filing reported	Very little need to go public

Is Anduril likely to be the next big AI IPO after Anthropic and OpenAI?

Anduril is one of the strongest eventual IPO candidates, but its next move under discussion is another private round, not a listing.

Anduril's business is already reaching a size that would support a major public listing. The company told Fortune that it expects about $4.3 billion of revenue this year, up from roughly $2.2 billion in 2025. Few private technology companies have reached that revenue level before going public.

Its capital needs are also unusually large. Anduril is building Arsenal-1 in Ohio, a roughly million-square-foot weapons manufacturing facility, while funding autonomous aircraft, missiles, underwater systems and the Lattice software platform. Those projects absorb far more capital than a normal software company needs.

An IPO would therefore make strategic sense eventually. Public equity can finance factories and long defense programs without repeatedly returning to venture investors.

Anduril still has another option, and investors keep making it attractive. Its last completed round, in May 2026, valued it at about $61 billion. In late October 2026, The Information reported that executives had discussed a possible fundraising in the first quarter of 2027 at around $120 billion. Anduril said no decisions had been made on any future financing, so that figure is talk, not a price.

That lowers the urgency of an IPO. Anduril is becoming a public-company-scale business while still funding itself like a private one.

We would put Anduril near the front of the later queue, but there is still a real gap between an obvious eventual candidate and a company preparing its listing.

Could Lambda or Together AI jump ahead of older AI startups?

Lambda now has the clearer IPO path of the two, with a planned 2027 listing reported in October 2026, while Together AI has the earlier but softer timing hint.

On 6 October 2026, TechCrunch, citing The Wall Street Journal, reported that Lambda was raising up to $4 billion at a $14.5 billion pre-money valuation, led by Coatue and Blackstone, in what could be its last private round before a planned 2027 IPO. The round had not closed, and Lambda had reportedly been expected to list in 2026 before pushing that back.

The same reporting said Lambda's backlog grew from $15 billion in June to $50 billion in September, much of it from a $35 billion commitment by Anthropic. Backlog is not revenue, and that much reliance on one customer is something public investors would study closely.

Lambda is also building the treasury machinery of a much larger company. After an earlier $926 million term loan rated investment grade by Moody's, it announced on 1 October 2026 a $1 billion investment-grade delayed-draw term loan. Its leadership, with Michel Combes as CEO and John Donovan as chairman, fits a company that relies on debt markets.

Together AI CEO Vipul Ved Prakash told Axios in July 2026 that an IPO in 2027 was possible, after an $800 million raise at an $8.3 billion valuation. No filing or firmer date has been reported since.

Lambda has the more concrete plan today. Either company could move up quickly if a confidential filing appears.

Are ElevenLabs and Glean already big enough to IPO?

ElevenLabs and Glean are both becoming large enough for public markets, but neither has shown the filing activity that would make an IPO look close today.

ElevenLabs has the more striking growth rate. The company ended 2025 around $350 million of ARR and said it had passed $500 million within the first four months of 2026. That means it added roughly $150 million of annualized recurring revenue in about one quarter. Adding $150 million of recurring revenue in one quarter is fast; we look at who is growing the fastest in our AI infrastructure market report.

It is also using private markets for liquidity. On 30 September 2026, ElevenLabs closed a $300 million employee tender offer at a $22 billion valuation, double its $11 billion February mark, led by Wellington and T. Rowe Price. That money went to selling shareholders, not the company, and it removes some of the pressure an IPO would otherwise relieve.

Glean's growth is slower in absolute dollars but clean for an enterprise software company. Glean says it took around nine months to go from $100 million to $200 million of ARR and only six more months to reach $300 million. Its Fortune 500 customer count also nearly doubled year over year.

There is one wrinkle in the Glean number. TechCrunch pointed out that part of Glean's pricing is consumption-based, so the $300 million figure is not perfectly equivalent to traditional contractual SaaS ARR. Public investors would separate fixed commitments from annualized consumption.

Both companies have enough momentum to become serious IPO candidates. Today, though, they still look like companies using private capital to maximize growth rather than companies getting ready to sell shares on Nasdaq.

Is Harvey close to an IPO, or is the legal AI company still too early?

Harvey is growing fast enough to enter the IPO conversation, but the legal AI company still looks at least one cycle behind in public-market readiness, and no filing or IPO schedule has been reported.

Harvey raised $550 million at a $15.5 billion valuation on 9 September 2026, co-led by Diffusion and Lightspeed, up from $11 billion in March. Earlier reporting put it at roughly $350 million of ARR. A $550 million round at $15.5 billion is big; we follow who has raised the most, and from whom, in our AI infrastructure report.

Those are serious numbers for a business founded only a few years ago. Harvey says 80% of Am Law 100 firms use it. In October 2026 it announced that CVS Health had deployed it across corporate affairs, and that it now runs inside a FedRAMP Class D certified cloud through Knox Systems, which opens doors in government.

The financial profile is less mature. Harvey is still unprofitable and is spending heavily on product development, international expansion and custom agents built for customers.

Another year of execution could improve Harvey's IPO case a lot. Public investors would want to see ARR keep compounding without hiring and infrastructure costs rising just as quickly.

Harvey belongs on the watchlist now. Calling it one of the companies closest to an IPO would still be premature.

Are huge private valuations actually delaying AI IPOs?

Huge private valuations and easier shareholder liquidity are delaying several AI IPOs because companies can get many of the benefits of being public without listing.

Databricks is the strongest example. Investor demand was reportedly three times the $5 billion it ultimately raised. When private investors value the company around $190 billion and provide billions in fresh capital, an IPO solves very little, and its CEO has reportedly pushed any listing past 2026.

Anduril is following a similar pattern. Its valuation moved from roughly $30 billion to about $61 billion, and a possible next round has been reported at around $120 billion. ElevenLabs went from $11 billion to $22 billion in about seven months through a tender rather than an IPO.

Liquidity has also become easier to create privately. Scale AI showed the extreme version when Meta invested more than $14 billion in the company. Scale said part of the proceeds would be distributed to shareholders and vested equity holders, giving them substantial liquidity while Scale remained independent.

This removes one of the old pressures that pushed mature startups toward the stock market. Employees can sell some shares, venture funds can return capital, and companies can raise multibillion-dollar rounds without publishing an S-1.

The irony is simple: the AI funding boom is creating dozens of companies large enough to IPO while giving many of them less reason to do it. Anthropic is the exception, and its pricing will test whether public markets can match private ones.

Could a weak Anthropic IPO shut the AI IPO window again?

A disappointing Anthropic IPO could delay several other private AI listings because it will be the first real public price for a frontier-model company.

The S-1 already gives investors the numbers they will argue about. Compute and infrastructure costs were $8.9 billion in the first half of 2026, more than half of revenue. Anthropic depends heavily on cloud partners and carries long-term infrastructure commitments that ordinary software companies do not.

Revenue presentation will matter too. Axios has highlighted differences in how Anthropic and OpenAI account for revenue distributed through partners. Investors now have Anthropic's own disclosures to test how much reported revenue turns into retained economics.

The valuation will amplify every concern. If the reported $1.5 trillion to $2 trillion range holds, good growth will not be enough. Public investors would have to believe the growth can stay exceptional for years while margins improve. The deal has not priced, so any reaction is still a scenario.

A strong reception would make life easier for OpenAI in 2027 and set a public benchmark for Lambda, Databricks, Anduril and the rest of the queue.

A weak debut would give well-funded private companies an easy reason to wait. Most of them have enough cash and investor demand to do exactly that.

So which private AI companies are closest to an IPO today?

Anthropic is the private AI company closest to an IPO, and it now stands alone at the front. OpenAI is second but has pushed its timing past 2026, and the distance to everyone else is substantial.

Anthropic has the hardest evidence we can find: a public S-1 filed on 26 October 2026, a Nasdaq application under ANT and more than 20 banks on the cover. Price range, share count and listing date are still not set.

OpenAI sits behind it. It has crossed the confidential-filing threshold and has major banks around the process, but Sam Altman said in September 2026 that it would not list this year.

Lambda and Databricks form the 2027 group. Lambda has a reported plan to list in 2027 after one last private round. Databricks has more than $7 billion of annualized revenue and positive cash flow, but management prefers private markets for now.

Anduril is the strongest large candidate behind them. Revenue is heading toward the mid-single-digit billions and its factories consume huge amounts of capital, but the next step under discussion is a private round reported at around $120 billion.

The next group is more fluid. Together AI has floated a possible 2027 IPO without a filing. ElevenLabs and Glean are reaching revenue levels where an IPO can become realistic quickly. Harvey is growing fast but still needs more financial maturity.

The key observation is how short the true IPO queue is. Plenty of private AI companies could go public. Only Anthropic is visibly doing it right now.

Rank	Private AI company	How close does it look today?	Why
1	Anthropic	In the final stretch	Public S-1 filed 26 October 2026, Nasdaq application as ANT; price range not yet set
2	OpenAI	Close, but not in 2026	Confidential S-1 filed, major banks involved, CEO ruled out a 2026 listing
3	Lambda	2027 candidate	Reported plan to list in 2027 after a raise of up to $4B; $1B investment-grade debt
4	Databricks	Ready whenever management wants	$7B+ revenue run rate, cash-flow positive, but 2026 reportedly ruled out
5	Anduril	Strong later candidate	~$4.3B expected revenue; possible private round reported at ~$120B, not decided
6	Together AI	Emerging candidate	CEO has publicly discussed a possible 2027 IPO; no filing reported
7	ElevenLabs	Fast-rising future candidate	$500M+ ARR; $22B valuation in a $300M employee tender
8	Glean	Fast-rising future candidate	$300M revenue run rate/ARR with accelerating enterprise adoption
9	Harvey	Plausible later candidate	$15.5B valuation, ~$350M ARR reported, still loss-making