AI boom faces a reality check as OpenAI, Anthropic head for IPOs

The artificial intelligence boom is approaching a crucial test as OpenAI and Anthropic prepare to enter public markets, where their valuations will face scrutiny from investors armed with detailed financial disclosures rather than private funding rounds and growth expectations.

For the past three years, some of the world's most consequential technology companies have largely been valued through private funding rounds, with each new round setting a fresh benchmark for their worth. But that could change once AI's biggest names become public companies.

Suyash Karn, co-founder and CEO of Interact AI, argues that the upcoming IPOs could mark the point when the market gets its first real look at the economics behind the AI boom.

“For three years, the most consequential companies in technology have been priced by people who were not allowed to sell. Private rounds set the number, the press repeated it, and the number only went up,” Karn said. “That era ends this October.”

Anthropic confidentially filed for an IPO on June 1, followed a week later by OpenAI. SpaceX, with xAI folded inside it, had already listed in June. Together, the three companies represent some of the biggest bets on the future of AI and technology.

“The three companies that between them define the AI era are walking, at nearly the same moment, into the one room where narrative does not set the price. Public markets do,” Karn said.

The numbers attached to these companies are enormous.

Karn points to Anthropic's last private round in May, which valued the company at $965 billion on roughly $47 billion in annualised revenue. Investors are now reportedly expecting an October debut at $2 trillion or more.

OpenAI, meanwhile, raised $122 billion in what Karn describes as the largest private round in Silicon Valley history, at an $852 billion valuation. Its CEO has reportedly set a floor of $1 trillion for any listing.

Together with SpaceX, these offerings could seek more than $200 billion from public markets, compared with the $45 billion raised by the entire US IPO market in 2025, according to Karn.

But Karn's argument is not that the AI boom is imaginary.

“So the honest question is not whether AI has progressed. It has, unambiguously,” he said.

Anthropic, he notes, has gone from roughly $1 billion in annualised revenue in late 2024 to a reported $47 billion by May. OpenAI is generating around $2 billion a month, with enterprise revenue approaching parity with consumer revenue.

“The models moved from novelty to line item in enterprise budgets. That part is real,” Karn said.

The question, therefore, is whether the enormous valuations being placed on these companies can withstand the scrutiny that comes with being publicly traded.WHAT HAPPENS WHEN THE NUMBERS ARE OUT?

The public market could provide a very different test from the private market.

Karn points to the experience of SpaceX after its listing. Institutions paid $135 at allocation, while retail investors paid as much as $161 in the first session. The stock then touched $225 within days, before falling below $105 after its first earnings report — a 53% drawdown in seven weeks — before recovering above its IPO price in mid-August.

“Nothing about the company's rockets changed between June and August,” Karn said. “What changed is that a real market, with sellers in it, got to vote twice, in opposite directions, on the same company.”

That, he argues, could be particularly relevant for OpenAI and Anthropic.OPENAI AND ANTHROPIC FACE DIFFERENT TESTS

While the two AI companies have often been viewed as similar bets on the future of artificial intelligence, their financial profiles are beginning to look different.

Anthropic is reportedly expecting its first operating profit, with Karn citing a figure of around $559 million on approximately $10.9 billion of second-quarter revenue. The company has also cautioned that the margin is partly influenced by a discount on its compute contracts.

OpenAI, on the other hand, is projected to lose around $14 billion this year and does not expect profitability before 2029, according to the article.

“For years the two were treated as interchangeable bets on the same future,” Karn said. “The IPO process is forcing a distinction the private market never demanded.”

He added: “One is a business that happens to burn cash while growing, the other is a mission that happens to have revenue. Public investors price those very differently.”WHY THE IPOs MATTER BEYOND OPENAI AND ANTHROPIC

Karn also points to another unusual feature of the current AI cycle: these companies do not necessarily need the money.

“Both companies raised more private capital in the past twelve months than most countries' entire venture ecosystems deploy in a decade,” he said.

“They did not need to go public for money. They filed anyway, because a confidential S-1 is an option, and options on a trillion-dollar window are worth holding.”

For Karn, that decision itself suggests that companies and investors cannot be certain that the current valuation environment will remain open indefinitely.

But the most important consequence of the IPOs, he argues, will not necessarily be the headline valuation.

It will be the information that comes with going public.

“When these S-1s go public, we get the first audited look at frontier AI economics, real gross margins after computing, real customer concentration, real unit costs,” Karn said.

“Every founder building on these models, every enterprise negotiating an AI contract, every policymaker allocating compute budgets has been working off vibes and leaked run rates. In a few weeks, we get the actual ledger.”

That could provide a much clearer picture of whether the economics behind the AI boom justify the valuations being placed on it.THE AI BOOM'S BIGGEST TEST

Karn believes the IPOs could ultimately produce two very different outcomes.

“If the listings hold, the bubble was a market,” he said. “If they trade like SpaceX did, up on faith, down on disclosure, the repricing will travel through every AI startup's next round, on every continent.”

“Either way, the guessing ends,” he added.

That is why he compares the moment to the dot-com era. The lasting lesson from that period, he argues, was not simply which internet companies survived, but how public markets eventually forced investors to confront the underlying economics of the sector.

“The dot-com era's most useful gift was not the companies that survived it. It was the moment public markets forced the internet's economics into the open, and everyone could finally build against real numbers,” Karn said.

“October is that moment for AI. Progress is not in question. Price discovery is.”

(Disclaimer: The views, opinions, recommendations, and suggestions expressed by experts/brokerages in this article are their own and do not reflect the views of the India Today Group. It is advisable to consult a qualified broker or financial advisor before making any actual investment or trading choices.)- EndsPublished By: Sonu VivekPublished On: Sep 11, 2026 18:06 IST