Skip to content

Lurking beneath the Anthropic, OpenAI and SpaceX IPO race is a colossal ‘reality check,’ experts say

Together, the three companies are expected to command valuations rarely seen in public markets.

Stock photo of the SpaceX logo, rendered against a colorful graph seen in the background.
SpaceX said in a regulatory filing that it intends to sell 555 million shares at $135 apiece in what would be the largest IPO ever. Photo by Matthew Modoono/Northeastern University

When AI startup Anthropic confidentially filed for an initial public offering, or IPO, last week, it intensified what many observers have framed as a mad rush to cash in on investor enthusiasm amid one of the most historic market booms the tech sector has ever seen. 

Following months of what many have described as an historic stock rally driven largely by a surge of investment in companies — such as Nvidia, AMD and Intel — that produce AI parts, a wave of AI IPOs may serve as a reality check for a market largely driven by expectations about the technology’s future potential, experts say. 

Anthropic joins SpaceX, which filed to go public last month, as two of the three most highly anticipated equity offerings in recent memory, with the third — OpenAI — expected to follow suit in the coming weeks. 

Together, the three companies are expected to command valuations rarely seen in public markets. SpaceX is widely expected to debut at a valuation exceeding $1 trillion, potentially making it the largest IPO in history, while OpenAI’s and Anthropic’s expected valuations appear to be closing in on $1 trillion, respectively. On Wednesday, SpaceX said in a regulatory filing that it intends to sell 555 million shares at $135 apiece, raising about $75 billion in what would be the largest IPO ever.

While SpaceX, whose primary business is the making of rockets, occupies a different corner of the technology sector than Anthropic and OpenAI, its operations depend on advances in machine learning, automation, autonomous decision-making and ultimately the broader artificial intelligence business, said Karthik Krishnan, an associate professor of finance at Northeastern University. 

Northeastern Global News, in your inbox.

Sign up for NGN’s daily newsletter for news, discovery and analysis from around the world.

Krishnan, who has studied IPOs, said that the process tends to follow a predictable pattern. When a company first goes public, its shares experience a “price pop,” or a precipitous rise, followed by a decline usually for some days and weeks, before the price stabilizes. 

With such an historic injection of investor capital — billions and billions of dollars in fresh funding — poised to enter the AI sector, some observers are questioning whether the lofty valuations attached to the three companies will ultimately hold up once public markets have more time to assess their long-term prospects.

“It’s hard to tell, at this stage, whether these are true valuations, or an expectation of what investors feel they can flip,” by reselling their shares to other buyers on the secondary market, Krishnan said.

IPO process, explained 

To make more sense of how historic the upcoming IPOs are, it helps to first understand how an IPO works.

After filing the necessary paperwork with regulators, companies seeking to go public typically embark on a multi-week “roadshow” in which executives and investment bankers meet with prospective investors to gauge interest in the offering, Krishnan said. This is what is commonly known as book building.

During the book building process, banks collect “indications of demand,” or interest in the shares being offered, and use that information to help determine the stock’s final offering price, Krishnan said. After weeks of meetings with investors, the investment banks and company executives involved in the process settle on an offering price.  Shares are then sold when trading begins on public markets.

Historically, Krishnan noted that access to IPO shares has been concentrated among large institutional investors, such as hedge funds and mutual funds, rather than mom-and-pop retail traders or individuals who trade on their own behalf.

Some companies, including SpaceX, have explored setting aside an unusually large chunk of shares specifically for retail investors. In SpaceX’s case, up to 30% of shares are potentially set for that purpose — “an unusual” nuance in the scheme, said Krishnan.

While the price pop helps build excitement and later inspires discussions around whether companies could have received more money for their shares, “retail investors will come in and buy it off your hands in the secondary market. So there may actually not be money left on the table,” he said.

A reality check

The enthusiasm about the three juggernaut IPOs has been fueled by the belief that artificial intelligence will transform the global economy and create enormous new markets for the companies leading the charge.

For Krishnan, so-called “pure-play” AI companies like Anthropic and OpenAI have already demonstrated the value in artificial intelligence as a standalone business.  

“Anyone of us who’s used ChatGPT knows this is not just a website with a ‘.com’ attached to it,” Krishnan said. “It’s a real thing, and we’re using it, and we’re paying for it. I think the economic value is there.”

But others note that today’s exuberance is reminiscent of the early 2000s, when many early internet companies saw their valuations soar before the dot-com bubble collapsed. 

Marc Meyer, the Robert J. Shillman Professor of Entrepreneurship and the Matthews Distinguished Professor at Northeastern, said the current enthusiasm around AI companies “feels similar” to that period, recalling how shares of Sun Microsystems, then a leading maker of servers and computing systems and software, plunged from about $70 to $5 within months of the bubble bursting. 

A self-described “product guy,” Meyer added that the companies that “have more value for me” are the ones that embed the technology into products and workflows customers already use rather than treating AI as a destination unto itself. He shared examples of agentic AI companies, such as ServiceNow, Salesforce or Qualcomm. 

Despite his view that AI can stand on its own, “Whether the valuation is reflective of the true economic value is not something I can say,” Krishnan said. “The price is where the demand is, right? At the end of the day, the price is set by who’s willing to buy it — and at what price,” Krishnan said. 

Tanner Stening is an assistant news editor at Northeastern Global News. Email him at t.stening@northeastern.edu. Follow him on X/Twitter @tstening90.