Why AI Training Costs Are Driving Companies to Go Public

Anthropic President: The Financial Reality Behind AI's Race to Go Public

The artificial intelligence industry is entering a new financial era — one defined not just by breakthrough technology, but by the staggering cost of building it. Anthropic Co-Founder and President Daniela Amodei made that reality clear on Thursday, June 4, when she spoke at the Bloomberg Tech conference and explained why companies at the frontier of AI development are increasingly looking toward public markets for capital.

Her comments come just days after Anthropic confirmed it had confidentially filed for an initial public offering (IPO) with the U.S. Securities and Exchange Commission (SEC), setting the stage for what analysts expect to be one of the largest stock market listings in recent history.

Training Costs: The Billion-Dollar Bottleneck

To understand why Anthropic and its peers are eyeing Wall Street, you first need to understand what it actually costs to build a competitive AI model. Training large language models (LLMs) like Claude requires massive computational infrastructure — thousands of high-end GPUs running continuously for weeks or months, enormous energy expenditures, and highly specialized engineering talent commanding top-of-market salaries.

Amodei did not mince words on the subject. Speaking at the Bloomberg Tech conference, she said: "It's a really big upfront cost to train the models and to serve inference on them. And I think that my guess is that over time, the sort of core set of companies that are working to advance the frontier are just going to need access to capital. And I think the public market is very well suited to that."

This statement encapsulates a core tension in the AI industry today. The frontier is expensive to maintain. Staying competitive means continuously training newer, larger, and more capable models — and each generation costs more than the last. Private funding rounds, even very large ones, may not be sufficient to sustain this pace indefinitely.

Anthropic's Path to the Public Markets

On June 1, 2025, Anthropic announced it had submitted a draft registration statement to the SEC, a move that gives the company the option to proceed with an IPO once the regulatory review process is complete. The company was careful to note that its listing plans remain contingent on market conditions and other factors, leaving open the timeline for when shares might actually begin trading.

When asked directly at the Bloomberg Tech conference whether Anthropic was racing to be the first AI company to hold an IPO, Amodei declined to comment — a calculated silence that itself speaks volumes about the competitive dynamics at play in the industry right now.

Just four days before the IPO filing announcement, on May 28, Anthropic disclosed that it had raised $65 billion in fresh funding. That figure underscores both the enormous investor appetite for AI exposure and the equally enormous capital requirements that make such fundraising necessary in the first place.

A Three-Way Race to the Stock Exchange

Anthropic is not alone in its march toward a public listing. According to reporting by the Financial Times, the company's IPO plans are part of a broader three-way race to go public that also includes OpenAI and SpaceX. If all three proceed, they would represent some of the largest stock market listings ever recorded — a seismic event for both the technology sector and the broader financial markets.

The competitive dimension of this race matters. Being first to market with a public listing could confer significant advantages, from brand recognition among retail investors to the ability to use publicly traded stock as currency for acquisitions and talent retention. It also establishes a public valuation benchmark that shapes how the entire AI industry is perceived and funded going forward.

Why the Public Market Makes Sense for Frontier AI

Amodei's argument for public markets is grounded in practical capital theory. While venture capital and private equity can write large checks, they operate under constraints — fund lifecycles, portfolio diversification requirements, and return expectations that may not align with the decade-long timelines required to achieve artificial general intelligence.

Public markets, by contrast, offer several structural advantages for capital-intensive, long-horizon businesses:

Scale: Public equity markets can absorb capital raises of a magnitude that no private round can easily match, giving companies the financial runway to pursue multi-year research agendas without constant fundraising cycles.
Liquidity: A public listing allows early investors, employees, and founders to realize returns, which in turn makes it easier to attract top talent willing to accept equity compensation.
Credibility and transparency: Public companies are subject to rigorous financial disclosure requirements, which can actually strengthen trust among enterprise customers who are evaluating AI vendors for mission-critical deployments.
Ongoing access: Once public, a company can return to markets through follow-on offerings, convertible debt, or other instruments whenever capital needs arise — a flexibility that private companies do not enjoy to the same degree.

What This Means for the AI Industry at Large

Amodei's candid remarks signal something important: the age of AI being funded primarily through secretive private rounds may be drawing to a close, at least for the largest players. As the compute and infrastructure requirements for frontier model training continue to scale, the pool of private capital capable of meeting those needs shrinks. Public markets become not just an attractive option, but a structural necessity.

This shift will have wide-ranging implications. Greater financial transparency around AI companies could fuel public and regulatory scrutiny of how these firms are spending their capital, what returns they are generating, and how they are managing the risks associated with increasingly powerful systems. It may also accelerate consolidation, as publicly traded AI companies gain the financial firepower to acquire smaller competitors and startups.

For investors, the prospect of publicly traded frontier AI companies represents an entirely new asset class — one with extraordinary upside potential but also significant uncertainty around timelines, competition, and the fundamental question of whether any given company will successfully navigate the path to transformative AI.

The Bottom Line

Daniela Amodei's remarks at Bloomberg Tech were brief, but they carried a clear message: building AI at the frontier is extraordinarily expensive, and the companies serious about maintaining their position will need the kind of capital that only public markets can reliably provide. With Anthropic's confidential IPO filing now on record and competitors like OpenAI preparing their own listings, the AI industry's financial center of gravity is shifting — from Sand Hill Road to Wall Street. For anyone watching the future of technology, this is a defining moment worth paying close attention to.