OpenAI and Anthropic File for Initial Public Offerings Amid Scrutiny Over AI Infrastructure Costs

2026-06-10

The BareStory

Artificial intelligence developers OpenAI and Anthropic have submitted confidential documents to U.S. regulators in preparation for initial public offerings. OpenAI filed its prospectus on Monday, following a similar move by Anthropic exactly one week prior.

The filings coincide with the industry's shift toward using "tokens"—data units measuring computing consumption—as the primary billing metric for artificial intelligence models. To support large-scale token generation, companies are taking on significant infrastructure expenses. According to corporate reports, OpenAI recently secured a computing agreement with chipmaker Cerebras extending through 2028. Meanwhile, Anthropic committed to paying SpaceX $1.25 billion monthly for three years to utilize its data center capacity. SpaceX, which reported that its AI division currently operates at a financial loss, is also slated to debut on the Nasdaq this week.

The escalating costs of AI development and the reliance on token-based billing have prompted discussions among financial and technology observers regarding long-term returns on investment. Various industry executives and analysts have noted that translating token usage into tangible business value remains an ongoing challenge for organizations.

This billing model has also drawn criticism from within the tech sector. Palantir Chief Executive Alex Karp stated on Wednesday that enterprise customers are dissatisfied with the leading AI laboratories. Karp alleged that these companies fail to understand practical business operations and focus excessively on burning through tokens to demonstrate productivity, a practice he labeled "tokenmaxxing." Despite his criticisms of the frontier labs, Karp disclosed that Palantir currently operates the majority of Anthropic’s publicly discussed projects.

Left Perspective

Shielding Against Speculative Extraction
Exposing Fictional Value Metrics
Defusing the Capital Black Hole

Right Perspective

Catalyzing Free Market Discovery
Fueling High-Risk Innovation Engines
Enforcing Ruthless Market Efficiency

Left Perspective

• Shielding Against Speculative Extraction Prioritizing the protection of retail investors, this camp views the simultaneous IPOs of OpenAI, Anthropic, and SpaceX as a mechanism to offload unsustainable infrastructure costs onto the public. By rushing to the Nasdaq while operating at steep financial losses—such as SpaceX's AI division—these tech giants are socializing their developmental risks while seeking to privatize their eventual gains. This aggressive push into public equities is interpreted less as a milestone of corporate maturity and more as a liquidity bailout for early venture capitalists.

• Exposing Fictional Value Metrics The industry-wide shift to "tokens" as a primary billing metric is scrutinized as a deliberate obfuscation of actual economic utility. By measuring success through computing consumption rather than tangible business returns, these laboratories create an artificial metric of progress to justify their staggering valuations. Palantir CEO Alex Karp's critique of "tokenmaxxing" validates consumer advocate fears that these companies are burning resources to simulate productivity rather than delivering real-world solutions.

• Defusing the Capital Black Hole There is deep systemic concern that massive capital concentration—exemplified by Anthropic's $1.25 billion monthly payment to SpaceX—diverts vital resources away from practical economic needs into a highly speculative ecosystem. Operating on a model of infinite cash burn to support data centers and chip agreements through 2028 creates dangerous structural vulnerabilities across the tech sector. The overarching risk is a catastrophic market correction when the disparity between astronomical computing costs and actual enterprise ROI inevitably collapses, leaving public stakeholders holding the bag.

Right Perspective

• Catalyzing Free Market Discovery Valuing systemic stability and capital efficiency, this camp views the transition of OpenAI, Anthropic, and SpaceX into public markets as a healthy, necessary step for transparent price discovery. Filing confidential IPO documents subjects these frontier laboratories to rigorous regulatory scrutiny and shareholder accountability, replacing opaque private funding with free-market discipline. This transition allows public capital markets to efficiently evaluate competing business models and reward the most viable paths to technological dominance.

• Fueling High-Risk Innovation Engines Massive infrastructure commitments—like OpenAI's 2028 Cerebras pact and Anthropic's multi-billion dollar SpaceX contract—are interpreted as essential, visionary capital expenditures. In a market realist framework, temporary financial operating losses are a necessary precondition for securing the immense computing scale required to build foundational technologies. Token-based billing is not viewed as a deception, but rather as a standardized, scalable commodity framework designed to efficiently price and trade an entirely new digital asset class.

• Enforcing Ruthless Market Efficiency Internal industry friction, such as Alex Karp’s condemnation of "tokenmaxxing," is welcomed as evidence of the free market naturally course-correcting toward tangible return on investment. Customer dissatisfaction serves as a vital economic signal, forcing AI laboratories to pivot from raw compute generation to delivering concrete, profitable business efficiencies. Ultimately, the fact that practical operators like Palantir successfully manage the majority of Anthropic’s public projects proves that market competition will seamlessly cull inefficiencies and translate raw computing power into sustained economic growth.

How it may affect me

As a U.S. reader:

• Retail investors will soon have the opportunity to buy shares in prominent artificial intelligence and technology companies like OpenAI, Anthropic, and SpaceX, opening new avenues for public investment but also exposing everyday shareholders to the steep financial losses currently carried by private investors.

• In the short term, businesses and consumers utilizing these artificial intelligence platforms may experience pricing and service models optimized for raw computing consumption rather than practical utility, potentially delaying the delivery of cost-effective, real-world technological solutions.

• The shift to public markets will subject these frontier laboratories to rigorous regulatory scrutiny and shareholder demands, which may ultimately force the companies to become more transparent and shift their focus toward delivering tangible business efficiencies.

• Over the long term, everyday public stakeholders and broader market stability could be impacted by a significant market correction if the billions of dollars being concentrated into artificial intelligence infrastructure fail to yield profitable returns on investment.

OpenAI and Anthropic File for Initial Public Offerings Amid Scrutiny Over AI Infrastructure Costs

The BareStory

Left Perspective

Right Perspective

How it may affect me

Read the story at