Two major artificial intelligence companies, OpenAI and Anthropic, are preparing to launch initial public offerings (IPOs) this year amid mounting concerns that the AI sector may be approaching a market bubble. Industry insiders and investors are closely watching these developments, which come at a time of increasing skepticism over AI’s promised transformative impact.

The enthusiasm surrounding these IPOs has been fueled by extensive media promotion of AI’s potential to revolutionize multiple industries, as well as by recent market activity including the highly anticipated SpaceX offering. SpaceX, a company closely associated with AI advancements, saw an initial share price increase following its June 12 IPO, although the gains subsequently diminished, with shares falling below their debut closing price within days.

Despite the excitement, some analysts warn that the rush to go public may signal that insiders believe the AI boom’s peak is near. Critics have pointed to several indicators suggesting that AI’s capabilities and cost benefits have been overstated. For instance, companies adopting AI-driven workforce reductions have sometimes reversed course, rehiring staff after cost savings failed to materialize. Research from Forrester and the British consultancy OrgVue highlighted the “AI boomerang effect,” describing this trend of workforce reinstatements following hasty AI implementation.

Pricing models for AI services have also created challenges. Firms such as Uber have publicly acknowledged that the productivity gains expected from AI investments have not matched the sizable expenditures. Uber’s entire 2026 budget for Anthropic’s Claude AI platform was reportedly exhausted by mid-March, underscoring concerns about the high and sometimes unpredictable costs associated with AI usage, especially as companies have moved from flat-rate subscriptions to metered “per-token” pricing.

Infrastructure constraints add further complications. The expansion of AI depends heavily on large-scale data centers, but many proposed facilities are facing delays or cancellations due to local opposition, regulatory hurdles, and capacity shortages in power grids and critical electrical equipment. Nearly half of the data centers planned in the U.S. for this year are expected to be postponed or scrapped, according to industry reports. JPMorgan Chase noted that most centers scheduled for completion next year have yet to commence construction.

Despite these headwinds, Anthropic filed a confidential IPO registration with the U.S. Securities and Exchange Commission on June 1, with OpenAI reportedly considering a similar move. Both companies carry massive private valuations—Anthropic at approximately $965 billion after a recent $65 billion funding round, and OpenAI valued at about $852 billion—yet neither has posted a profitable quarter. Published estimates indicate OpenAI recorded a $38.5 billion loss in 2025 on revenues of $13.1 billion, raising questions about sustainability and funding sources as the companies prepare to access public capital markets.

Investors should also consider inherent risks associated with IPOs. Historically, such public offerings have often served as liquidity events favoring founders and early investors rather than immediate opportunities for new investors. Past market cycles, including the late 1990s dot-com craze, illustrate how enthusiasm can lead to excessive valuations followed by significant losses. Notable investment figures like Warren Buffett have long cautioned against participation in IPOs, citing them as favoring insiders rather than outside shareholders.

The upcoming AI IPOs, like SpaceX’s, appear to be leveraging a wave of investor enthusiasm despite persistent questions about the near-term viability and profitability of AI technology. These offerings may reflect a critical juncture where insiders seek to capitalize on current market conditions before potential declines. As AI firms continue to raise vast sums and confront growing operational challenges, the unfolding IPOs will test whether confidence in AI’s future remains justified or if the sector faces a more pronounced correction.