OpenAI files for IPO, following Anthropic

The confidential S-1 route gives OpenAI pricing flexibility and shields early financials from competitors during the quiet period, but it also means the public won't see burn rate, revenue concentration, or compute cost disclosures until the registration goes effective, which is where the real story lives.

The timing sits directly alongside Alphabet's announced $80 billion capital raise for AI infrastructure (covered here from TechCrunch, June 1). That raise signaled that compute scale, not model capability alone, determines competitive durability. OpenAI going public is partly a response to that same pressure: the company needs a capital formation mechanism that can match or outlast hyperscaler balance sheets over a multi-year horizon. Two frontier labs filing for IPOs within days of each other also changes the fundraising calculus for every private AI company below them, since institutional capital that flows into liquid public AI equities is capital that does not go into Series B and C rounds.

Watch whether OpenAI's S-1, once public, shows gross margins above 40 percent on its API business. If it does, the profitability-timeline concern embedded in the summary softens considerably. If margins are below that threshold at current revenue scale, the public-market scrutiny the summary flags becomes an active constraint on R&D spending within 18 months of listing.