A new phase in AI: from private rockets to public markets
The latest wave of AI company IPO plans refers to the simultaneous push by leading artificial intelligence firms to file listing documents, raise public capital, and lock in valuations while also facing growing pressure from enterprise buyers over costs, performance, and long-term flexibility in their AI architectures. OpenAI has confirmed that it confidentially submitted a draft S-1 to the SEC, giving it the option to launch an OpenAI IPO filing between September and November and placing it alongside Anthropic and SpaceX in a once-in-a-generation queue of frontier-tech listings. Collectively worth more than $3.5 trillion in private markets, these firms helped define the AI boom through products like ChatGPT and Claude. Their move toward Wall Street signals a shift from hype-driven private valuations to public scrutiny of revenue quality, profitability timelines, and how sustainable current enterprise AI pricing wars will be.
Inside OpenAI’s unusual IPO path and strategic risks
OpenAI’s draft S-1 highlights both its scale and its structural quirks. The for-profit arm, OpenAI Group PBC, is a public benefit corporation controlled by the OpenAI Foundation, with Microsoft holding a significant minority stake and employees and investors owning the rest. Revenue has reportedly grown 12x in two years to roughly USD 24 billion (approx. RM110.4 billion) annualized, but gross margins are near 33% and the company expects to lose about USD 14 billion (approx. RM64.4 billion) in 2026, with no cash-flow break-even before 2030. According to OpenAI’s own blog, the filing is partly about flexibility: going public sooner if conditions demand it while retaining the option to stay private longer. Forrester warns that despite today’s lead, OpenAI could become “AI’s BlackBerry FIFO,” if it fails to keep enterprise customers from switching to cheaper or more flexible providers.
Anthropic, SpaceX, and the clustering of AI company IPO plans
OpenAI is not alone in heading to public markets. Anthropic reportedly confidentially filed on June 1 with a valuation near USD 965 billion (approx. RM4.44 trillion), while SpaceX is conducting an IPO roadshow around USD 1.75 trillion (approx. RM8.06 trillion). Together with OpenAI’s last reported USD 852 billion (approx. RM3.92 trillion) valuation, this cluster represents more than USD 3.5 trillion (approx. RM16.13 trillion) of private-market value ready to test public investor appetite. The grouping compresses multiple frontier-tech bets—frontier AI labs and a launch provider critical for satellite and space infrastructure—into a single IPO window. For investors, this is a chance to own the perceived core of the AI boom; for the companies, it is a way to fund enormous compute and R&D needs while giving early backers liquidity. Yet the timing also exposes them to the same market questions about sustainability, competition, and enterprise churn.
Enterprise AI pricing wars: DeepSeek vs OpenAI costs
Beneath the headline valuations, enterprise AI pricing wars are intensifying. As models grow larger and agentic workflows expand, token-based billing makes costs unpredictable for coding and knowledge-work tasks. Some enterprises have seen budgets exhausted within months when staff engage in “tokenmaxxing” with long prompts and loops. At the same time, OpenAI and Anthropic have reportedly raised enterprise plan prices while adding limits on total token consumption, pushing customers toward cheaper options. One high-profile example is Microsoft exploring a self-hosted version of DeepSeek’s V4 model for Copilot Cowork, moving away from default reliance on OpenAI and Anthropic. The DeepSeek vs OpenAI costs comparison highlights a broader trend: enterprises prefer metered architectures they can control, with lower per-token prices and the ability to swap models. This erodes incumbents’ pricing power even as they head into public markets seeking premium valuations.
Consolidation, customer churn, and what comes after the IPO rush
AI market consolidation now means a small group of frontier labs and platforms sit atop most enterprise AI spending, yet their grip is less secure than it appears. Forrester urges buyers not to lock into long contracts or single-vendor architectures, warning that the company defining a category can be “most painfully displaced by it.” OpenAI, Anthropic, and their peers must persuade consumers to adopt their agents, convince enterprises to embed them into workflows, and keep pace in the race toward more capable systems. Meanwhile, cheaper models such as DeepSeek’s V4 show that credible alternatives can emerge quickly, especially when open-source foundations lower switching costs. The IPO rush reflects strong investor appetite but masks fragile economics: low margins, high capital needs, and rising churn risks. The next phase of the boom will hinge less on who lists first and more on who offers sustainable, transparent, and fairly priced enterprise AI.
