OpenAI wants to go public. Wall Street isn’t sure it wants OpenAI.
The company behind ChatGPT has been signaling IPO intentions for months, but a growing chorus of institutional investors is raising hard questions about valuation, profitability, and the sustainability of a business model that burns cash at an extraordinary rate. According to The Information, OpenAI’s IPO hopes face a skeptical investor community that isn’t buying the hype at the prices being floated.
The skepticism is warranted.
The Numbers Don’t Add Up — Yet
OpenAI reportedly hit an annualized revenue run rate of $12.7 billion as of early 2025, according to reporting from Bloomberg. That’s impressive growth. But revenue isn’t profit. The company’s costs are staggering — compute infrastructure alone runs into billions annually, and OpenAI has acknowledged it expects to lose money for years. In its last known fundraising round in late 2024, the company raised $6.6 billion at a $157 billion valuation, as reported by Reuters. Some internal projections now reportedly target a valuation north of $300 billion for a public offering.
That’s a lot to ask from public market investors who’ve watched the AI trade cool considerably since early 2024. The Nasdaq’s AI darlings have come back to earth. Investors who got burned on speculative tech listings — remember WeWork, remember Rivian’s post-IPO collapse — are asking tougher questions now.
Here’s the core tension: OpenAI’s valuation assumes it will dominate an industry that doesn’t yet have clear unit economics. Training frontier AI models costs hundreds of millions per run. Inference — actually serving responses to users — isn’t cheap either. And competition is intensifying from Google DeepMind, Anthropic, Meta’s Llama models, and open-source alternatives that are closing the performance gap fast.
According to The Information, some prospective investors have pushed back on the idea that OpenAI can sustain its growth trajectory without margins improving dramatically. The company’s enterprise business is growing, but so is customer concentration risk — a handful of large clients reportedly account for a disproportionate share of revenue.
So where does the margin expansion come from? OpenAI bulls point to subscription revenue from ChatGPT Plus and the new Pro tier, API usage fees, and enterprise contracts. But inference costs scale with usage. More customers mean more compute. And OpenAI doesn’t own its own infrastructure — it relies heavily on Microsoft Azure, a relationship that comes with both benefits and dependencies.
The Microsoft Question
Microsoft’s $13 billion investment in OpenAI gives it a 49% stake in the company’s profits, access to its models, and enormous influence over its direction. For public market investors, this creates a structural concern. How do you value a company whose most important partner is also its landlord, its largest investor, and increasingly a competitor? Microsoft has been building its own AI capabilities, integrating Copilot across its product line, and has the resources to develop or acquire alternatives if the relationship sours.
OpenAI’s ongoing conversion from a nonprofit to a for-profit entity adds another layer of complexity. The New York Times reported that the restructuring has drawn legal challenges and scrutiny from state attorneys general. Until that conversion is finalized, the corporate structure remains unusual — not the kind of clean story public market investors typically favor.
And then there’s the competitive picture. Anthropic raised $2 billion from Amazon in early 2025. Google is pouring resources into Gemini. Meta gives away its models for free. xAI, Elon Musk’s startup, raised $6 billion and is building massive compute clusters. The moat around OpenAI’s technology is narrower than it was eighteen months ago. Benchmarks show competitors matching or exceeding GPT-4 class performance on key tasks.
None of this means OpenAI is a bad company. It’s a technically impressive organization with genuine product-market fit — ChatGPT has over 200 million weekly active users, per the company’s own disclosures. But product-market fit and IPO-readiness are different things.
The broader AI market is also sending mixed signals. Enterprise adoption is real but slower than the hype suggested. A McKinsey survey found that while 65% of organizations report using generative AI regularly, many are still in pilot phases, struggling with integration, accuracy, and ROI measurement. The killer enterprise use case — the one that justifies $300 billion valuations — hasn’t materialized yet for most companies.
Private market investors have different risk tolerances than public ones. Venture funds and sovereign wealth funds that participated in OpenAI’s last round can afford to wait a decade for returns. Mutual funds, pension managers, and retail investors buying into an IPO cannot. They need a credible path to profitability on a reasonable timeline, and OpenAI hasn’t publicly provided one.
Sam Altman has said the company plans to become profitable eventually. Eventually is not a financial projection.
There’s also the question of timing. IPO windows open and close. The current macroeconomic environment — persistent inflation concerns, trade policy uncertainty, and volatile equity markets — doesn’t scream favorable conditions for a massive tech offering. The Financial Times has noted that the 2025 IPO pipeline remains thin compared to historical norms, with several high-profile companies delaying their listings.
OpenAI may ultimately go public and trade well. It has brand recognition, a massive user base, and genuine technological capability. But the investor community’s skepticism isn’t irrational. It reflects hard-learned lessons from previous cycles where transformative technology and sustainable business models turned out to be very different things.
The smart money is waiting. Not because AI isn’t real. Because the price has to be.
WebProNews is an iEntry Publication