Oracle and CoreWeave defended OpenAI after reports that ChatGPT missed user and revenue goals, but investors are still selling stocks tied to its infrastructure development. The market is now asking a tough question: if OpenAI is headed for an IPO at anything close to a $1 trillion valuation, can ChatGPT's subscriptions and business sales grow fast enough to cover the costs of running the models?
OpenAI is still private, but the sale of partners has given the public markets a way to price the doubt. Reuters reported that Oracle, CoreWeave and other AI-related stocks fell after a Wall Street Journal report said OpenAI missed revenue and user goals, with Oracle down 3.4% and CoreWeave down 2.8%. The decline wasn't just one company missing internal numbers. It was about whether the development of the AI infrastructure is funded before the revenue that can take a long time to arrive.
Standardization makes the issue difficult to reverse. OpenAI said on March 31, 2026 that it had closed a funding round with $122 billion in committed capital at a post-cash value of $852 billion. Reports also said the company was preparing for an IPO that could be worth up to $1 trillion, although any final timeline or valuation remains uncertain. (openai.com)
That's a huge number of business models being tested in real time. ChatGPT is recognized worldwide, but social market measurement requires more than a popular brand. It needs proof that user demand can be sustainable revenue, that paying customers stay, and that margins survive the cost of chips, cloud capacity, data centers, power and model development.
The reported user target goes directly to the subscription model. Forbes reported that OpenAI missed an internal goal of reaching 1 billion active ChatGPT users by the end of 2025 and also fell short of several monthly revenue goals. If those reports are accurate, the warning isn't that OpenAI doesn't scale. The caveat is that usage may not turn into paychecks quickly enough for the valuation attached to the company.
AI is not ordinary software. A traditional software platform can often add customers without a significant increase in costs. A heavy user of AI incurs costs every time they ask a model to write code, analyze files, create media, run agents or handle work tasks. The better the product, the more people use it – and the infrastructure OpenAI needs behind it.
That makes ChatGPT subscriptions harder to track than it first appeared. A fixed monthly plan provides OpenAI with recurring revenue, but usage varies greatly. A regular subscriber may be more profitable. A power user can use much more of the computer than the subscription price suggests. If OpenAI continues to add more capable tools without raising prices enough, revenue could grow while margins remain under pressure.
There is also a churn problem. Some users pay for ChatGPT when they have a lot of work, then cancel, downgrade or switch when another model develops. Google Gemini, Anthropic Claude and Microsoft Copilot are all putting pressure on OpenAI's pricing power. In the IPO, investors will want to know whether ChatGPT behaves like a sticky software revenue or product users flow in and out of it.
Corporate income can support greater valuations, but it brings a different cost structure. Enterprise customers want security, administrative controls, data management, purchasing policies, integration and proof that AI saves time or money. Those contracts can be larger and more stable than consumer subscriptions, but they require support, product depth, compliance guarantees and negotiated prices.
This is where Oracle and CoreWeave become part of the same measurement story. Oracle sells cloud capacity. CoreWeave sells AI infrastructure. Nvidia and AMD sell chips. If OpenAI's revenue continues to rise, those providers are sitting on the back of a strong growth cycle. If OpenAI's targets falter, investors are starting to question whether some of the data center and cloud spending is built on demand that may arrive later, or on weaker margins, than expected.
CoreWeave's response captured that tension. It defended OpenAI while insisting that it serves a broad customer base. That is the message infrastructure providers need to send. OpenAI is important, but no supplier wants to appear too dependent on a single independent company's ability to hit aggressive targets.
The question of margin is becoming increasingly difficult to avoid. If OpenAI pays more for cloud capacity, chips and power, then a larger share of the value of AI can flow to suppliers. If OpenAI can charge more for registration, business tools, developer access and agent workflow, it saves a lot of economics. The case for an IPO depends on which side captures the margin.
Private funding cycles can depend on scarcity, strategic prospects and long-term belief. Public markets have little patience. They will ask about revenue quality, churn, business retention, gross margin, computer debt, cloud commitments and capital requirements. The private valuation of $852 billion can be trusted for a while. A listed company must report the numbers.
OpenAI's fundraising gives it great resources, but it also raises the standard of proof. At this rate, ChatGPT can't just be popular. It should be a revenue engine big enough to support the most expensive infrastructure races in business history.
The question of the IPO is not whether OpenAI is important. Whether public investors will pay nearly a billion dollars before they see enough evidence that the revenue model works at scale. The answer depends on how quickly OpenAI can move from users to paying customers, from consumer subscriptions to business contracts, and from high usage to heavy margins in computing.
The risk for Oracle, CoreWeave, Nvidia and AMD is not that OpenAI stops growing. Whether growth is less predictable, more expensive to provide, or slower to monetize. In today's AI standards, even the slightest doubt about the speed of revenue can send billions to the companies that provide the design.
OpenAI can still access public markets in large numbers. But the target reported to have been exceeded has changed the debate. The company now has to show that ChatGPT is not just a successful product. It must demonstrate that it can turn high consumption, stiff competition and high demand for infrastructure into long-term capital.
That's a test that investors will keep coming back to. If ChatGPT's revenue can exceed the computer bill, OpenAI's valuation issue remains. If it can't, the pressure won't stop with OpenAI. It will spread through the AI stocks created by its growth.
More from Finance Monthly: Meta Cuts 8,000 Jobs to Pay for AI
