The Wall Street Journal reported on April 28 that OpenAI missed several monthly revenue targets in early 2026 and failed to reach its internal goal of one billion weekly active users — and that the shortfall has put the company’s CFO and CEO at odds over a spending strategy that has already committed roughly $600 billion to future compute infrastructure.

The story broke as OpenAI’s competitor Anthropic simultaneously reported that its Claude models now hold approximately 40% of enterprise LLM API spending, up from a smaller share just eighteen months ago. OpenAI, which once dominated that market with a 50% share, has slipped to around 27% according to multiple analyst estimates circulating in late April.

What the WSJ Actually Said

The Journal’s core claim is that OpenAI missed multiple monthly revenue targets in Q1 2026 after growth for ChatGPT slowed toward the end of 2025. The company also fell short of a self-imposed milestone to reach one billion weekly active ChatGPT users.

The most pointed detail in the report: CFO Sarah Friar has privately expressed concern that the company’s revenue trajectory may not support the scale of compute contracts it has entered into. OpenAI has committed itself to $250 billion in Azure spending through 2032, and has been making even broader infrastructure commitments alongside the Stargate project. Those bets only make financial sense if revenue roughly doubles each year.

Friar and CEO Sam Altman issued a joint statement describing the WSJ characterisation as “ridiculous” and pointed to what they called “breakout Codex growth” in the enterprise coding market. But the fact that two senior executives felt they needed to publicly respond to a revenue report suggests the story landed harder than a standard media correction would warrant.

OpenAI has, by any conventional standard, grown at a historic pace. The company raised $122 billion in its most recent funding round at an $852 billion valuation. It is planning a public listing in 2026. The revenue miss is relative — a gap between the company’s own aggressive internal targets and what it actually delivered, not a sign of a failing business.

Why Anthropic Is Winning Enterprise

The more strategically significant part of this story is not OpenAI’s shortfall — it is why Anthropic is gaining ground.

Enterprise AI buyers in 2025 and 2026 evaluated vendors on three things: model capability, API reliability, and safety posture. Anthropic scored well on all three in workloads that matter most to large organisations. Claude’s Constitutional AI approach made it easier for procurement and legal teams to approve deployments. Its coding benchmark performance won over developer-heavy organisations. And its enterprise contract terms — particularly around data handling and model availability guarantees — gave risk-averse buyers more confidence than some alternatives.

The result: Anthropic now counts more than 1,000 enterprise customers each spending over $1 million per year, a figure that reportedly doubled between February and April 2026 alone.

Google’s decision to invest up to $40 billion in Anthropic — announced just days before the WSJ revenue story broke — also changes the infrastructure equation. Anthropic’s historical constraint has been compute capacity. With Google’s cloud backing and TPU access now locked in, the reliability argument that OpenAI once used as a wedge against Anthropic has weakened considerably.

The Broader Market Shift

This story is part of a structural change in how enterprise AI spending is being allocated. In 2024, most enterprise AI budgets defaulted to OpenAI — largely because IT leaders knew the brand, executives had approved it, and risk-averse procurement teams defaulted to the market leader.

That default is eroding. Enterprise buyers now have more options, more internal expertise to evaluate them, and more tolerance for deploying non-OpenAI models in production. Anthropic, Google DeepMind, and Mistral have all moved from “alternative” to “serious contender” status with enterprise procurement teams.

The competitive pressure benefits buyers. When two well-funded vendors are competing aggressively for enterprise contracts in the lead-up to dual IPOs, pricing flexibility increases, SLA commitments improve, and support quality rises. A company locked into a single AI vendor in 2023 probably got worse terms than a company going through competitive evaluation in 2026.

What This Means for Business

If you are mid-evaluation: This is a good moment to ask both OpenAI and Anthropic for their most current enterprise SLA commitments, pricing tiers, and data handling agreements. Both vendors are motivated to close enterprise contracts ahead of IPO. Leverage that.

If you are already on OpenAI: The revenue miss does not mean OpenAI’s products have degraded. GPT-5.x and Codex remain strong tools. What it does mean is that OpenAI’s internal growth anxiety may translate into more aggressive enterprise sales outreach — and potentially better terms if you are in a renewal cycle.

If you are running production AI workloads: The CFO-compute spending tension at OpenAI is worth monitoring. If the company’s revenue trajectory continues to miss internal targets while its infrastructure commitments remain locked in, there is some risk of pricing changes, capacity reallocation, or enterprise tier restructuring. Diversifying across at least two model providers is no longer just a technical best practice — it is a business continuity question.

If you have not started yet: The market is more competitive than it has ever been, which means getting into AI automation right now comes at better pricing and with better vendor support than it did eighteen months ago. Waiting for “the dust to settle” means waiting through the best buyer’s market the enterprise AI sector has seen.

The headline of the WSJ story is about OpenAI stumbling. The more useful frame for business leaders is that the AI platform market just became more competitive, more mature, and more negotiable — and that is a good thing for anyone building seriously on it.

---

If you want the playbook other teams are using with Claude and Codex right now, grab the free Working With Claude field guide. Download it here.