OpenAI Missed Internal Growth Targets for ChatGPT Users and Revenue

OpenAI failed to meet internal growth targets for ChatGPT user base and revenue for the first time in a long while. According to Wall Street Journal reporting cited by Bloomberg and Reuters, the company did not achieve its own targets for new user acquisition and sales, and ChatGPT failed to reach 1 billion weekly active users by the end of 2025. Within OpenAI, this triggered not just frustration over a missed KPI, but a more serious question — is commercial growth keeping pace with the scale of infrastructure the company is already building and contracting for years ahead.

Here's a crucial detail: this is not about declining demand in absolute terms. According to OpenAI's own data as of March 31, 2026, ChatGPT already has over 900 million weekly active users and more than 50 million subscribers. In December, the company also reported that over 1 million corporate clients worldwide use its business products directly. In other words, the scale remains enormous and stands out as exceptional compared to most technology companies. But when you set an internal goal of 1 billion weekly active users by a specific date, the gap between "very large audience" and "target audience per plan" becomes financially significant, because hiring, sales, marketing, and most importantly, computing power are pre-planned based on these expectations.

According to WSJ, OpenAI fell short on several monthly revenue targets in 2026. The publication links this, among other factors, to Anthropic strengthening its position in programming and the enterprise segment, and the growing popularity of Google Gemini complicating subscriber retention. This is an important shift. Until recently, OpenAI could grow almost automatically on the back of general interest in generative AI, but now the market has become noticeably more competitive and segmented. User audience no longer guarantees dominance in the most profitable categories.

Corporate clients compare not just model quality, but total cost of ownership, reliability, integrations, data control, agent capabilities, and ease of deployment into real workflows. Against this backdrop, infrastructure costs appear particularly sensitive. According to Reuters, over the past year OpenAI has committed to approximately $600 billion in data center construction over the coming years. Wall Street Journal also reports that CFO Sarah Friar expressed concern in conversations with leadership: if revenue doesn't grow fast enough, the company may struggle to pay future contracts for computing power.

There are also reports that the board is scrutinizing recent data center deals more carefully and asking questions about further scaling of computing resources. For a business dependent on increasingly expensive clusters, chips, energy, and long-term infrastructure contracts, even moderate deviation from internal forecasts quickly becomes a strategic problem.

Meanwhile, the picture still looks very strong from the outside. At the end of March, OpenAI announced raising $122 billion in new capital, and in early April announced that its enterprise business already accounts for over 40% of its revenue and could match its consumer business by the end of 2026. In other words, the company doesn't look weak in the classical sense. Rather, this is about a conflict between two speeds: on one hand, near-unprecedented product growth and client base expansion; on the other, an even more aggressive pace of infrastructure commitments, requiring flawless execution of commercial plans.

At these scales, even a modest shortfall in revenue or user base stops being cosmetic and starts affecting the investment narrative, partnerships negotiations, and preparation for potential IPO. The main conclusion is simple: leadership in AI is no longer measured solely by user count, demo quality, or the loudness of new releases. Now it's critical for OpenAI to prove that it can simultaneously retain audience, win corporate budgets, and convert the colossal demand for AI into cash flow that justifies data center costs.

If this works out, the current miss will remain an episode against the backdrop of overall growth. If not, the market will begin viewing every new round, computing power contract, and model launch through the same lens: is the business model keeping pace with infrastructure ambitions?