OpenAI is losing research staff as it doubles down on enterprise ChatGPT sales at the expense of longer-term AI research projects, according to people familiar with the matter speaking to the Financial Times.
The departures follow CEO Sam Altman's recent acknowledgment that enterprise sales are a 2026 priority to counter Anthropic's enterprise momentum. Despite ChatGPT approaching 900 million weekly users, OpenAI lags in enterprise AI sales where competitors have gained ground.
The enterprise numbers tell a different story
OpenAI's enterprise push is already showing results. ChatGPT enterprise usage increased 8x in 2025, with structured workflows up 19x and reasoning token consumption jumping 320x. The company claims 92% of Fortune 500 companies now use OpenAI technology in some form. Heavy AI users report saving more than 10 hours per week.
The company offers ChatGPT Teams, Business, and Enterprise tiers with shared projects and group memory features designed for organizational deployments.
Research continues, but the priority shift is clear
OpenAI maintains it hasn't abandoned research entirely. The company forecasts small AI discoveries in 2026 and is expanding into the EU via OpenAI for Europe. A recent Microsoft deal provides flexibility for the enterprise push while supporting $1.4 trillion compute infrastructure plans.
What this means in practice: OpenAI is betting enterprise revenue will fund future research rounds. The company is solving what Altman calls "application problems" rather than pure research challenges.
The timing is notable. This comes as research teams increasingly evaluate self-hosted open source alternatives like Llama and Mistral for sensitive data and budget constraints. Academic institutions and enterprise research teams are weighing API costs against locally deployed models, particularly for domain-specific fine-tuning.
History suggests prioritizing commercial products over research eventually constrains innovation. OpenAI's challenge: maintaining its technical edge while chasing enterprise contracts that competitors are already winning.