OpenAI's ChatGPT Focus Sparks Internal Tension, Senior Departures

New reporting from the Financial Times reveals significant internal strain at OpenAI as the company shifts resources toward ChatGPT product development at the expense of other research initiatives. According to eight current and former employees, teams working on the Sora video generation system and DALL-E image model have experienced neglect and underinvestment, prompting several senior researchers to leave the organization.

The strategic shift appears driven by ChatGPT's commercial success, with the chatbot generating estimated $1.6 billion in annual revenue according to recent analyst reports. Meanwhile, OpenAI's research division has seen its budget plateau despite ambitious projects like Sora - which demonstrated text-to-video generation capabilities in early prototypes - requiring substantial computational resources.

Three concrete impacts have emerged from this reallocation:

1. Resource starvation: Sora team members reportedly waited weeks for critical GPU access
2. Talent attrition: At least four senior researchers from multimedia teams departed in Q4 2025
3. Roadmap delays: DALL-E 4 development has been postponed indefinitely

This tension reflects a broader industry pattern where AI labs balance pure research against commercial pressures. OpenAI's recent partnership with Snowflake to embed ChatGPT in enterprise data systems exemplifies the product-first approach that's allegedly causing internal friction. The company's compute allocation strategy now prioritizes ChatGPT inference workloads (serving existing customers) over research experimentation (developing new capabilities).

Technical staff expressed particular concern about falling behind competitors in multimodal AI. While OpenAI demonstrated early leadership with DALL-E (2021) and Sora (2024), rivals like Midjourney and Runway ML have released multiple iterative improvements to their image and video systems in the past 18 months.

The reported changes come amid growing pressure to demonstrate sustainable revenue following Microsoft's additional $10 billion investment in 2025. OpenAI's enterprise ChatGPT business now accounts for nearly 80% of revenue according to analyst estimates, creating strong incentives to concentrate resources on maintaining that growth.

Critical unanswered questions remain:

• Whether this represents a permanent shift or temporary reallocation
• How the research/product balance might change post-IPO
• If departing talent is migrating to competing labs

OpenAI declined to comment on the specific allegations but reiterated its commitment to both applied and theoretical AI development. The situation highlights the difficult tradeoffs facing AI labs as they transition from research organizations to commercial entities - a tension likely to intensify as investor expectations collide with scientific ambitions.