Uday Ruddarraju steps into CTO, Compute role at OpenAI

Uday Ruddarraju announced that he is moving into the CTO, Compute role at OpenAI. The title matters: this is a specialized infrastructure leadership position, not a claim that he has become OpenAI's companywide chief technology officer.

Ruddarraju said his first 12 months at OpenAI involved bringing new capacity online and building systems across compute, networking, storage and machine learning. Current reporting from Financial Express and Times of India follows his first-party announcement; OpenAI has not published a separate corporate newsroom release for the appointment.

Ruddarraju described a compute roadmap spanning both software and physical infrastructure:

• •large-scale distributed systems and production reliability;
• •networking and storage for frontier-model training;
• •hardware and manufacturing;
• •data-center construction across civil, mechanical and electrical engineering; and
• •recruitment across the infrastructure stack.

He linked that work to systems used to train frontier models such as GPT-5.6 and said the company intends to build what he called the world's largest compute footprint. Those are forward-looking company statements, not independently measured capacity results.

At frontier-model scale, compute is no longer just a procurement line item. Training and serving models depends on how quickly an organization can secure power and chips, connect clusters, move data, recover from hardware failures and keep utilization high without sacrificing reliability. Giving this work a dedicated CTO title signals that OpenAI is treating infrastructure design as an executive engineering function.

The appointment also separates two related responsibilities. Ruddarraju's announcement focuses on building and operating the compute stack. Sachin Katti separately announced a role covering compute strategy and GPT infrastructure, including capacity access, economics and investment returns. That division suggests OpenAI is formalizing leadership around both the engineering and commercial sides of its infrastructure expansion.

The immediate news is the leadership change, not proof that OpenAI has already delivered the scale described in the post. The most useful signals to watch are measurable ones: new capacity brought online, reliability improvements, custom-hardware milestones, data-center delivery dates and changes in training or inference economics. Until those results are disclosed, the roadmap should be read as an ambition attached to a newly elevated role.