Indian-origin Ruddarraju takes over as compute CTO at OpenAI

OpenAI has promoted Uday Ruddarraju to chief technology officer (CTO) of compute, placing the Indian-origin infrastructure executive in charge of one of the company’s most consequential and capital-intensive operations.

Ruddarraju announced the appointment in a LinkedIn post, roughly a year after joining OpenAI as head of compute and infrastructure from Elon Musk’s xAI.

His LinkedIn profile now lists him as CTO of Compute from July 2026. OpenAI has not issued a separate announcement about the promotion.

The appointment reflects how access to computing capacity has become as important to frontier AI companies as model research and software development. OpenAI is expanding beyond buying cloud services and chips into the construction and operation of an infrastructure network involving data centers, power systems, networking, storage and custom hardware.

“My first 12 months at OpenAI have been incredibly rewarding,” Ruddarraju wrote. He said the compute team had worked to bring capacity online while developing the systems across computing, networks, storage and machine learning needed to train frontier models such as GPT-5.6.

Ruddarraju said OpenAI intends to build “the world’s largest compute footprint” and is hiring across distributed systems, hardware, manufacturing and data-center construction, including civil, mechanical and electrical engineering.

The breadth of that hiring points to a change in the nature of AI development. Building increasingly capable models no longer depends mainly on researchers designing better algorithms. It also requires securing chips, electricity, land, cooling systems and high-speed networks, then getting thousands of computing units to operate reliably as one system.

Compute becomes the constraint

Ruddarraju takes the role as OpenAI accelerates Stargate, its long-term infrastructure program.

OpenAI said in April that it had surpassed its original target of securing 10 gigawatts of US AI infrastructure by 2029, after adding more than 3 gigawatts during the preceding 90 days. The company is evaluating further locations as it plans beyond that initial target.

That expansion builds on projects announced with Oracle, SoftBank and other partners. In September 2025, OpenAI said five additional US sites, together with its Abilene campus and projects involving CoreWeave, represented nearly 7 gigawatts of planned capacity and more than $400 billion of investment over three years.

OpenAI and its partners have since continued to add projects. In June, the company broke ground on a 1-gigawatt data-center campus in Saline, Michigan, developed with Oracle, Related Digital and Walbridge. The facility is expected to create more than 2,500 union construction jobs and 450 permanent jobs on site, according to OpenAI.

The challenge facing Ruddarraju will therefore extend far beyond operating servers. OpenAI must coordinate cloud providers, chipmakers, energy companies, construction groups and local authorities while managing rapid changes in hardware and model design.

Large AI clusters must also remain reliable despite failures in individual chips, network links or storage systems. A disruption during a large training run can waste scarce computing capacity and raise development costs. Bringing facilities online late can similarly delay model development or constrain the number of customers OpenAI can serve.

OpenAI describes compute as the center of its commercial and technological cycle. More capacity allows it to train stronger models and serve more users, while increased usage generates revenue that can be reinvested in infrastructure.

Before joining OpenAI, Ruddarraju served as head of infrastructure engineering at xAI, where he worked on the systems behind the company’s Colossus supercomputer. His earlier career included engineering roles at eBay and Robinhood, according to his professional profile and published accounts of his career.

His move from xAI to OpenAI in July 2025 came amid intense competition among AI companies for engineers capable of designing and operating large-scale computing systems.

That competition is likely to intensify as frontier-model developers pursue larger training runs and expand the use of AI in consumer and enterprise applications.