OpenAI Restructures Infrastructure Division: Uday Ruddarraju Takes the Reins as Sachin Katti Transitions to Advisory Role

SAN FRANCISCO — In a major leadership shakeup highlighting the fierce competition for elite artificial intelligence infrastructure talent, OpenAI has consolidated its computing and infrastructure divisions under Uday Ruddarraju. A seasoned engineering leader who previously spearheaded infrastructure at Elon Musk’s xAI, Ruddarraju assumes total control over the artificial intelligence pioneer’s sprawling hardware and data center strategy.

The structural pivot, first confirmed by Bloomberg on Wednesday, sees Sachin Katti stepping down from his day-to-day oversight of the infrastructure team after a high-intensity tenure defined by multi-billion-dollar expansions. While Katti will transition out of his active management role in the coming weeks, he will remain with OpenAI in an advisory capacity. According to company insiders, the executive leadership framework directly beneath the top tier will remain stable, with middle management and operational leads maintaining their positions to ensure continuity across ongoing projects.

The consolidation marks a pivotal moment for OpenAI as the company pivots from pure conceptual research to the industrial-scale deployment of next-generation physical infrastructure. With hundreds of billions of dollars hanging in the balance, the ability to secure power, chips, and land has become just as critical to the AI race as algorithmic breakthroughs.

Main Facts

The core of OpenAI’s infrastructure leadership is now unified under a single executive banner. Uday Ruddarraju, who joined OpenAI in July 2025 as chief technology officer for compute, now oversees both compute procurement and data center engineering operations.

Sachin Katti, who arrived at OpenAI late last year from Intel—where he served as chief technology and AI officer—is relinquishing his comprehensive infrastructure portfolio. Katti’s tenure was characterized by massive capital deployment, including the oversight of OpenAI’s massive $30 billion data center build in Georgia, the company’s first fully self-designed and managed site. Furthermore, Katti managed the sprawling 8-gigawatt infrastructure project in Ohio, developed in partnership with SB Energy and Nvidia. Regulatory filings later revealed that Nvidia had guaranteed up to $105 billion in lease and power payments for the Ohio initiative.

Concurrently, Katti’s former oversight of "GPT-Infra"—an internal initiative utilizing OpenAI’s own models to accelerate semiconductor design—has been reassigned to Richard Ho, the company’s chief chip architect. This transition comes amid public skepticism from Ho, who remarked earlier this month that he does not yet fully trust automated AI models to independently execute critical chip design changes.

Ruddarraju inherits an infrastructure footprint that is almost entirely concentrated within the United States. International ambitions have hit significant roadblocks; notably, OpenAI paused its "Stargate UK" initiative in April due to prohibitive British electricity prices—which run more than four times higher than US averages—coupled with unresolved regulatory and copyright frameworks. Consequently, the vast majority of OpenAI’s projected $856 billion compute bill shown to investors will be deployed domestically across massive American facilities.

Chronology of Infrastructure Leadership

The restructuring is the culmination of a turbulent eighteen months of executive comings and goings within OpenAI’s infrastructure and hardware divisions, driven by the unprecedented demands of scaling frontier AI models.

• Late 2024: Sachin Katti joins OpenAI from Intel, bringing deep networking and AI hardware expertise to oversee the company’s burgeoning data center ambitions.
• March 2025: Katti formally assumes leadership over the Stargate infrastructure initiative, steering long-term hardware strategy and site acquisitions.
• July 2025: Uday Ruddarraju joins OpenAI, subsequently taking on the role of chief technology officer for compute. Around the same time, Katti takes charge of GPT-Infra, applying OpenAI models to internal hardware design tasks.
• July 2025: OpenAI formally announces a 230-megawatt European data center project in Norway—its sole publicly confirmed continental European site.
• April 2025: OpenAI officially pauses its Stargate UK project, citing structural economic hurdles, with The Guardian later reporting that the company never physically visited the proposed site in northeast England.
• Summer 2025: A wave of departures hits the infrastructure team. Chris Malone, OpenAI’s head of data centers, departs the company, prompting a quiet executive search. Concurrently, key Stargate architect Peter Hoeschele leaves for Meta alongside colleagues Keith Heyde and Shamez Hemani.
• August 2025: When confirming Chris Malone’s departure, OpenAI quietly acknowledges it has already filled the compute CTO role, though leadership refuses to publicly name Ruddarraju at the time.
• October 2025: Richard Ho publicly expresses reservations regarding AI-driven chip design changes under the GPT-Infra framework.
• October 2026: Bloomberg reports that OpenAI is officially consolidating compute and infrastructure under Uday Ruddarraju, with Sachin Katti stepping down from operational duties to serve as an advisor.

Supporting Data and Financial Footprint

The sheer scale of OpenAI’s infrastructure commitments dwarfs traditional technology capital expenditures, demanding a level of coordination that necessitates a unified leadership structure.

The financial exposure managed by Ruddarraju is staggering. Having previously cut his teeth at Elon Musk’s xAI—where he directly managed the massive compute clusters required to train the Grok 3 model across a sprawling fleet of 100,000 enterprise GPUs—Ruddarraju brings battle-tested operational experience to OpenAI’s hardware ambitions.

Official Responses and Executive Perspectives

The transition has been characterized by public diplomacy from both outgoing and incoming leaders, emphasizing collaboration and strategic realignment rather than internal friction.

Reflecting on his new dual-hatted role, Uday Ruddarraju praised his predecessor in statements provided to Bloomberg:

"Sachin has done an incredible job maintaining the commercial partnerships."

Ruddarraju noted that he anticipates a seamless handover process, adding that the consolidation will allow him to dedicate increased executive bandwidth to the commercial monetization of OpenAI’s computing assets.

For his part, Sachin Katti expressed optimism regarding the future of the organization under Ruddarraju’s unified command, signaling his readiness for new professional challenges.

"I’m proud of what we’ve built and have total confidence in Uday and the team," Katti stated, noting that he is eager to explore new engineering frontiers following his advisory stint.

Meanwhile, internal friction remains visible in adjacent hardware units. Richard Ho, overseeing the specialized chip design initiatives inherited from the restructuring, has maintained a pragmatic, cautious stance on the integration of generative AI into silicon engineering. Ho’s public admission that he does not yet fully trust automated models to execute silicon design modifications highlights the friction points between software-driven development and the unforgiving physical tolerances of semiconductor manufacturing.

Strategic Implications for the AI Industry

The decision to merge compute and infrastructure under a single leader carries profound implications for the broader artificial intelligence landscape.

1. Elimination of Silos

Historically, tech companies often split infrastructure (power, real estate, cooling, and networking) from compute (chip procurement, cluster architecture, and topology optimization). By placing both domains under Uday Ruddarraju, OpenAI is eliminating internal friction. Building a multi-gigawatt facility requires synchronizing silicon delivery schedules directly with electrical grid activation and structural engineering; a unified chain of command prevents costly delays.

2. The Talent War and the xAI Connection

Recruiting Ruddarraju from xAI—and subsequently elevating him over internal figures—signals OpenAI’s aggressive posture in the market for elite infrastructure talent. As frontier labs push past the 100,000-GPU threshold toward millions of interconnected processors, the individuals who know how to keep massive, fault-prone clusters running continuously are among the most valuable assets in Silicon Valley.

3. Domestic Focus and International Realities

The effective abandonment—or indefinite postponement—of major international projects like Stargate UK underscores a hard economic truth: the energy and regulatory grid in the United States, particularly in pro-development states like Ohio and Georgia, currently offers a more viable runway for hyperscale AI expansion than Western Europe. With British energy costs penalizing industrial operations, OpenAI’s infrastructure destiny is firmly anchored to American soil.

As Ruddarraju steps into his expanded role, the pressure will be immediate. With hundreds of billions of dollars committed to utility providers, chipmakers, and real estate developers, OpenAI must prove that its physical infrastructure can scale rapidly enough to support the next generation of artificial general intelligence.