Silicon Valley is witnessing a new and decisive phase in the fierce battle for rare talent, as news reports indicate that three senior executives who played a pivotal role in establishing OpenAI’s “Stargate” data center initiative are making the final move to Meta. This advanced strategic step reflects the scale of intense competition among major tech companies to control the AI infrastructure sector, which has become the true cornerstone in shaping the future of global technology. With the increasing reliance on large language models, the need has become extremely urgent to attract engineering minds capable of designing and managing highly complex computing systems.
Details of prominent engineering leadership departures
Sources familiar with the matter revealed that Peter Hoeschele, who was instrumental and central in launching the ambitious “Stargate” project, is preparing to join Meta in a move that has surprised many observers. He is not alone; he is accompanied by Shamiz Hemani, who handled computing strategy and business development, and Anup Saharan, another prominent leader from the supercomputing division. These leaks confirm recent reports by some tech publications about the departure of this trio from OpenAI, representing a major and significant loss for the infrastructure development team at the company that launched the generative AI revolution, leaving it facing new challenges to rebuild its leadership structure.
Comprehensive restructuring and a change in course
These significant departures coincide with a series of sweeping administrative changes within OpenAI’s infrastructure department. Last March, the company appointed former Intel executive Sachin Kati to oversee the computing clusters for the Stargate project. This appointment came after the company shifted its strategic course, preferring to lease more AI servers from major cloud service providers rather than bearing the exorbitant costs and grueling engineering effort of building all its data centers from scratch. Under Kati’s leadership, the company reorganized its infrastructure team into a completely new three-tiered management structure in a flexible attempt to adapt to the rapid and changing demands of the tech market.
Accelerating the pace of infrastructure investments
This intensive and clever talent acquisition aligns with Meta’s aggressive push to expand its own artificial intelligence infrastructure. In January, the company projected its 2026 capital expenditures to be between $115 billion and $135 billion, roughly double the $72 billion it spent in 2025, according to its fourth-quarter earnings report data. CEO Mark Zuckerberg described 2026 as a pivotal and historic year for AI, clearly indicating that these massive investments are primarily aimed at building what he called “personal superintelligence.” The company is currently building several large-scale data centers across the United States and is evaluating potential expansion at the Abilene, Texas site—the same strategic location where both OpenAI and Oracle backed away from expansion plans in March. Alongside infrastructure engineers, the company has separately hired several distinguished AI researchers to strengthen its development team.
Challenges of mega projects and shifting alliances
The “Stargate” project was officially announced at the White House in January 2025, with financial and technical backing from major entities including SoftBank, OpenAI, Oracle, and MGX. The project was promoted as a historic and unprecedented $500 billion initiative aimed at building highly advanced AI infrastructure across the United States. However, this ambitious initiative faced successive setbacks and obstacles, most notably the cancellation of the planned expansion for the main Abilene campus, which aimed to increase capacity from 1.2 gigawatts to about 2 gigawatts of power. This sudden cancellation came after negotiations among the founding companies collapsed due to deep disagreements over funding and changing capacity expectations. Following this, Microsoft stepped in to salvage the situation by building two new data center buildings at the same Abilene site through a strategic partnership with developer Crusoe.
The future of the industry amid fierce competition
The brain drain and movement of talent from one organization to another reflect a state of widespread repositioning in the advanced technology industry. Companies with massive financial resources are fiercely competing to secure the engineering expertise needed to build unprecedented computing capabilities. These companies are well aware that the real and future battle is not limited to developing software algorithms, but extends to the actual ability to run these algorithms with high efficiency, which requires formidable infrastructure and minds capable of managing and developing it continuously. As this desperate race for talent continues, leading organizations will be forced to offer exceptional incentives and alluring work environments to retain their engineers, while major entities will continue to seize every possible opportunity to consolidate their absolute dominance over the global tech market, paving the way for qualitative leaps in the speed and efficiency of future data processing.
Frequently Asked Questions
Who are the executives who moved to Meta?
The list of transferees included Peter Hoeschele, Shamiz Hemani, and Anup Saharan, who are among the most prominent leaders and founders of the data center project at their former company.
What is the scale of Meta’s infrastructure investments?
The company expects its capital expenditures to reach between $115 billion and $135 billion in 2026 to enhance AI computing capabilities.
What are the reasons behind the recent setbacks of the Stargate project?
The project faced notable setbacks due to disagreements over funding and capacity expectations, leading to the cancellation of major expansion plans at the Abilene, Texas site.
