SpaceX has moved to reset the leadership of its artificial intelligence data center division after civil-engineering concerns and consistent reliability issues emerged across its Tennessee and Mississippi facilities. The company pulled seasoned executives from its rocket and satellite internet programs and dropped them into the AI infrastructure unit, a move that comes as Elon Musk has expressed frustration with the performance of the company’s AI infrastructure and as reliability concerns have mounted across its existing facilities.

The Information first reported the reorganization on Tuesday, citing people familiar with internal discussions at the company. Facilities in Tennessee and Mississippi sit at the center of the trouble, having run for extended stretches without the backup cooling and power redundancy that data centers of this scale typically require. That gap, according to the report, increased the risk of outages at facilities supporting SpaceXAI’s AI workloads. The personnel shift underscores the growing operational challenges SpaceX faces as it works to expand its AI compute infrastructure while maintaining reliability.

What Triggered the Sudden Leadership Reset

The root of the problem traces back to how quickly SpaceX assembled its AI infrastructure in the first place. SpaceXAI built out capacity at an unusually rapid pace, leaving some facilities operating for months without backup cooling and power systems in place. Reliability gaps of this nature rarely stay hidden for long once training workloads scale, and SpaceX’s internal uptime target of 99.9% became increasingly difficult to hit. Temporary fixes, including mobile gas turbines and more than 100 portable chillers, kept the lights on but could not substitute for permanent infrastructure.

The temporary measures contributed to outages at the Macrohard facility that interrupted the training of new AI models, a costly outcome for any operation racing to keep pace with rivals like OpenAI and Anthropic. The leadership changes instead brought executives from SpaceX’s rocket and satellite businesses into the AI infrastructure operation, alongside an engineer from Neuralink and new civil-engineering hires.

Rocket and Starlink Veterans Move Into AI Infrastructure

SpaceX responded by reassigning executives from its rocket and satellite businesses to the AI infrastructure operation. Wesley Salandro, who previously served as vice president of production for the Falcon and Dragon rocket programs, has now joined the AI infrastructure team to bring manufacturing-grade rigor to data center construction. Logan McConnell, who ran product support operations at the Starbase launch site in Texas, has also been reassigned into the unit. Their appointments bring additional experience from SpaceX’s rocket and satellite operations into a data center organization dealing with civil-engineering and reliability concerns. The company has additionally brought in an engineer from Neuralink, broadening the talent pool beyond its traditional aerospace bench. SpaceX is also advertising multiple civil-engineering positions, reinforcing its focus on expanding the engineering team supporting its rapidly growing AI infrastructure.

A Wave of Departures Set the Stage

The leadership overhaul did not happen in a vacuum; it followed a string of high-profile exits from the AI infrastructure team. Jake Palmer, who led physical infrastructure for SpaceXAI, departed in late July, removing one of the unit’s most senior operational figures. Data center leaders Zach Wells and Pablo Mendoza also left the company around the same period, thinning out the team responsible for day-to-day site management. Other departures carried a competitive sting: Brent Mayo moved to OpenAI, while Liz Balke joined Anthropic, both direct rivals in the race for AI compute dominance. Losing talent to competitors while simultaneously managing reliability concerns added further pressure to a team already undergoing significant changes. The rocket and Starlink transfers therefore represent a significant reinforcement of the AI infrastructure team following a series of senior departures.

Colossus, Macrohard, and the Roots of the xAI Deal

Understanding the current crisis requires tracing the division’s unusual origin story. The unit at the center of the shake-up did not start inside SpaceX at all; it was xAI, the artificial intelligence company Elon Musk founded in 2023. SpaceX acquired xAI in February 2026 in a transaction that valued xAI at $250 billion and SpaceX at $1 trillion, creating a combined company valued at $1.25 trillion, and formally rebranded the AI operation as SpaceXAI five months later.

The division’s flagship compute operations are centered in Memphis, Tennessee, while natural-gas turbines supporting the Colossus 2 facility have been installed just across the state line in Southaven, Mississippi. Musk later expanded capacity with the Colossus 2 facility, which has also been referred to as ‘Macrohard,’ reflecting the scale of xAI’s rapidly expanding compute operation. That rapid-build approach resulted in facilities relying heavily on temporary power and cooling equipment while permanent infrastructure was being developed.

Community Groups and Regulators Push Back on Gas Turbines

The infrastructure strain has not stayed confined to internal operations; it has drawn outside scrutiny as well. The NAACP and other community groups have sued over the Colossus cluster’s reliance on dozens of unpermitted natural-gas turbines used for backup power, according to TechCrunch. Local residents and community groups have raised concerns about air pollution and regulatory compliance surrounding the rapidly expanding computing infrastructure. SpaceXAI confirmed the expansion of its Memphis-area footprint while beginning a phased removal of temporary turbines under an agreed order with the Mississippi Department of Environmental Quality. That parallel expansion and turbine-removal process highlights the regulatory and infrastructure challenges accompanying the company’s rapid growth. Meanwhile, the legal exposure adds another layer of complexity to a leadership team already managing engineering fires.

Michael Nicolls Inherits the Turnaround

Ultimate responsibility for the AI infrastructure operation now falls to Michael Nicolls, a senior vice president for Starlink engineering whom Musk installed as president of the newly renamed SpaceXAI back in April. The facilities Nicolls inherited had leaned heavily on temporary equipment, including mobile gas turbines, Tesla Megapack batteries, and more than 100 mobile chillers, Tesla Megapack batteries, and more than 100 portable chillers, a temporary setup that made it more difficult to keep the Macrohard facility running as consistently as traditionally built data centers. That patchwork contributed to outages that interrupted AI model training runs, while the site’s uptime rate remained below an internal goal of at least 99.9%.

Nicolls now oversees a team reinforced with rocket-program discipline, tasked with converting stopgap measures into durable systems. SpaceX is separately building turbine-blade manufacturing capacity in Texas that Musk said could shorten generator deployment timelines by as much as 18 months, Tom’s Hardware reported. If achieved, that expected reduction in turbine-component delivery times could help SpaceX bring new power-generation capacity online more quickly.

Investors Register a Modest but Telling Reaction

Markets took notice of the reorganization, though the reaction remained measured rather than alarmed. Shares of SpaceX (SPCX) were trading near $143 around the time of writing, but the move alone does not establish how investors interpreted the leadership shake-up. The stock remains comfortably above its $135 IPO price, indicating that public market confidence in SpaceX’s broader trajectory has not eroded. However, shares remain below their all-time high near $225, although the price gap cannot by itself be attributed specifically to concerns about execution risk in the AI infrastructure buildout. Compute reliability has emerged as an important operational consideration as SpaceX expands its AI infrastructure alongside its broader businesses. The relatively limited stock move provides no clear evidence that investors have materially repriced SpaceX specifically because of the data-center leadership changes.

Can SpaceX Still Hit Its Two-Gigawatt Target?

The most consequential question raised by the shake-up concerns timing. Plans for additional data centers beyond SpaceX’s Memphis hub, including proposed projects in Texas, have already slipped according to the original report. That slippage raises legitimate doubts about whether SpaceX can still hit its stated goal of more than two gigawatts of computing capacity by the end of the year. Reaching that figure would require SpaceX to continue expanding capacity while addressing the reliability and construction challenges affecting its existing and planned facilities. The addition of manufacturing veterans like Salandro could accelerate construction discipline over time, but new leaders typically need weeks or months to fully absorb site-specific complexity. Whether SpaceX closes the gap before year-end will depend on the pace of new data-center construction, the availability of power and cooling infrastructure, and the company’s ability to maintain reliable operations as capacity expands.

Why the Stakes Extend Beyond a Single Data Center

This episode carries weight well beyond Memphis and Southaven because it tests whether SpaceX’s aerospace-grade execution culture can transfer cleanly into the data center business. Rocket manufacturing and AI infrastructure operate under different construction and deployment demands, while SpaceXAI has emphasized rapid deployment as a key advantage in its terrestrial compute strategy. SpaceX now appears to be reinforcing that balance by moving veterans from Falcon, Dragon, and Starbase into the AI infrastructure operation.

Furthermore, the outcome will shape how seriously competitors and regulators take SpaceX’s compute ambitions going forward, particularly as legal challenges over the Colossus turbines continue to unfold. If Nicolls and his reinforced team can improve reliability while the company addresses its regulatory obligations and continues expanding capacity, SpaceX may still close in on its two-gigawatt target. If not, the company could face additional execution challenges in an AI infrastructure market where both reliable operations and available capacity are increasingly important.