2 Sources
2 Sources
[1]
Jensen Huang Says Elon Musk's GPUs‑in‑Space Idea Isn't Crazy -- But 3 Huge Barriers Remain
Nvidia just delivered one of the biggest quarters in tech history. The chipmaker reported $68 billion in revenue, up 73 percent year over year, a staggering figure that underscores just how integral its GPUs have become to the artificial intelligence boom. As cloud providers and startups alike race to build AI models, Nvidia has positioned itself at the center of the arms race. But during the Nvidia earnings call, CEO Jensen Huang looked beyond the blowout numbers. Instead of dwelling solely on record demand, he turned to a more speculative question asked by an attendee on the call: what happens when Earth itself becomes the bottleneck? As AI workloads surge, terrestrial data centers are running up against real-world constraints, including power grid limitations, land scarcity, and the immense cooling requirements of dense GPU clusters. Those pressures have pushed a once-fringe idea into more serious industry discussion: moving compute into orbit.
[2]
Jensen Huang Thinks Orbital Datacenters Have Poor Economics Right Now, Says Cooling Is The Bottleneck - NVIDIA (NASDAQ:NVDA)
Nvidia Corp. (NASDAQ:NVDA) CEO Jensen Huang shared his views on orbital data centers during the chipmaker's Q4 earnings call, outlining economic challenges associated with the phenomenon. Poor Economics, Abundance Of Energy During the company's earnings call with investors and analysts, Huang was asked about his views on space-based AI compute. "The economics are poor today, but it is going to improve over time," he said. He also added that space had an "abundance of energy" and "plenty of space" in orbit, outlining positive factors in the race for solar-powered AI satellites. However, he acknowledged that there was no airflow in space, which means that "the only way to dissipate heat is through conduction," he said. He added that radiators needed for the heat dissipation were "fairly large." Space-based hardware isn't novel for the chipmaker. "MPS is already the world's first GPU in space, Hopper is in space," Huang said. Gene Munster Weighs In Investor Gene Munster of Deepwater Asset Management took Huang's comments as a positive signal for the future of AI datacenters in space. In a post on the social media platform X on Wednesday, Munster summarized Huang's views on the topic. "My take: Jensen message is orbital data centers are difficult today and worth pursuing," the investor said in the post. Nvidia's Earnings The chipmaker posted record revenue figures in Q4, reporting $68.13 billion, up 73% from last year's fourth quarter and beating the Street consensus of $66 billion. Nvidia's EPS earnings of $1.62 also beat market estimates of $1.53 per share. Elon Musk's Goal Benzinga Edge Rankings show that Nvidia scores well on the Momentum, Growth and Quality metrics. It also boasts a positive price trend in the short, medium, and long term. Price Action: NVDA surged 1.41% to $195.56 at Market close on Wednesday. Check out more of Benzinga's Future Of Mobility coverage by following this link. Photo courtesy: jamesonwu1972 On Shutterstock.com Market News and Data brought to you by Benzinga APIs To add Benzinga News as your preferred source on Google, click here.
Share
Share
Copy Link
Nvidia CEO Jensen Huang addressed the feasibility of space-based AI compute during the company's earnings call, acknowledging that orbital datacenters currently have poor economics. While space offers abundant energy and room for expansion, cooling remains the critical bottleneck as heat dissipation in zero airflow requires large radiators.
Nvidia CEO Jensen Huang tackled one of the tech industry's most speculative questions during the chipmaker's recent earnings call: can orbital datacenters solve the mounting constraints facing terrestrial data center limitations? The answer, according to Huang, is complex. While he didn't dismiss the concept of GPUs in space outright, he painted a realistic picture of the technical and economic hurdles that make space-based AI datacenters impractical today
1
.
Source: Benzinga
The discussion emerged against the backdrop of Nvidia's record earnings, with the company reporting $68.13 billion in revenue—a 73 percent increase year over year—and earnings per share of $1.62 that beat market estimates
2
. These staggering figures underscore how central Nvidia has become to the AI boom, as cloud providers and startups race to build increasingly powerful AI models.As AI workloads surge, ground-based infrastructure is hitting physical limits. Power grid limitations, land scarcity, and the immense cooling requirements of dense GPU clusters are creating real constraints for data centers. These pressures have elevated what was once a fringe concept into serious industry conversation
1
. The question of moving compute into orbit addresses a fundamental challenge: as demand for AI computing continues to accelerate, where will the necessary infrastructure actually fit?During the earnings call, Jensen Huang was direct about the current state of orbital datacenters. "The economics are poor today, but it is going to improve over time," he stated
2
. He acknowledged that space offers abundant energy and space in orbit—two critical advantages that could theoretically address power and real estate constraints plaguing ground-based facilities.However, cooling remains the most significant technical barrier. Without airflow in space, heat dissipation becomes exceptionally difficult. "The only way to dissipate heat is through conduction," Huang explained, noting that the radiators required for this process are "fairly large"
2
. This cooling challenge represents a fundamental physics problem that can't be easily engineered away with current technology.Related Stories
Interestingly, space-based hardware isn't entirely new territory for Nvidia. Huang revealed that "MPS is already the world's first GPU in space, Hopper is in space"
2
. This indicates the chipmaker has been testing its technology in orbital environments, gaining practical insights into the challenges and opportunities of space-based computing.
Source: Inc.
Investor Gene Munster of Deepwater Asset Management interpreted Huang's comments as cautiously optimistic. "My take: Jensen message is orbital data centers are difficult today and worth pursuing," Munster posted on social media platform X
2
. This perspective suggests that while immediate implementation faces substantial obstacles, the concept merits continued research and development.The implications for AI computing are significant. If terrestrial constraints continue to tighten while space-based solutions improve economically and technically, the industry could see a gradual shift toward hybrid infrastructure models. For now, companies should monitor advances in space cooling technology, launch cost reductions, and regulatory frameworks governing orbital data centers. The connection to Elon Musk's broader space ambitions adds another dimension to this emerging conversation about where the next generation of AI infrastructure will physically reside.
Summarized by
Navi
[1]
17 Nov 2025•Technology

31 Oct 2025•Technology

22 Oct 2025•Technology

1
Business and Economy

2
Technology

3
Policy and Regulation
