The chip maker recorded its largest-ever weekly increase in market capitalization
Nvidia, led by CEO Jensen Huang, booked a record weekly gain in market value on Friday.
Nvidia's stock rose on Friday to secure its best weekly performance in more than a year, following a major endorsement for its artificial-intelligence chips and further confirmation that hyperscaler spending is robust.
The chip maker's (NVDA) upcoming Vera Rubin chip architecture was touted as "the best" by SpaceX $(SPCX)$ CEO Elon Musk on his company's first-ever earnings call Tuesday afternoon, when Musk said that SpaceX has "decided to build exclusively on Nvidia" moving forward.
"We think it's the best AI computer, and we greatly value our close cooperation and partnership on many levels with Nvidia," Musk said.
Nvidia's stock closed up 2.3% on Friday afternoon, and rose 11.6% this week. That made for its best weekly gain since May 16, 2025, when it rose 16.1%, according to Dow Jones Market Data. The chip maker also added $562 billion to its market capitalization this week, which is the company's largest-ever weekly gain on record, the data showed.
SpaceX's earnings report followed Big Tech results that were already fueling Nvidia's stock, as Google $(GOOGL)$ $(GOOG)$, Amazon.com (AMZN) and Meta Platforms (META) all guided their capital expenditures higher for the year.
Aside from Musk's praise, "the bigger story is how much the major hyperscalers are spending on AI infrastructure right now," Kate Leaman, chief market analyst at AvaTrade, told MarketWatch, adding that Nvidia remains the top supplier of graphics processing units.
Willy Lee, a principal at investment firm Neostellar, told MarketWatch that while Google's first bout of quarterly negative free cash flow spooked investors, Microsoft's $(MSFT)$ "more measured approach" to capex and its accelerating cloud growth showed investors a more positive view of all that spending.
With Nvidia's own earnings report expected later this month, "some traders are clearly positioning ahead of that," Leaman said in emailed comments. Expectations are already high, however, so a pullback is possible, she noted.
Meanwhile, Musk said SpaceX's "understanding with Nvidia is that we will receive a very significant percentage" of its GPUs next year, which is when SpaceX plans to install close to 10 gigawatts of compute capacity.
BNP Paribas analyst Stefan Slowinski said Musk's commentary raises a question of whether availability of GPUs "could once again emerge as a key constraint on AI-infrastructure deployments" in the next 12 to 18 months. That would both offer risks and opportunities, he said in a note.
Compute providers that lack a strong strategic relationship with Nvidia could fall behind, said Slowinski, who does not cover SpaceX. However, customers that do have access to Nvidia's allocations, or that can scale custom silicon projects, "stand to benefit disproportionately," he added.
In Slowinski's view, CoreWeave (CRWV), Nebius Group (NBIS) and Oracle $(ORCL)$ would be among the best-positioned cloud providers to receive a meaningful share of Nvidia's chips in the event of supply shortages, though those companies face a possible risk of Nvidia prioritizing the growth potential of its SpaceX relationship.
SpaceX's emphasis on its allocation from Nvidia "further reinforces that the AI-compute shortage is unlikely to ease, even through 2027," Slowinski said.
Additionally, SpaceX's suggestion that memory output is growing at about 20% each year while AI demand is growing at a 200% clip supports the idea that supply is still severely lagging demand, he noted.
"If even directionally accurate, this would suggest that the current favorable pricing environment is unlikely to normalize any time soon," Slowinski said.
-Britney Nguyen
Comments