SpaceX commits exclusively to Nvidia chips as AI infrastructure spending hits $18.4 billion

6 Sources

Share

Elon Musk declared SpaceX will exclusively use Nvidia's Vera Rubin GPUs for its AI infrastructure buildout, targeting up to 10 gigawatts of computing power by 2027. While Nvidia stock surged 4%, SpaceX shares tumbled 12% after revealing second-quarter capital expenditures of $18.4 billion that exceeded analyst expectations.

SpaceX Commits Exclusively to Nvidia for AI Infrastructure

Elon Musk announced during SpaceX's first earnings call as a public company that the rocket manufacturer will exclusively use Nvidia's chips for its AI infrastructure expansion

1

. "We've decided to build exclusively on Nvidia because we think the Vera Rubin architecture is the best architecture," Musk stated on the earnings call

2

. This commitment marks a significant shift from May 2025, when Musk indicated SpaceX would use both AMD and Nvidia processors in its hardware buildout

1

.

Market Reactions Diverge Sharply

The announcement triggered contrasting stock movements that highlight a defining pattern in AI infrastructure spending. Nvidia shares climbed more than 3% initially and were up 4% in subsequent trading, marking their fifth consecutive day of gains

1

2

. Meanwhile, SpaceX shares dropped more than 7% initially and fell 12% overall after the company revealed second-quarter capital expenditures of $18.4 billion that significantly exceeded analyst estimates

1

2

. The diverging moves underscore how hyperscalers face investor pressure for AI spending while chipmakers get rewarded as money flows their way

1

.

Massive Computing Power Targets and Revenue Implications

Musk outlined ambitious plans for SpaceX's AI computing platform expansion. The company intends to end 2026 with over 2 gigawatts of compute capacity and expects cumulative online compute to reach between 5 and 10 gigawatts by the end of 2027

4

. When pressed about his confidence in these projections, Musk acknowledged uncertainty, stating his "best guess" for annual revenue each gigawatt of Nvidia's Rubin systems could generate was "somewhere between 30 and $50" billion, adding "This is just a guess"

1

. He also conceded that if a quarter of projects slip, SpaceX would likely land "close to 15" gigawatts rather than the 20-gigawatt target he initially floated

1

.

Analysts Calculate Potential Nvidia Revenue Windfall

Wall Street analysts quickly translated Musk's commitment into potential revenue projections for Nvidia. Chris Caso of Wolfe Research calculated that SpaceX would add approximately 6 gigawatts of additional Nvidia computing power in 2027, which at $35 billion per gigawatt "translates to close to $200 billion in potential revenue for NVDA in calendar year 2027"

1

. Melius Research analyst Ben Reitzes noted this spending intention "could finally get Jensen Huang to formally forecast that he has a line of sight to $1T in revenue well before 2030," potentially substantiating Nvidia CEO Jensen Huang's March prediction that flagship processors could generate $1 trillion in annual revenue through 2027

1

. Nvidia reports fiscal second-quarter results on August 26, when Huang could either confirm or adjust these demand projections

1

.

JPMorgan Highlights Deepening Partnership

JPMorgan emphasized that SpaceX expects to receive a "very significant percentage" of Nvidia's GPU shipments in 2027

3

. The bank highlighted that SpaceX has decided to exclusively deploy Nvidia's Vera Rubin GPUs across its data centers and will also use the Vera Rubin architecture in its planned StarMind AI satellites, further deepening the partnership

3

. This supports SpaceX's plans to rapidly scale AI compute capacity from more than 2 gigawatts by the end of 2026 to between 5 and 10 gigawatts by the end of 2027

3

.

Vera Rubin NVL72 System Deployment Plans

SpaceX plans to launch Nvidia's Vera Rubin NVL72 system starting next year

4

. Musk praised the design, stating "We think the design of the NVL72 VR computer is a much better design than is typical than, say, having a standard rack-style design, and so we expect to actually deploy this on the ground, as well as in orbit"

4

. This deployment strategy extends beyond traditional data centers into orbital AI compute applications.

Orbital AI Compute and StarMind AI1 Satellite Initiative

The SpaceX StarMind AI1 announcement represents a shift from passive relay satellites to active AI inference platforms in space

5

. SpaceX is partnering with Nvidia to design the StarMind AI1 satellite compute payload, with each satellite embedding Nvidia Rubin GPUs and Vera CPUs for datacenter-class compute in orbit

5

. This transforms SpaceX from primarily a launch provider into an AI chip factory in the sky, offering AI-as-a-Service from orbital infrastructure

5

.

Key Questions About Financing and Execution

Wall Street analysts are now debating critical implementation details. "This commentary will be all anyone talks about, along with debates around Elon's ability to actually deliver it within a realistic timeframe, how it will be financed, and what role Nvidia may play in that financing," Melius Research analyst Ben Reitzes wrote

1

. The questions center on how quickly SpaceX can build out this AI infrastructure, financing mechanisms, and whether arrangements might involve circular deals potentially including Nvidia itself

1

. Orbital compute introduces physics-based constraints including power budgets, radiation hardening, and thermal management that ground-based Vera Rubin racks don't face, creating real execution risks

5

.

Industry Commentary and Broader Implications

Jim Cramer, who has been consistently bullish on Nvidia over recent years, reposted Musk's commitment with a five-word response: "and there it is," underscoring the significance of the exclusive partnership

3

. The announcement validates the entire Nvidia Vera Rubin NVL72 ecosystem with ripple effects spanning orbital compute, neocloud infrastructure, and space-grade hardware

5

. Space investment hit $338 billion in H1 2026, up 1,436% year-over-year, and StarMind transforms the sector from launch economics to AI inference platforms in orbit

5

. With Wednesday's gains, Nvidia shares are up 18% since the start of the year, though still about 7% off their May highs

2

.

Today's Top Stories

© 2026 TheOutpost.AI All rights reserved