Nvidia's Revenue Forecast is so Huge That Wall Street Wonders if SpaceX is the Reason

Dow Jones
Yesterday

SpaceX plans to spend tons of money on chips and recently announced it will only buy Nvidia's. But other high-profile companies need more processing power as well.

SpaceX recently said it would only use Nvidia's data-center chips.

Nvidia delivered a "shockingly bullish" outlook for its next fiscal year, in the words of one analyst - but the sheer level of expected growth has some investors wondering what's behind such upbeat projections.

Nvidia (NVDA) expects to grow revenue by more than 70% in the fiscal year that ends in January 2028. Now, a key question for Wall Street is whether that forecast is "simply a function" of SpaceX's (SPCX) plan to use only Nvidia's data-center chips, Mizuho analyst Daniel O'Regan said in a note to clients.

With SpaceX aiming to deploy close to 10 gigawatts of computing power by the end of next year, the company is increasingly relevant to Nvidia, said Kate Leaman, chief market analyst at AvaTrade. SpaceX CEO Elon Musk has said his company expects a significant share of Nvidia's upcoming Rubin chips next year, "which underlines just how material the relationship is becoming," Leaman noted.

While SpaceX and Nvidia have worked together for roughly a decade, the companies haven't always been as close as they now appear to be. Nvidia is both a major supplier to and investor in SpaceX, owning $21 billion of the company's stock as of June 30.

SpaceX also accounted for 5% of Nvidia's fiscal second-quarter revenue, according to Gene Munster, managing partner at Deepwater Asset Management. That's up from 3% in the prior quarter, he said.

Despite the companies' strengthening relationship, Leaman would caution against tying Nvidia's growth forecast too closely to a single customer.

Nvidia's various hyperscaler customers altogether drove $49 billion of its $89 billion in data-center revenue in the latest quarter, according to the company's 10-Q filing. Additionally, sovereign artificial-intelligence efforts, neoclouds, enterprises, startups and other nonhyperscaler customers are driving a large part of Nvidia's data-center business, she noted.

"That's a lot of demand coming from a lot of directions," Leaman said in emailed comments.

She pointed out commentary from Nvidia's management that its revenue forecast factors in that there won't be enough supply to meet all its demand. Nvidia estimates its cloud-industry backlog is now above $2 trillion.

"The SpaceX commitment is significant, and the exclusivity angle is a real competitive signal," Leaman said. But zooming out, Nvidia has more demand than it can keep up with, she added, and the AI race will be focused on whether supply is able to catch up.

Neither Nvidia nor SpaceX immediately responded to MarketWatch's requests for comment.

Another question is over whether Nvidia will offer Musk financial support to reach his goals tied to computing power. Analysts expect SpaceX to lean heavily on debt over the next decade, and Mizuho estimates the company will need between $300 billion and $500 billion to meet its 2027 compute target.

Earlier this month, Nvidia announced a partnership with financial firms including Apollo Global Management $(APO)$, Blackstone (BX) and Goldman Sachs (GS) to raise more than $500 billion in third-party capital to finance the AI infrastructure build-out. While much is still unknown about the effort, Mizuho questioned if it was targeted at SpaceX.

The partners said their work would enable the build-out across Nvidia's "ecosystem," including frontier AI labs. Kevin Cook, a senior stock strategist at Zacks Investment Research, said Nvidia's aim to make its graphics processing units an investable asset class offers a case for why it delivered such strong revenue guidance.

The members of the partnership want to see sovereign nations and enterprises build out their own AI offerings, Cook said, and that will ultimately benefit Nvidia as the market for its computing hardware expands.

Brian Colello, an analyst at Morningstar, said it's possible to think of Nvidia's computing infrastructure like a car: It can be financed by a customer, repossessed and sold back to another customer or used internally. "It would take a lot of things to go wrong for this to really backfire on Nvidia," he added of the financing arrangement.

-William Gavin -Britney Nguyen

 

At the request of the copyright holder, you need to log in to view this content

Disclaimer: Investing carries risk. This is not financial advice. The above content should not be regarded as an offer, recommendation, or solicitation on acquiring or disposing of any financial products, any associated discussions, comments, or posts by author or other users should not be considered as such either. It is solely for general information purpose only, which does not consider your own investment objectives, financial situations or needs. TTM assumes no responsibility or warranty for the accuracy and completeness of the information, investors should do their own research and may seek professional advice before investing.

Most Discussed

  1. 1
     
     
     
     
  2. 2
     
     
     
     
  3. 3
     
     
     
     
  4. 4
     
     
     
     
  5. 5
     
     
     
     
  6. 6
     
     
     
     
  7. 7
     
     
     
     
  8. 8
     
     
     
     
  9. 9
     
     
     
     
  10. 10