Musk Signals Chip Preference on SpaceX’s First Public Earnings Call as GPU Rivals Battle for AI Share

The race to supply artificial intelligence compute has largely been a two-company contest between Nvidia and Advanced Micro Devices, and a high-profile comment from Elon Musk on SpaceX’s debut earnings call as a publicly traded company has added a new layer to the rivalry.

Nvidia, which closed at $223.67 (down 0.73%) and carries a market capitalization of roughly $5.56 trillion, is shipping its Blackwell architecture at scale while preparing Vera Rubin as its next rack-scale platform. AMD, whose shares rose 2.81% to $516.13 and which holds a market cap near $780 billion, is currently selling MI300-based systems and has its MI400 accelerator generation lined up as a follow-on aimed at keeping pace in the data center.

On the demand side, the largest AI builders are refusing to commit exclusively to either vendor. Companies including Alphabet, Microsoft, Oracle, and Meta Platforms, along with Anthropic and OpenAI, have been deploying hardware from both Nvidia and AMD within the same infrastructure stacks. The dual-sourcing strategy reflects constrained accelerator supply and the fact that different workloads favor different chip architectures — waiting on a single vendor is not an option at the scale of current AI build-outs.

Against that backdrop, Musk — SpaceX’s chief executive — used the company’s first earnings call as a public company to make clear which of the two silicon suppliers he favors, inserting himself directly into one of the technology sector’s most closely watched competitive storylines.

The stakes are substantial. Combined, Nvidia and AMD account for the vast majority of merchant GPU supply for AI training and inference, and their divergent market caps — roughly $5.56 trillion versus $780 billion — underscore how far ahead Nvidia remains in the accelerator market even as AMD presses forward with each new product generation.

What to watch

• AMD’s progress toward MI400 system deployments and any customer announcements.
• Nvidia’s ramp of the Vera Rubin rack-scale platform following Blackwell.
• Details from SpaceX’s upcoming earnings reports and any further comments from Musk on compute suppliers.
• Upcoming quarterly results from Alphabet, Microsoft, Oracle, and Meta for updated AI infrastructure spending signals.

Source: original release