Musk Says Hiring Nvidia's Jensen Huang Would Be 'an Honor'

Elon Musk said Wednesday night that it would be "an honor" to hire Nvidia CEO Jensen Huang, answering a one-line suggestion posted on X by the markets site ZeroHedge. The exchange consists of exactly two posts and no announcement followed. Neither Huang nor Nvidia had responded on X as of Wednesday night.

ZeroHedge posted "Elon please hire Jensen" at 6:22 p.m. ET. Musk replied two hours later: "That would be an honor." The post, dated November 12, 2026, carries no conditions, no role, and no indication that any conversation has occurred between the two executives. The reply nonetheless touched one of the closest supplier relationships in the technology industry, and that relationship carries some of the largest numbers in it.

What does the Musk-Huang exchange actually rest on?

The foundation is SpaceXSI's stated dependency on Nvidia silicon. On SpaceXSI's August earnings call, Musk said: "going forward, we've decided to build exclusively on NVIDIA because we think the Vera Rubin architecture is the best architecture." That commitment is contractual, not rhetorical.

On October 26, SpaceXSI disclosed a binding commitment to enter into a roughly $40 billion lease of GPUs and related equipment with an Apollo-managed entity. SpaceXSI said the deal will deliver "the next generation of NVIDIA Vera Rubin chips." The leases are expected to begin in the first half of 2027. The scale of the operation behind the commitment is measurable: SpaceXSI reported 1.7 gigawatts of installed compute capacity on September 30, up from 0.6 gigawatts a year earlier — nearly a tripling of capacity in twelve months.

That is the market context for a throwaway hiring joke. A $40 billion lease, an exclusive architecture decision, and a founder who says he wants the other company's chief executive on his payroll — all within a single supplier relationship that neither party has yet commented on beyond the posts themselves.

How large is Musk's own chip-building plan?

Musk is simultaneously constructing a chip supply of his own, and the numbers dwarf most national semiconductor projects. On August 6, SpaceXSI and Tesla announced that Terafab, a vertically integrated complex covering logic, memory, packaging and testing, will be built in Grimes County, Texas. The companies put the initial investment at about $16.8 billion. They described a plant of more than 100 million square feet that will employ at least 3,000 people.

The demand justification the companies gave is larger still. SpaceXSI and Tesla said their combined chip demand is expected to exceed 1 terawatt of compute, a figure they described as "significantly larger than the current global supply." In the same statement, they called themselves "deeply appreciative of our current chip suppliers" — a group that Nvidia, given the August exclusivity decision and the October lease, effectively heads.

Musk has attached a timeline and a caveat to Terafab. On the October 20 earnings call, he said Terafab aims for volume production around the third quarter of 2028. He also said he puts "big error bars" on those internal targets, an acknowledgment that semiconductor fabrication schedules routinely slip.

What has Huang said about Musk?

Huang's public assessment of Musk, delivered in 2024, is directly relevant to why the two names travel together. Nvidia said xAI and Nvidia built the 100,000-GPU Colossus cluster in Memphis in 122 days, with 19 days passing from the first rack on the floor to the start of training. On the BG2 podcast, Huang called the effort "superhuman." He said Musk "is singular in his understanding of engineering and construction and large systems and marshaling resources."

That record explains part of the supplier relationship: Musk's companies consume Nvidia's chips at a pace few customers can match, and Nvidia gets showcase deployments in return. Colossus in Memphis remains the reference case — a 100,000-GPU cluster stood up in 122 days.

How do the two timelines run against each other?

The schedules now run in parallel, and the sequence is fixed. The Vera Rubin leases begin in the first half of 2027. Musk said on the October 20 call that he aims to have Terafab's first test chip — an SRAM-type chip built at the research fab in Austin — working in approximately May or June of that same year.

Until then, Nvidia remains the supplier Musk has said he is building on exclusively. The test chip in Austin and the $40 billion in Vera Rubin leases will overlap in the same window, which means Musk's stated plan has his own fabrication capacity and his leased Nvidia capacity scaling at the same time rather than one replacing the other.

The exchange with ZeroHedge changes none of this. It remains two posts on X: a request at 6:22 p.m. ET and a reply two hours later saying it "would be an honor." What it demonstrates is the degree to which the industry's largest compute buildout — 1.7 gigawatts installed, 1 terawatt projected, $40 billion leased, $16.8 billion committed to a Texas fab — runs through a relationship between two named individuals whose companies have not yet responded to the suggestion that one hire the other.