Nvidia’s Vera-Rubin Architecture Hits Full Production Amid Capital Rotation to SpaceX

Over the past month, two very different Silicon Valley stories have collided in Nvidia’s share price. On one side sits a company rolling out its next-generation hardware platform, building the chips for autonomous AI agents, desktop artificial intelligence, and even humanoid robots. On the other sits the gravitational pull of Elon Musk’s SpaceX, which is preparing a public listing that could instantly make it the seventh-largest US-listed company by market cap — and forcing hedge funds to sell big tech names to free up cash.

The result: Nvidia’s stock closed at €177.28 on Friday, sitting precisely on its 50-day moving average, a level that has historically acted as a tipping point. That marks a loss of roughly 2% for the week and a drop of about 11% from the all-time high of €202.50 reached in mid-May.

The operational story, however, could hardly be more different. The Vera-Rubin platform — named after the astronomer who discovered dark matter — is now in full production. Built around an 88-core processor on a 3-nanometer node paired with an entirely new GPU generation, the architecture is designed to handle what Nvidia’s CEO Jensen Huang calls “agentic workloads”: autonomous systems that execute multi-step tasks rather than simply answering queries. Nvidia is moving decisively from being a supplier of training chips to becoming the architect of the entire inference and agent-logic pipeline.

That ambition extends well beyond the data center. With the RTX Spark platform, Nvidia is directly challenging the x86 processor dominance that has defined the PC market for decades. By coupling its own Grace CPUs with Blackwell graphics, the company is positioning itself as the provider of integrated AI chips for everyday computing. The DGX Station for Windows is the most visible product in this push — a device that signals Nvidia’s intention to own the interface where AI developers and enterprise users work, not just the hidden engine in the cloud. Analysts are already talking about a “CPU super-cycle” in which traditional processors give way to bundled AI silicon, reshaping industry margins.

A third pillar is Physical AI — intelligence with a body. Nvidia has unveiled the Isaac GR00T platform for humanoid robots and announced partnerships with LG Group and SK Telecom to build so-called AI factories. National telecom operators and manufacturers are emerging as the next major market for data-center-scale infrastructure.

Should investors sell immediately? Or is it worth buying Nvidia?

Investors distracted by the SpaceX rotation may be overlooking the financial momentum. In the first quarter of fiscal 2027, Nvidia generated $81.6 billion in revenue, and management has guided for $91 billion in the current quarter. The infrastructure build-out is not slowing — it is simply changing shape as the market shifts from model training to deployment.

Analyst sentiment remains constructive. DA Davidson has reiterated its buy rating with a $300 price target and added the stock to its “best-of-breed” list. Hedge-fund manager Dan Loeb, despite Nvidia already approaching a $5 trillion market capitalization, argues the shares are still undervalued and sees no growth deceleration on the horizon. The consensus analyst target of €258.25 implies roughly 44% upside from current levels.

Technically, the picture is nuanced. The relative strength index at 48.3 suggests neutral momentum — neither overbought nor oversold. The stock still sits more than 10% above its 200-day moving average of €162.31, so the long-term trend remains intact. But Friday’s close directly on the 50-day line makes that level critical. Should it hold, the all-time high comes back into view. If it breaks, the consolidation may deepen.

A potential catalyst arrives on June 24, when Nvidia holds its virtual annual general meeting. Shareholders will vote on executive compensation, ten director positions, and proposals related to greenhouse-gas emissions and diversity policies.

Underlying demand shows no signs of abating. At a recent Bank of America technology conference, 37 semiconductor executives painted a consistent picture: supply is ramping quickly, but it still cannot keep pace with the staggering appetite for chips that power the next wave of AI. The near-term rotation out of tech into SpaceX is a liquidity event, not a fundamental verdict. Nvidia’s broader transformation — from chip seller to infrastructure builder — remains the plot that matters.