FEATURED

NVIDIA’s Vera Rubin Is Shipping This Fall

There’s a moment in every major platform cycle where the market knows something big is coming and still somehow underestimates it. We might be in one of those moments right now with NVIDIA.

Vera Rubin is in full production.

That’s not a roadmap item anymore. It’s not a teaser slide from a keynote. NVIDIA confirmed at GTC Taipei and Computex in May and June 2026 that the Vera Rubin platform is ramping into full production, with partner availability in the second half of this year. AWS, Google Cloud, Microsoft Azure, OCI, CoreWeave, Lambda, Nebius, and Nscale are among the first confirmed cloud partners to deploy Vera Rubin-based instances. Shipments to customers are expected to begin this fall, with enterprise on-prem allocation extending into 2027 given current demand queues.

Slight tangent, but it matters: a lot of investors are fixated on the stock’s near-term drift and missing the bigger picture. NVDA is up only about 7% year-to-date as of late June 2026, lagging the Philadelphia Semiconductor Index by a wide margin, even as the underlying business has arguably never been stronger. That kind of gap tends to resolve eventually. Usually in one direction.

The Numbers That Matter Right Now

For fiscal Q1 2027, reported on May 20, 2026, NVIDIA posted $81.6 billion in revenue, up 85% year over year. Data center revenue alone hit $75.2 billion, up 92% year over year, accounting for the vast majority of total sales. GAAP gross margin came in at 74.9%, with non-GAAP gross margin at 75.0%. Non-GAAP EPS was $1.87, beating consensus estimates in the range of $1.75 to $1.77. And the Q2 FY2027 guide was set at $91 billion in revenue, with no assumed Data Center compute revenue from China baked in.

Jensen Huang’s words on the earnings call were direct: “Agentic AI has arrived” and “the AI factory buildout is accelerating at extraordinary speed.” That’s the framing NVIDIA is leaning into hard right now.

Meanwhile, sovereign AI generated over $30 billion in fiscal 2026 revenue, more than tripling year over year, and now represents roughly 14% of NVIDIA’s total annual revenue. France deployed 18,000 Grace Blackwell systems. The UK, Netherlands, Canada, and Singapore have committed to large-scale deployments. That’s a demand category that barely existed a few years ago, and it’s now a material line item on the income statement.

What Vera Rubin Actually Does

The Vera Rubin NVL72 rack trains mixture-of-experts models with one-fourth the GPUs required on the prior Blackwell platform. Agent throughput at scale is 10x higher than the Grace Blackwell generation. Cost per token drops by a factor of ten. For the hyperscalers building out agentic AI infrastructure, those efficiency gains are not incremental. They’re structural.

The supply side is worth watching. Jensen Huang has said NVIDIA has enough supply for very robust growth, but the company remains supply-constrained. More than 350 supply chain partners across 30 countries, including 150 in Taiwan alone, are currently ramping Vera Rubin production. That’s a meaningful signal about the scale of the build underway.

Analyst sentiment sits at a strong buy consensus, with average price targets around $300 per share. The stock is currently around $200, off from its all-time high of $236 set in mid-May 2026. That gap between target and price reflects a market that broadly agrees on direction but disagrees sharply on timing and magnitude.

The Part People Gloss Over

The bear case is real and worth taking seriously: hyperscalers are investing in custom chips, China revenue is effectively zero in NVIDIA’s own forward guidance, and there’s a genuine debate about whether demand durability justifies current valuation multiples. These are not fringe concerns.

But the counter is equally real. NVIDIA remains the only platform running every major frontier AI model. Inference is growing faster than training. Sovereign and enterprise demand are diversifying the customer base away from hyperscaler concentration. And Vera Rubin ships directly into that demand environment starting this fall.

What’s interesting is the distance between where the stock is and where the business is. The income statement confirmed the demand. The platform transition is underway. The stock has lagged the semiconductor sector by a wide margin all year. H2 2026 is when Vera Rubin starts landing at scale. That’s the situation investors are looking at right now.

Worth a close look before that window opens.