Nvidia’s, Record

Nvidia’s Record Quarter Masks a Strategic Pivot: Vera CPU Shines as Inferencing Chip Fades

Published on 05/28/2026 at 06:04 | Redaktion boerse-global.de

Nvidia's record Q1 revenue $81.6B, $80B buyback; stock down 1.7% cancels Rubin CPX, signs $20B Groq deal, pivots to platform AI. Vera CPU outpaces AMD/Intel.

Nvidia’s Record Quarter Masks a Strategic Pivot: Vera CPU Shines as Inferencing Chip Fades Illustration mit AI erstellt übermittelt durch boerse-global.de
Nvidia’s Record Quarter Masks a Strategic Pivot: Vera CPU Shines as Inferencing Chip Fades Illustration mit AI erstellt übermittelt durch boerse-global.de

Nvidia posted a record $81.62 billion in first?quarter revenue — a 85.2% surge year?on?year — and authorised a fresh $80 billion share buyback, yet the market reacted with caution. Shares slipped to EUR 181.32, down 1.7% on the day and 5.6% on the week, as investors digested a quiet but significant shake?up in the company’s product roadmap. The data centre segment alone contributed $75.25 billion, growing 92%, and management guided for around $91 billion in the current quarter. But the headline numbers obscure a deeper transformation: Nvidia is dialling back its reliance on standalone chip iterations and betting instead on platforms, licensing, and industrial AI workflows.

The clearest sign of the shift is the fate of the Rubin CPX inferencing chip. Initially slated for the second half of the year with 128 GB of GDDR7 memory, the project has all but disappeared. Supply?chain sources report no visible orders for the required memory or substrates, and the chip was absent from the roadmap presented at GTC 2026. Industry observers now consider it effectively cancelled, even as Nvidia avoids labelling it a retreat. In its place, the company has struck a $20 billion licensing agreement with Groq 3 LPX, focusing on low?latency real?time inferencing rather than raw GPU throughput. The pivot dovetails with a push into agentic AI: Nvidia unveiled PhysicsNeMo as the cornerstone of a platform for agent?based digital development, promising to triple productivity and cut development cycles by 60% for industrial partners such as McLaren Automotive.

On the hardware front, however, Nvidia is far from stepping back. The first Vera?based Rubin racks have already been delivered to OpenAI, Anthropic, SpaceX and Oracle, with a broad market launch scheduled for the second half of 2026. The Vera CPU itself — an 88?core ARM?based processor with 176 threads — posted striking benchmarks this week. In Phoronix tests it led the AMD Epyc 9575F by 11% and the Intel Xeon 6980P by 55%, while delivering memory bandwidth of 1.2 TB/s and a STREAM TRIAD utilisation rate of 90%. Nvidia selected the workloads carefully, a caveat analysts will note, but the numbers underscore a processor that could crack a $200 billion addressable market. The company expects CPU?related revenues to reach $20 billion by the end of fiscal 2027. The Vera platform is also designed to cut inferencing costs by 90% while using 75% fewer GPUs than the current Blackwell generation — projections that, if confirmed, would reshape data?centre economics.

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

CEO Jensen Huang used the moment to push back against a narrative he calls lazy. Managers blaming AI for layoffs, he said, are “irresponsible” and “lazy”, arguing that AI has been deployed productively for only about six months and cannot logically account for job cuts that occurred two years ago. The real driver of restructuring, he insisted, is the enormous capital outlay for AI infrastructure — not the technology itself. Meanwhile, Nvidia’s physical footprint continues to expand. In Taiwan, which Huang called the “epicentre of the AI revolution”, annual spending has multiplied tenfold over five years to roughly $150 billion. A new Taipei headquarters is set to break ground in 2026 and open by 2030. In Israel, Nvidia secured a ten?year lease for 29,000 square metres in Yokneam, part of a planned 160,000?square?metre campus in Kiryat Tivon.

The financial firepower for all this is evident. The first?quarter revenue record rests on $119 billion in outstanding purchase commitments, illustrating the deep ties in Nvidia’s supply chain. The company also expanded its collaboration with Marvell Technology through a $2 billion investment aimed at co?developing custom XPUs and networking solutions for NVLink Fusion. Capital returns continue: the quarterly dividend was raised to $0.25 per share alongside the new $80 billion buyback authorisation. Yet the stock remains roughly 9% below its 52?week high of EUR 201.05 (mid?May), and the relative strength index hovers near 40 — a level that historically has been seen as an entry point. Whether the Vera benchmarks and early Rubin rack deliveries can reignite momentum will become clearer with the next earnings release in August. For now, Nvidia is placing its bets not just on the next chip, but on a broader AI infrastructure play that spans silicon, licensing and industrial platforms.

Ad

Nvidia Stock: New Analysis - 28 May

Fresh Nvidia information released. What's the impact for investors? Our latest independent report examines recent figures and market trends.

Read our updated Nvidia analysis...

Disclaimer regarding our articles: No investment advice, no buy or sell recommendation. Information on prices, companies, and markets is provided without guarantee; changes are possible at any time. Stock market transactions can lead to substantial losses. Our articles are created and reviewed in whole or in part automatically with the support of AI.

en | US67066G1040 | NVIDIA’S | boerse | 69429916 |