Nvidia's Vera Rubin enters production as analyst flags 62% upside

Nvidia's Vera Rubin platform has entered production, Chief Executive Officer Jensen Huang said July 15, extending the company's lead as rivals Advanced Micro Devices Inc. and Broadcom Inc. invest billions to challenge its AI chip dominance.

"Vera Rubin is progressing well and has entered production," Huang said in a statement, without disclosing a specific timeline for customer shipments.

The Vera Rubin architecture succeeds Nvidia's Blackwell platform, which began ramping in late 2024. Super Micro Computer Inc. introduced a blueprint based on the Vera Rubin NVL4 platform at the ISC 2026 conference in June, with each Scalable Unit housing as many as 1,152 Rubin graphics processing units and 576 Vera central processing units, using liquid cooling and Nvidia's Quantum-X800 InfiniBand networking. The platform is built on TSMC's advanced process node and uses next-generation high-bandwidth memory, though Nvidia has not disclosed the specific node or memory configuration.

Nvidia's Data Center revenue surged 92% year over year to $75.2 billion in the fiscal first quarter ended April, while total revenue hit a record $81.6 billion — an 85% increase. The company guided for second-quarter revenue of about $91 billion, signaling continued demand for its AI infrastructure. The Zacks consensus estimate for fiscal 2027 revenue stands at $385.5 billion, implying 78.5% annual growth.

Despite reports of a slight delay in the Rubin rollout, one analyst still expects 62% upside for Nvidia shares, citing the company's widening competitive moat through its integrated hardware-software ecosystem. Nvidia has expanded partnerships with Google Cloud to deploy Vera Rubin-powered AI instances, with Marvell Technology Inc. through NVLink Fusion for custom AI infrastructure, and with Coherent Corp., Corning Inc. and Lumentum Holdings Inc. to improve optical networking for next-generation data centers.

The Vera Rubin production milestone comes as Nvidia's rivals accelerate their own AI chip efforts. AMD's Data Center segment revenue rose 57% year over year to $5.78 billion in its first quarter, driven by demand for its Instinct GPUs and EPYC processors, while the company has strengthened its open-source ROCm software platform to attract developers. Broadcom has also been investing aggressively in custom AI accelerators for hyperscale customers.

Nvidia's software ecosystem — including open-source platforms Dynamo, Agent Toolkit and Nemotron — encourages developers to build AI applications on Nvidia hardware, creating switching costs that make it harder for customers to migrate to competing platforms. The company's broad partner network and integrated chip-to-software stack create a structural advantage that rivals have yet to replicate at scale.

Nvidia shares trade at about 35 times forward earnings. The production ramp of Vera Rubin, combined with the company's expanding software moat and partner ecosystem, suggests the market has not fully priced in the platform's revenue contribution for fiscal 2028, according to analysts tracking the stock.

This article is for informational purposes only and does not constitute investment advice.