Nvidia's Vera Rubin AI Platform Faces Minor Delay but Target Price Raised

Nvidia's next‑generation AI platform, Vera Rubin, is expected to enter full production this year, but a modest delay has been reported due to thermal‑dissipator issues and a hold‑up in qualifying SK Hynix's HBM4 memory. Analyst John Vinh of KeyBanc noted that the setback is limited, raising his target price for Nvidia shares from $310 to $330 and maintaining an Overweight rating. He cited stronger‑than‑expected demand, a 69% jump in CoWoS supply forecasts for 2027 and plans to ship the Rubin Ultra model beginning late 2027, with interim shipments of B300 GPUs.

The platform is designed for AI research labs, cloud hyperscalers and high‑performance workloads, featuring a POD‑scale architecture that integrates five racks into a single supercomputer. Early adopters such as Digi Power X have committed $35 million to purchase Vera Rubin capacity for GPU‑as‑a‑Service offerings. Despite the minor delay, analysts expect Nvidia's AI data‑center revenue to keep expanding, supported by projected data‑center capital expenditures that could rise from $650 billion to over $1 trillion. Wall Street forecasts revenue growth of about 41% for the coming year, and the stock has risen roughly 7.8% year‑to‑date, closing at $203.53 after a recent dip.