Nvidia Vera CPU Reaches 4 Top AI Labs Ahead of Rubin Launch

Bank of America raises its 2030 AI infrastructure forecast to $1.7 trillion as Nvidia’s next-generation Vera CPU begins shipping to key customers.

The first wave of Nvidia Corp.’s (NASDAQ:NVDA) next-generation Vera CPUs has arrived at the AI labs of Anthropic, OpenAI, SpaceX, and Oracle, a move that signals the start of the industry’s next major hardware upgrade cycle. The shipments come as Bank of America nearly doubled its price target on memory maker Micron and sharply raised its outlook for the entire AI hardware sector, citing unprecedented demand from cloud providers and AI companies.

"AI demand is scaling faster than traditional enterprise software adoption curves," Bank of America analyst Vivek Arya said in a note to clients. The firm lifted its 2030 forecast for artificial intelligence infrastructure spending for the third time in eleven months, raising it to $1.7 trillion from a previous $1.4 trillion.

The Vera CPU is a key component of Nvidia’s upcoming Vera Rubin platform, which is slated for a full launch in the second half of this year. The new Rubin systems are specified to include 288 gigabytes of HBM4 memory per chip, a significant jump from current-generation hardware. In response to the accelerating demand, BofA raised its price target on Nvidia to $320 from $300 and also increased targets on Advanced Micro Devices Inc. (NASDAQ:AMD) to $500, Marvell Technology Inc. (NASDAQ:MRVL) to $200, and Micron Technology Inc. (NASDAQ:MU) to $950.

The aggressive analyst revisions are a direct reflection of enormous capital commitments from the world’s largest technology companies. The aggregate cloud capex pool is now expected to reach 95 to 100 percent of operating cash flow from 2026 through 2028 for major providers, a level BofA notes is far higher than the historical 35 to 50 percent ratio and even surpasses the peak spending of the 5G telecom buildout.

Cloud Backlog Exceeds $1.9 Trillion

The spending is backed by staggering contractual obligations. Microsoft Corp. (NASDAQ:MSFT) recently disclosed approximately $627 billion in commercial remaining performance obligations (RPO), up 99 percent year-over-year, which includes an incremental $250 billion contracted from its partner OpenAI. OpenAI itself is reportedly projecting revenues to grow from $30 billion in 2026 to $283 billion by 2030.

Similarly, Amazon.com Inc.’s (NASDAQ:AMZN) AWS division reported a $364 billion backlog, a figure that excludes a recent deal with Anthropic valued at over $100 billion. Alphabet Inc.’s (NASDAQ:GOOGL) Google Cloud saw its backlog nearly double quarter-over-quarter to more than $460 billion. Oracle Corp. (NYSE:ORCL), one of the first recipients of the Vera CPU, reported a 325 percent year-over-year increase in its RPO to $553 billion.

Memory and Optics See Major Upgrades

The performance of next-generation AI systems depends heavily on specialized memory and high-speed optical components, leading Bank of America to upgrade several key suppliers. The firm’s price target on Micron was lifted to $950 from $500, with its forecast for the high-bandwidth memory (HBM) market now reaching $168 billion by 2030, up from just $35 billion in 2025. "We view memory supply elasticity as now structurally lower," Arya wrote, suggesting pricing power will remain with sellers like Micron for years as new factory supply won't arrive until 2027-2028.

Marvell Technology received the second-largest percentage price target increase, moving to $200 from $125. BofA sees the market for ethernet transceivers growing by $10 billion by 2028, with Marvell positioned to capture a significant share due to its dominance in the DSPs required for 800G and 1.6T optical modules. Coherent Corp. (NYSE:COHR) also saw its target lifted to $400 from $365, benefiting from its strong position in transceivers and a new deal to supply high-performance lasers for Nvidia's future optical systems.

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