Key Takeaways
- Wells Fargo expresses confidence in NVDA before GTC 2026, referencing historical data showing post-event outperformance versus SOX index reaching up to 45% over three months
- Bank of America maintains its Buy rating with a $300 price target, highlighting NVDA’s current valuation at 17x forward PE represents a historical low
- Nvidia plans to showcase its second-generation co-packaged optic switch along with developments in its Feynman GPU line and Kyber NVL576 racks
- Agentic AI applications are creating unprecedented CPU demand — Nvidia’s Vera CPU has entered production with deployment at Meta data centers, targeting full-scale rollout in 2027
- CPU market faces supply constraints, with AMD and Intel reporting lead times extending to six months alongside price increases exceeding 10%
Nvidia (NVDA) approaches its annual GTC conference next week with substantial Wall Street attention focused on the event. Scheduled for March 16–19 in San Jose, California, the conference represents a potential catalyst for the stock and the wider semiconductor industry.
Wells Fargo analysts, under Aaron Rakers’ leadership, stated they are “NVDA buyers ahead of the event.” The firm highlighted historical trends showing robust stock performance during the three-month period following previous GTC conferences, with NVDA delivering approximately 30% average outperformance versus the SOX semiconductor index, spanning a range from +12% to +45%.
Bank of America analyst Vivek Arya maintained a Buy rating alongside a $300 price target. He observed the stock currently trades around 17x forward earnings — approaching a historical low — following a successful Blackwell product ramp that delivered an estimated $500 billion in cumulative sales.
CEO Jensen Huang is scheduled to present a keynote address at 2 p.m. ET on Monday. He will also lead an industry panel on Wednesday afternoon. Companies participating in main stage sessions include OpenAI, Google DeepMind, Meta, Microsoft, and Tesla.
Regarding product announcements, Nvidia is anticipated to introduce its second-generation co-packaged optic switch, incorporating Taiwan Semiconductor’s co-packaged optic technology. Volume production is projected to scale in 2027, targeting approximately 80,000 units. The company may also share progress updates on its Feynman GPU line and the Kyber NVL576 rack system.
Wells Fargo anticipates Nvidia will refresh its pipeline outlook, potentially increasing its cumulative revenue projection from $500 billion to over $600 billion through 2026. Rakers also questioned whether Nvidia will revise its estimate of $3–$4 trillion annually in global AI infrastructure spending by 2030.
CPUs Gain Prominence in AI Infrastructure
Beyond GPUs, a significant transformation is emerging. Agentic AI — task-oriented AI that coordinates workflows across multiple agents — demands a different computational architecture than traditional AI inference. This shift is elevating demand for central processing units to unprecedented levels.
Nvidia’s head of AI infrastructure, Dion Harris, told CNBC this week that “CPUs are becoming the bottleneck in terms of growing out this AI and agentic workflow.” The company’s Vera CPU is in production and is already deployed at Meta data centers as part of a multiyear deal announced in February. Nvidia plans to expand that deployment in 2027.
Thousands of standalone Nvidia CPUs are also running at the Texas Advanced Computing Center and Los Alamos National Lab. Bank of America projects the CPU market could more than double, from $27 billion in 2025 to $60 billion by 2030.
At GTC, Nvidia is expected to display a CPU-only rack on the showroom floor — demonstrating the company’s commitment to standalone CPU deployments.
Supply Constraints Intensify
The broader CPU market faces mounting pressure. AMD and Intel have both alerted customers to supply shortages, with delivery lead times extending up to six months and prices climbing more than 10%, according to Reuters.
AMD’s head of data center Forrest Norrod told CNBC that demand increases over the past six to nine months have been “unprecedented.” Intel said inventory is expected to reach its lowest level this quarter, though the company anticipates supply improvement through Q2 2026.
Currently, Nvidia reports it has avoided significant CPU shipment delays. Harris said the company’s supply chain has successfully managed demand, partially because most of its CPUs ship alongside GPUs in full rack-scale systems.
Mercury Research estimates Nvidia held a 6.2% share of the server CPU market in Q4 2025, trailing Intel at 60% and AMD at 24.3%. Other stocks that could experience movement based on GTC announcements include AMD, Taiwan Semiconductor, Broadcom (AVGO), Intel, and Marvell (MRVL).
