Wells Fargo reiterated its Overweight rating and $265 price target on NVIDIA (NASDAQ: NVDA), following clarification around the company’s relationship with privately held AI hardware firm Groq.
The firm said speculation that NVIDIA was spending more than $20 billion to acquire Groq was resolved after Groq formally announced a non-exclusive licensing agreement with NVIDIA. Wells Fargo noted the agreement effectively functioned as an acqui-hire, with Groq founder Jonathan Ross, President and COO Sunny Madra, and other Groq employees set to join NVIDIA to help scale high-performance, low-cost AI inference globally.
The analyst described the move as strategically focused on latency-sensitive and deterministic AI inference, providing NVIDIA access to Groq’s specialized hardware architecture and compiler software. Wells Fargo also questioned how the deal reflected NVIDIA’s views on high-bandwidth memory, noting Groq’s LPUs rely solely on on-chip SRAM, which the firm said can deliver up to 10 times the performance of HBM.
The note further compared the Groq agreement with NVIDIA’s prior $900 million acqui-hire of Enfabrica, which Wells Fargo viewed as targeting evolving memory architectures and optimization. Additional questions included whether combining NVIDIA’s NVLink C2C interconnect with Groq’s architecture could enhance inference-optimized systems, and how Groq should be positioned relative to NVIDIA’s Rubin CPX platform, introduced in September and designed for large-context inference workloads using 128GB of GDDR7 memory, with availability expected in late 2026.
