Morgan Stanley sees more gains ahead for Nvidia and Broadcom , as the momentum in artificial intelligence continues. The bank reiterated its overweight thesis on AI poster child Nvidia. Morgan Stanley’s $250 price target, up from $235, corresponds to 41% upside from Nvidia’s Friday closing price. The chipmaker has gained 32% in 2025. “We continue to see NVIDIA maintaining dominant market share, as threats are becoming overstated, though we aren’t sure exactly what will turn sentiment around,” analyst Joseph Moore wrote. “Customers’ biggest anxiety for the next 12 months is their ability to procure enough NVIDIA product generally, and Vera Rubin specifically.” NVDA YTD mountain NVDA YTD chart Moore said that his multiple assumption still trades at a discount to peer Broadcom but at a premium to the broader semiconductor group, “albeit a narrowing one as the absolute level of margins and revenue make multiple expansion more challenging.” The analyst also stood by his overweight rating for Broadcom and raised his price target to $443 from $409. This revised forecast is approximately 10% higher than where shares of Broadcom closed on Friday. Broadcom stock has gained 74% this year. AVGO YTD mountain AVGO YTD chart Moore highlighted Broadcom’s large AI exposure as a positive, also applauding the company’s growth potential. He pointed to Broadcom’s tensor processing unit, or TPU, as a tailwind. “Google’s Tensor processor supply chain — designed and sold by AVGO — is getting revised higher, but this is slightly at the expense of other Broadcom customers — and the homegrown version is also a major focus,” he wrote. “Even before the positive reviews of Gemini created the current wave of enthusiasm, we heard multiple sources of TPU being revised up, from analog companies, from memory companies, and from ODM partners.” However, he added the caveat that this is replacing other chip expectations for Broadcom. “Specifically, we believe that MTIA builds for customer Meta — still volume expected 2H26 — have been deferred somewhat, and replaced by Meta’s expected use of TPU,” he added. “Unconfirmed press reports have Meta, as well as Open AI, using TPU, at least in part to gain familiarity with ASIC usage for eventual migration to internal ASIC.”
