Analyst Ratings

Morgan Stanley Names Nvidia, Broadcom and Astera Labs as Its Top Chip Stocks for 2026

The bank is dismissing AI-bubble concerns and betting the next leg of semiconductor growth broadens beyond GPUs into memory and fab utilization — with one small-cap name emerging as its standout pick.

NEW YORK · By Nathan Ellis · Technology & Markets Contributor · Published

Last updated

Nathan Ellis contributes to Global Markets Review as an independent journalist and holds no positions in individual securities covered by the publication.

Morgan Stanley has named Nvidia, Broadcom and Astera Labs as its top semiconductor picks for 2026, arguing in a research note that AI-driven demand for data center chips remains the most important theme in the sector for a third consecutive year. The call comes as skepticism about an "AI bubble" has grown louder across parts of Wall Street, and the bank is using the note to push back directly against that narrative.

The case for Nvidia and Broadcom

Morgan Stanley analyst Joseph Moore, ranked among the top analysts on Wall Street by several industry trackers, set a $250 price target on Nvidia, implying meaningful upside from current levels. Moore has argued that while he understands longer-term skepticism around AI spending, 2026 should remain strong with good visibility extending into 2027 — a dynamic he believes benefits AI semiconductor stocks broadly, with the next leg of growth increasingly favoring memory chips and fab utilization rather than GPUs alone. The firm sees Nvidia delivering the strongest returns in cloud computing specifically as the company's next-generation Vera Rubin platform ramps through the second half of 2026, and argues the market continues to undervalue Nvidia's role in cloud infrastructure more broadly.

Broadcom is Morgan Stanley's other preferred large-cap processor name, with an Overweight rating and a price target near $462 implying roughly a third of potential upside. The bank considers Broadcom central to AI infrastructure buildouts even as it flags potential bottlenecks in the growth of custom ASIC chips — the application-specific processors hyperscalers increasingly use as an alternative to off-the-shelf GPUs. Moore has praised Broadcom's management team as among the most credible in the sector at hitting its own numbers, a factor that matters increasingly to investors wary of overly aggressive AI capital-spending guidance across the industry.

The standout small-cap: Astera Labs

Beyond the two large-cap names, Morgan Stanley singled out Astera Labs as its top small-cap pick for data-center exposure. The company, which builds connectivity chips linking GPUs, memory and networking equipment inside AI data centers, is positioned around an emerging thesis: that hyperscale AI operators will continue relying on third-party connectivity and memory-expansion silicon rather than folding every function into in-house custom chips. Astera's planned rollout of UA Link-based scale-up connectivity solutions in 2026 aligns directly with Morgan Stanley's broader focus on open-standard infrastructure as the next phase of AI data center architecture, reinforcing the bank's view that the company can participate in both scale-up and scale-out segments of the buildout.

Free newsletter

Market Daily

One email before the US open: the overnight moves, the day's earnings, and the one number worth watching. No tips, no hype.

Where the bank is more cautious

Morgan Stanley maintained Equal-Weight ratings on AMD and Marvell, acknowledging room for upside in both names but flagging greater uncertainty relative to its top picks. Intel drew a more cautious view still, with the bank pointing to ongoing challenges in the company's foundry operations as a key overhang. In memory chips, Morgan Stanley named Micron as its preferred name, with Sandisk also carrying an Overweight rating, while Applied Materials and Taiwan Semiconductor topped its list in chip equipment. Among analog chipmakers, the bank favors NXP Semiconductors for its balance of growth and value.

What this means for investors

The note is notable less for naming Nvidia and Broadcom as favorites — both have topped similar lists from multiple banks throughout 2026 — and more for its explicit argument that semiconductor gains are set to broaden into memory and equipment names as the AI buildout matures. For investors already holding concentrated AI-chip positions, Morgan Stanley's inclusion of Micron, Applied Materials and Taiwan Semiconductor alongside its core picks suggests diversifying exposure across the supply chain, rather than concentrating purely in GPU makers, may be the more durable way to play the sector into 2027.

Frequently asked questions

What is Morgan Stanley's Nvidia price target?
Analyst Joseph Moore carries a $250 target on Nvidia, implying meaningful upside from recent levels.
Why does Morgan Stanley like Astera Labs?
It is the bank's top small-cap data-centre pick, on the thesis that hyperscalers keep buying third-party connectivity and memory-expansion silicon rather than building everything in house.
Which chip stocks is the bank cautious on?
AMD and Marvell are Equal-Weight, and Intel draws a more cautious view on continued foundry challenges.

Related reading