Morgan Stanley has reinstated Nvidia as its top pick among semiconductor stocks, arguing that the next constraint on artificial intelligence is no longer just the chips themselves, but the power and capital needed to build and run the data centers that host them.
The bank's analysts say AI demand remains strong, but the industry is hitting a new ceiling: how quickly companies can construct and energize data centers and secure the financing to do so. In this environment, buyers are shifting focus from raw chip counts to what the bank calls “AI output per watt” — essentially, how much useful computing work can be squeezed from a fixed electricity budget.
Why power and financing now matter more
For years, the AI boom has been measured in GPU shipments and wafer supply. But as data center buildouts accelerate, electricity availability and grid connection timelines have become the gating factors. Utilities, permitting processes, and the sheer scale of power required for AI workloads are slowing projects, even when chips are available.
Morgan Stanley argues that Nvidia has an edge in this new reality. If its systems can generate more “tokens” — the chunks of text or data an AI model produces — per gigawatt of power, customers can scale capacity faster without waiting for additional grid connections. That efficiency advantage, the bank says, could become a key differentiator as the industry grapples with energy constraints.
The bank also highlights Nvidia's diverse customer base, which includes Big Tech “hyperscalers,” newer cloud providers, governments, and enterprises. This mix gives Nvidia multiple pools of demand, even as some projects face delays from permitting, power hookups, or financing hurdles.
What the numbers say
Morgan Stanley's analysis points to a significant gap between supply and demand. The bank says Nvidia's fiscal 2028 outlook for roughly 70% revenue growth looks “supply-constrained,” with underlying demand closer to 100%. In other words, the company could sell far more than it can produce, even with its aggressive expansion plans.
The note also flags that Nvidia's stock trades at around 15 times its forecast fiscal 2028 earnings. That valuation, the bank suggests, reflects a market that is beginning to price in the power and financing constraints as much as the chip supply story.
What it means for investors
For everyday investors, Morgan Stanley's stance reframes how to think about Nvidia's prospects. The stock is no longer just a bet on GPU sales; it's increasingly a bet on the company's ability to deliver more usable AI capacity per unit of electricity. That shifts the metrics worth watching.
Instead of focusing solely on wafer supply or unit shipments, investors may want to track evidence of efficiency improvements and faster deployment times. If Nvidia can demonstrate better performance-per-watt, that could support earnings forecasts and justify a higher valuation. Conversely, if power constraints slow customer adoption, even strong chip sales might not translate into the expected revenue growth.
The broader backdrop is also relevant. Rising bond yields from AI data center financing could raise the cost of capital for the entire sector, making efficiency and time-to-deploy even more critical. Meanwhile, major cloud providers are investing heavily in power and water infrastructure to ease bottlenecks, a sign that the industry is taking these constraints seriously.
For Nvidia specifically, the bank's view suggests that the company's competitive moat extends beyond chip design. Its ability to integrate hardware, software, and system-level efficiency could be what sets it apart as the AI buildout enters a more energy-constrained phase.
The bottom line
Morgan Stanley's upgrade is a reminder that the AI trade is evolving. The next leg of growth may hinge less on how many chips Nvidia can ship and more on how much AI output it can deliver per watt of power. For investors, that means paying attention to efficiency metrics, customer deployment pace, and the broader energy and financing landscape.
As always, no single analyst call should drive an investment decision. But understanding the shifting dynamics of the AI supply chain can help investors make more informed choices about their exposure to this fast-moving sector.
