Morgan Stanley Reinstates Nvidia as Top Pick Amid AI Infrastructure Surge
Key Takeaways
- Morgan Stanley has named Nvidia its top semiconductor pick for 2026, citing a significant disconnect between the company's record-breaking fundamentals and its stagnant stock performance.
- Analyst Joseph Moore highlighted $660 billion in projected AI infrastructure spending and long-term prepaid contracts as evidence that the AI growth cycle is far from peaking.
Mentioned
Key Intelligence
Key Facts
- 1Morgan Stanley set a $260 price target for NVDA, implying 47% upside from current levels
- 2Hyperscalers are projected to spend over $660 billion on AI infrastructure in 2026
- 3Nvidia's earnings expectations for the current quarter rose 38% over the past six months
- 4Hyperscalers are signing three-year supply contracts with full upfront prepayments
- 5Nvidia currently trades at approximately 18 times projected 2027 earnings
| Metric | ||
|---|---|---|
| Status | New Top Pick | Former Top Pick |
| Recent Performance | Down 3% YTD (2026) | Up 300-900% (period) |
| Valuation Basis | 18x 2027 Earnings | Sustainability Concerns |
Analysis
The semiconductor landscape is witnessing a strategic pivot as Morgan Stanley reinstates Nvidia (NVDA) as its top industry pick, displacing Micron Technology (MU). This move by lead analyst Joseph Moore signals a conviction that the market has unfairly discounted the AI giant's long-term trajectory. Despite a 3% decline in share price during the early months of 2026, Nvidia's underlying business metrics have continued to accelerate, creating what Moore describes as a "surprisingly good entry point" for institutional investors. The core of the thesis rests on a stark divergence: while Nvidia’s earnings expectations for the current quarter have been revised upward by 38% over the last six months, the stock has remained largely range-bound, trading at a modest 18 times projected 2027 earnings.
This valuation reset comes after a period where memory-focused stocks like Micron enjoyed astronomical gains of 300% to 900%, leading analysts to question the sustainability of further upside in the memory sector. In contrast, Nvidia's relative underperformance in late 2025 and early 2026 has de-risked the stock, even as the demand for its H-series and Blackwell-generation GPUs remains insatiable. The broader SaaS and cloud ecosystem is currently anchored by the capital expenditure of "hyperscalers"—the likes of Microsoft, Google, and Amazon—who are projected to funnel over $660 billion into AI infrastructure throughout 2026.
If Nvidia can maintain its current pace of innovation while the $660 billion in infrastructure spending materializes, the $260 price target set by Morgan Stanley—representing a 47% upside—may prove conservative.
One of the most compelling data points provided by Morgan Stanley involves the nature of these infrastructure investments. Rather than cautious, quarter-by-quarter purchasing, hyperscalers are increasingly locking in three-year supply contracts. Crucially, some of these agreements involve full upfront prepayments, a "durability signal" that suggests the AI spending cycle is not the ephemeral bubble some skeptics feared. These prepayments provide Nvidia with a level of revenue visibility that is rare in the historically cyclical semiconductor industry. It effectively counters the narrative that the AI build-out will hit a "digestion phase" in the second half of 2026.
What to Watch
Furthermore, the analyst addresses the looming threat of custom silicon. While hyperscalers are developing their own AI accelerators (ASICs) to reduce dependency on Nvidia, the transition is proving slower and more complex than anticipated. Nvidia’s software moat—specifically its CUDA platform—continues to make its hardware the default choice for developers deploying large language models. The integration of hardware and software remains a significant barrier to entry for rivals like AMD or internal chip projects at major cloud providers. Moore's report suggests that the "ASIC threat" is currently more of a talking point than a material headwind to Nvidia's market share.
Looking ahead, the market's focus will likely shift toward the execution of the Blackwell ramp-up and the subsequent product roadmap. If Nvidia can maintain its current pace of innovation while the $660 billion in infrastructure spending materializes, the $260 price target set by Morgan Stanley—representing a 47% upside—may prove conservative. For the SaaS and Cloud sectors, this reinforces the reality that the "picks and shovels" of the AI era are still in high demand, and the foundational layer of the cloud stack is being rebuilt for a multi-year AI-first cycle. The shift in "Top Pick" status from memory to compute reflects a broader market realization: while memory is essential, the high-margin value remains concentrated in the logic and processing power that Nvidia uniquely provides at scale.