Wall Street Analysts Double Down on Nvidia: Why "Buy" Ratings Persist Despite Valuation Fears and Rising Competition in 2025
- Why Are Analysts Ignoring Nvidia’s Sky-High Valuation?
- Google’s TPU Threat: Overblown or Overdue?
- The $13B Wild Card: Nvidia’s 2026 CPX Chip
- China Risks: A $275 Speed Bump?
- FAQ: Nvidia’s 2025-26 Outlook
Despite swirling concerns about stretched valuations and intensifying competition from Broadcom, AMD, and Google’s TPUs, Wall Street’s top analysts remain steadfast in their bullish outlook for Nvidia (NVDA). Bank of America, Bernstein, and Jefferies all reaffirmed their "Buy" ratings this week, with price targets as high as $275. The consensus? Nvidia’s upcoming Blackwell and Rubin chips—paired with a $500B+ demand pipeline—could cement its dominance in AI infrastructure through 2026. Here’s why the smart money isn’t backing down.
Why Are Analysts Ignoring Nvidia’s Sky-High Valuation?
At first glance, Nvidia’s 25x 2026 P/E ratio seems frothy—until you consider its PEG ratio of 0.5, dwarfing the "Magnificent Seven" average of 2x. Bank of America’s Vivek Arya (ranked #270 on TipRanks with a 58% win rate) argues the stock’s still cheap given its tech moat: "Blackwell’s 2024 architecture isn’t even comparable to today’s Hopper-based GPUs," he noted after meeting Nvidia’s IR team. Management claims Blackwell-powered LLMs (due early 2026) will deliver generational leaps in tokens-per-watt, with MLPerf benchmarks already backing those claims. Arya’s $275 target implies 30% upside, banking on $500B+ in combined Blackwell/Rubin demand—a figure thatrecent OpenAI and Anthropic LOIs.
Google’s TPU Threat: Overblown or Overdue?
Bernstein’s Stacy Rasgon (#144 rank, 67% accuracy) tackled the elephant in the room: Google’s Gemini 3 reportedly matches Nvidia’s TPU performance. "The street’s missing two key points," Rasgon countered. First, cloud providers face steep switching costs—Nvidia’s programmable platforms still dominate 83% of AI infrastructure (per TradingView data). Second, Blackwell’s 2026 rollout puts Nvidia "2 years ahead of Google’s roadmap," per management. Even if TPUs gain traction, Rasgon estimates(NVLink 6, due late 2026) could lock in hyperscalers through 2027.
The $13B Wild Card: Nvidia’s 2026 CPX Chip
Jefferies’ Blayne Curtis (#58 rank, 64% win rate) spotlighted an underappreciated catalyst: Nvidia’s inference-focused CPX chip launching late 2026. "This isn’t just about training anymore," Curtis told clients, projecting $13B in 2027 CPX revenue as hyperscalers ramp inferencing spend. His $250 target assumes Blackwell-based LLMs (early 2026) and Vera-Rubin GPUs (mid-2026) will sustain 17-20% annual EPS growth. "At 18x forward earnings, NVDA’s priced like atech stock—not the AI arms dealer it is," he quipped.
China Risks: A $275 Speed Bump?
All three analysts downplayed China export fears, noting H200 shipments (25% discounted vs. U.S. prices) represent
FAQ: Nvidia’s 2025-26 Outlook
What’s driving Nvidia’s $500B demand projection?
The $500B combines Blackwell GPU orders, Rubin CPU demand, and networking solutions—excluding recent OpenAI/Anthropic deals that could add billions more.
How does Nvidia’s valuation compare to peers?
At 25x 2026 earnings, NVDA trades at half the PEG ratio of mega-cap tech peers (0.5 vs. 2.0 average).
When will Blackwell chips impact revenue?
First LLMs using Blackwell are expected early 2026, with full-scale adoption likely by 2027.
