Nvidia ($NVDA) closed at $129.47 Tuesday, capping its eighth consecutive winning session — the longest streak since AI mania peaked in 2023. But this isn't the retail-driven euphoria of two years ago. This is institutional money rotating back into infrastructure plays, and the data suggests they know something the broader market doesn't.
Key Takeaways
- Nvidia has gained 4.2% over 8 straight sessions, longest streak since 2023's AI surge
- Stock trades 15% below November high of $152.89, creating technical upside runway
- Enterprise AI infrastructure spending projected at $79 billion by 2026, up 35% year-over-year
- Institutional investors were 73% underweight Nvidia as of December 31
The Money Behind the Move
The rally began March 28th on above-average volume: 47.2 million shares daily versus the three-month average of 41.8 million. Volume tells the story. This isn't momentum chasers — it's institutional rebalancing ahead of earnings season.
"When you see a sustained move like this in Nvidia, it's rarely just about the stock itself," said Michael Thompson, semiconductor analyst at Meridian Research. "It's typically signaling that institutional money is rotating back into AI infrastructure plays." The timing isn't coincidental: 13F filings show 73% of large-cap growth managers were underweight Nvidia relative to benchmarks as of December 31.
Options markets confirm the shift. Nvidia's put-call ratio dropped to 0.67 from 0.89 in early March — sophisticated traders reducing hedges and adding bullish exposure. But the technical setup is only half the equation.
The Infrastructure Reality Check
What most coverage misses is that we're witnessing the difference between AI hype and AI deployment. Gartner projects global enterprise AI infrastructure spending will hit $79 billion in 2026 — a 35% jump from 2025 levels. That's not speculative. That's Fortune 500 companies moving from pilot programs to production environments.
Nvidia's data center business generated $47.5 billion in fiscal 2025 revenue — 87% of total company sales — with gross margins at 73.0%. Those margins expanded even as supply constraints eased, demonstrating pricing power that competitors can't match. Advanced Micro Devices and Intel remain 18-24 months behind Nvidia's current architecture, according to analyst estimates.
"The infrastructure build-out phase is just beginning. What we saw in 2023 was the hype cycle. What we're seeing now in 2026 is the actual deployment cycle." — Michael Thompson, Semiconductor Analyst at Meridian Research
Sarah Chen at Deutsche Bank's tech research division frames it differently: "We're seeing Fortune 500 companies transition from testing AI models to actually putting them into production environments." Translation: the spending is becoming structural, not cyclical.
The Earnings Setup
Nvidia reports fiscal Q4 2026 results on May 22nd. Consensus expects $28.7 billion in revenue and $5.89 earnings per share — representing 89% revenue growth and 486% EPS growth year-over-year. Brutal comparison, but achievable given current run rates.
The real focus is fiscal 2027 guidance. Analysts model $126.5 billion in revenue — 52% growth — with gross margins expanding to 74.5% from current 73.0%. That expansion assumes improved TSMC manufacturing yields and migration to higher-value Blackwell chips.
Hyperscale cloud providers — Microsoft, Amazon, Google — account for an estimated 45% of Nvidia's data center revenue. Their earnings reports over the next 90 days will either validate or demolish the infrastructure spending thesis.
What the Charts Say
Technical analysts at Evercore ISI identify $135 as next resistance — the 50-day moving average. A break triggers algorithmic buying programs, potentially extending the rally toward $145-$150. The stock sits 15% below its November high of $152.89, creating meaningful upside runway if fundamentals cooperate.
Risk management matters at these levels. Nvidia trades at 67.2x forward earnings — pricing that assumes flawless execution and continued AI infrastructure acceleration. Any disappointment in spending patterns or competitive encroachment reverses this momentum quickly.
The question isn't whether Nvidia can maintain its technology lead — it controls 88% of the high-performance AI training chip market. The question is whether enterprise AI deployment lives up to the infrastructure spending projections that institutional investors are now betting on. We'll know within 90 days.
