Key Takeaways
- Rental costs for Nvidia’s H100 GPUs have jumped nearly 40% since October, climbing from $1.70 to $2.35 per hour as available GPU capacity has essentially vanished from the market.
- Even Nvidia’s latest Blackwell GPU lineup faces severe supply constraints, with delivery schedules now pushed out to mid-2026, contrary to predictions that newer technology would relieve demand for previous-generation hardware.
- The top four cloud providers — Alphabet, Microsoft, Meta, and Amazon — plan to allocate approximately $700 billion toward AI infrastructure investments in 2026, with Nvidia commanding 85-90% of the GPU marketplace.
- Recent Chinese regulatory approval for H100 chip distribution, with licenses granted to multiple buyers, creates a potential $25 billion yearly revenue stream that hasn’t been factored into current forecasts.
- NVDA shares are valued at 15.7x forward earnings — beneath the 3-year mean of 19.4x — while analyst consensus price targets average $273.57, suggesting approximately 55% potential appreciation.
The cost to rent Nvidia’s H100 graphics processors continues climbing rather than declining, with hourly rates jumping nearly 40% since October — from $1.70 to approximately $2.35 per hour. This pricing data comes from SemiAnalysis, which surveyed over 100 industry stakeholders to track market trends.
Available GPU inventory has essentially disappeared from the market. Organizations that secured capacity early continue retaining their allocations despite escalating costs. Some enterprises are turning to premium-priced spot instances on platforms such as AWS just to obtain access.
This supply-demand imbalance extends beyond legacy hardware. Nvidia’s latest Blackwell GPU series confronts identical constraints, with current lead times extending into mid-2026. This contradicts earlier industry assumptions that more powerful, energy-efficient next-generation processors would reduce demand — and consequently pricing — for products like the H100.
The surge in requirements stems from diverse AI workloads, spanning media-creation platforms at organizations like ByteDance and Google to expanding deployment of language models such as Anthropic’s Claude.
Cloud Giants’ Capital Commitments Establish Revenue Baseline
Fueling the capacity shortage is an unprecedented capital commitment wave. Alphabet, Microsoft, Meta, and Amazon collectively plan to invest approximately $700 billion in AI infrastructure throughout 2026. These represent confirmed spending allocations, not speculative forecasts.
Microsoft has indicated that roughly two-thirds of its infrastructure expenditures target GPUs and CPUs. Given Nvidia’s 85-90% dominance of the GPU sector, the majority of these dollars flow to Nvidia. Even assuming semiconductors represent merely 20% of total AI infrastructure expenses, this translates to over $140 billion in annual processor purchases from just these four major customers.
Nvidia reported Q4 revenues of $68.13 billion, representing 73% year-over-year expansion, while its Q1 forecast of $78 billion exceeded analyst expectations by more than $5 billion. Fiscal 2027 revenue growth currently stands at an estimated 71%.
Nevertheless, NVDA shares have declined roughly 6.5% year-to-date, pressured by broader macroeconomic uncertainties related to energy inflation and general market risk aversion. The stock presently trades at 15.7x forward earnings — below both its 3-year historical mean of 19.4x and AMD’s forward multiple of 18.9x, despite Nvidia maintaining substantially superior market position, profit margins, and expansion rates.
Chinese Market Access and Vera Rubin Platform Expansion Offer Additional Catalysts
One potentially significant revenue driver absent from current projections: Chinese authorities have granted Nvidia authorization to distribute H100 processors, with licenses awarded to several purchasers. Wells Fargo analysts estimate this opportunity could generate $25 billion or more annually. This revenue potential wasn’t incorporated into Nvidia’s latest financial guidance.
Regarding product development, the forthcoming Vera Rubin platform provides ten times greater performance per watt than Blackwell and approximately 50 times more tokens per watt versus the previous Hopper architecture. Shipments are scheduled to commence during the second half of 2026.
Nvidia recently finalized a $2 billion equity stake in Marvell Technology (MRVL) on March 31, expanding its NVLink ecosystem to accommodate Marvell’s specialized AI processors — currently utilized by Amazon, Alphabet, and Microsoft. NVDA shares rose over 5% following that announcement. Marvell surged 13%.
Analyst consensus on NVDA remains at Strong Buy, with 41 Buy recommendations, one Hold, and one Sell rating issued over the past three months. The mean price target stands at $273.57.
