Nvidia’s Record Quarter Signals Momentum, But Growth Outlook Cools

What happened

Nvidia posted another record quarter but the data-center line landed a touch light versus the Street, and the next-quarter revenue guide—while slightly above consensus—looked cooler compared with prior blowout ramps.

Key numbers

• Total revenue (Jul qtr): $46.7B (record; roughly in line)
• Data-center revenue: $41.1B (vs. ~$41.3B expected; ~89% of sales)
• Net income: $26.4B (+59% y/y)
• Next-qtr guide: ~$54B (slightly above consensus)

Why it matters

After a multi-quarter surge, investors are hypersensitive to any sign that AI compute demand is normalizing. Small misses and modest guides can resize expectations in a stock that’s priced for exceptional growth.

Drivers

• Hyperscalers and AI labs (OpenAI, Microsoft, Amazon, Alphabet, Meta) continue building training and inference capacity.
• Blackwell GPUs are ramping; management called demand “extraordinary.”
• China headwinds (export limits/H20 halts) offset by strength elsewhere; optionality if an approved, reduced-capability Blackwell variant ships into China.

China & policy overhang

• H20 sales paused/restricted; booked ~$180M in legacy H20 to China and ~$650M to a non-China customer.
• If approvals land, finance expects $2B–$5B potential data-center upside in the coming quarter.

Strategy positioning

Nvidia has evolved into a full-stack AI infrastructure company—GPUs, systems, networking, and software—focused on maximizing data-center throughput per dollar. Jensen Huang estimates AI leaders will spend $3T–$4T over five years; Nvidia targets a dominant slice.

Risks to watch

• Export controls/geopolitics; timing of China-eligible products
• Supply cadence vs. cloud deployment schedules
• Competition (AMD/custom silicon), and market sensitivity to small quarterly variances

Bottom line

The long-term AI build-out remains intact. Near-term, the stock trades on increments in data-center momentum and China visibility. Expect more quarter-to-quarter chop; execution on Blackwell ramps and inference software should drive the next leg.

Correction note: Management now estimates hyperscaler AI capex of $3T–$4T over five years and suggested Nvidia could capture as much as 70% of that revenue (prior 35% reference was corrected on Aug. 27).