TLDR
- Nvidia (NVDA) reported Q2 revenue of $46.7 billion, up 56% year-over-year, beating Wall Street estimates
- Company posted zero H20 processor sales to China during the quarter due to US export controls
- Data center revenue reached $41.1 billion but fell short of analyst expectations at $41.34 billion
- Stock dropped 3.3% in after-hours trading despite beating earnings and revenue forecasts
- Board approved additional $60 billion in share repurchases with no expiration date
Nvidia delivered another quarter of robust financial performance, reporting second-quarter revenue of $46.7 billion that exceeded Wall Street expectations. The world’s largest publicly traded company by market cap posted earnings per share of $1.05, surpassing the $1.01 analyst estimate.
NVIDIA Q2 2025 EARNINGS:
– $46.7B revenues, +56% YoY
– $26.4B net income, +59% YoY
– 72% gross margins, +20% QoQ
– $60B buybacks
– $54B revenue guide Q3The most exciting thing about this quarter is that Nvidia did NOT include China sales in guidance.
Outstanding. pic.twitter.com/HohSwsHDVQ
— amit (@amitisinvesting) August 27, 2025
Revenue climbed 56% compared to the same quarter last year when the company reported $30.04 billion. This marks the ninth consecutive quarter where year-over-year revenue growth has topped 50%, a streak that began in mid-2023 with the start of the AI boom.
The quarter represented sequential growth of 6% from the previous period. Net income jumped 59% to $26.42 billion, translating to $1.05 per share compared to 67 cents per share in the year-ago period.
Despite the strong numbers, Nvidia shares fell 3.3% in after-hours trading on Wednesday. The decline came as investors focused on data center revenue missing expectations.
Data Center Business Drives Results
Nvidia’s data center division generated $41.1 billion in revenue, falling just short of the StreetAccount estimate of $41.34 billion. The segment still posted 56% growth from the previous year.
CFO Colette Kress broke down the data center numbers further during the earnings call. She said $33.8 billion came from compute sales, primarily GPU chips, which declined 1% from the first quarter.
The networking portion of the data center business brought in $7.3 billion. This segment nearly doubled compared to the same quarter last year as customers bought more parts needed for complex AI systems.
Blackwell chip sales, representing Nvidia’s latest generation of processors, rose 17% from the previous quarter. The company previously disclosed that Blackwell products reached $27 billion in sales and now account for roughly 70% of data center revenue.
China Export Controls Impact Operations
The company confirmed it sold zero H20 processors to China-based customers during the quarter. The H20 represents a modified version of Nvidia’s flagship H100 chip, specifically designed for the Chinese market to comply with US export restrictions.
Export controls imposed by the Trump administration brought H20 sales to China to a complete stop. The restrictions included export licenses and fees totaling approximately $5.5 billion.
However, Nvidia did benefit from releasing $180 million worth of H20 inventory to a customer outside of China during the quarter. Kress indicated the company could potentially ship between $2 billion and $5 billion in H20 revenue in future quarters if geopolitical conditions allow.
The gaming division reported $4.3 billion in sales, representing 49% growth from the year-ago period. This segment previously served as Nvidia’s largest revenue source before AI demand transformed the company’s business mix.
Nvidia’s robotics division posted $586 million in sales during the quarter, showing 69% annual growth. While management has highlighted robotics as a future growth opportunity, it remains a small portion of overall business.
The company’s board authorized an additional $60 billion in share repurchases with no expiration date. Nvidia bought back $9.7 billion of its own stock during the quarter.
Looking ahead, Nvidia expects third-quarter revenue of $54 billion, plus or minus 2%. This guidance excludes any potential H20 shipments to China and exceeds analyst expectations of $53.1 billion.
Kress told analysts that Nvidia anticipates between $3 trillion and $4 trillion in global AI infrastructure spending by the decade’s end. Large cloud providers currently represent about half of Nvidia’s data center customer base, with companies like Meta, Alphabet, Microsoft, and Amazon investing tens of billions quarterly in AI infrastructure buildouts.
The quarter’s profit margin reached approximately 72.4%, reflecting the company’s pricing power in the AI chip market. Nvidia maintains its position as the dominant supplier of processors needed for training and running artificial intelligence models.
