NVIDIA Corp Earnings - Q2 2027 Analysis & Highlights
NVIDIA reported record Q2 2027 results driven by extraordinary AI infrastructure demand, with the company guiding 70% revenue growth for fiscal 2028 despite supply constraints, while expanding its full-stack AI factory platform and making significant strategic investments in frontier AI labs.
Key Financial Results
Total revenue reached $96 billion, more than doubling year-over-year with growth accelerating for the fourth consecutive quarter.
Record operating income and EPS were delivered in the quarter.
GAAP and non-GAAP gross margins were both 75%, largely unchanged from the prior quarter due to similar product mix.
GAAP and non-GAAP operating expenses increased 10% and 11% sequentially, primarily due to high compute infrastructure costs and compensation and benefits costs.
Non-GAAP effective tax rate of 16% increased from a year ago primarily due to higher revenue.
Record $26 billion returned to shareholders in Q2, consisting of $20 billion through share repurchases and $6 billion through quarterly dividend of $0.25 per share.
Year-to-date shareholder returns of 60% of free cash flow, with intention to increase and return excess free cash flow net of strategic uses going forward.
Business Segment Results
Data-center revenue increased 18% quarter-over-quarter to $89 billion with strong contributions from both Hyperscale and ACIE sub-segments.
Hyperscale revenue of $49 billion grew 13% sequentially, driven by sustained strength in Blackwell as new GPU capacity comes online.
ACIE revenue of $40 billion increased 25% sequentially and 138% year-over-year, driven by neocloud capacity additions and demand from enterprises, AI startups, and sovereigns.
Networking business achieved record quarter with revenue growing 18% sequentially, with Spectrum-X Ethernet growing 2.6x year-over-year.
Grace CPU revenue on trailing 12-month basis exceeded $5 billion, with the company now in full production of next-generation Vera CPU.
Sovereign AI business grew 35% sequentially and more than tripled year-over-year in Q2, primarily through regional neoclouds.
Enterprise on-prem revenue in automotive vertical reached $8 billion on trailing 12-month basis, while financial services, manufacturing and healthcare combined contributed $7 billion.
Capital Allocation
Inventory increased to $32 billion as the company prepared for the Vera Rubin launch.
Days of sales outstanding increased to 60 days, reflecting extended payment terms for large purchases by certain investment grade customers to be shipped over multiple quarters.
Nearly $50 billion invested in frontier AI labs, representing a meaningful commitment but a small fraction of expected free cash flow over the same period.
Partnerships established with six leading infrastructure capital providers (Apollo, BlackRock, Blackstone, Brookfield, Goldman Sachs, and KKR) to establish financing platforms that will raise over $500 billion of third-party capital.
Partnership with SoftBank Energy announced to exclusively host NVIDIA compute at Portsmouth campus, with initial deployment expected to support 4.25 gigawatts of AI factory capacity for OpenAI.
Selective credit enhancement provided for nearly 2 gigawatts of compute for another frontier AI lab.
Industry Trends and Dynamics
Surge in AI demand driving global infrastructure buildout supported by expanding and diverse set of growth opportunities spanning hyperscalers, AI labs, AI natives, enterprises and sovereign customers.
Cloud industry backlog now greater than $2 trillion, with CapEx by top five hyperscalers expected to reach nearly $800 billion in 2026 and $1.3 trillion in 2027.
Hyperscale customers delivered strong financial results with accelerating revenue growth and expanding margins.
Neocloud partners expected to exit year with 8 gigawatts in total installed capacity, up from approximately 3 gigawatts at end of 2025.
Global VC funding in AI exceeded $400 billion in first half of 2026, surpassing $265 billion raised in all of 2025, with roughly 70% spent on compute.
Nearly 20 AI native companies now exceed $1 billion in annualized run rate revenue, up from 13 companies in Q4 of last year, with vertical enterprise software logging fastest growth.
Agentic AI adoption is accelerating, with demand for data-center CPUs rising as AI agents run continuously in the background.
Both closed and open models are skyrocketing in use, with nearly all open models running on NVIDIA.
Competitive Landscape
NVIDIA's architecture runs every model with the company growing share as closed and open model adoption grows.
NVIDIA runs leading closed models including OpenAI, Anthropic, Grok, Meta, and Gemini, as well as leading open models including TML, Mistral, Qwen, Kimi, GLM, DeepSeek, MiniMax and Nemotron.
NVIDIA is great at training, inference, and agentic workloads with one platform fungible for every model and workload, durable for entire life cycle of AI.
Combination of performance, fungibility and durability makes NVIDIA the productive and financeable compute infrastructure.
Full-stack AI factory platform expanding share of data-center TAM, with revenue opportunity per gigawatt growing from $18 billion with Hopper to $25 billion with Blackwell to $40 billion with Vera Rubin.
Vera Rubin delivering 30x higher throughput per megawatt and 35x lower token cost relative to Grace Blackwell Ultra.
Groq 3 LPX demonstrating nearly 4x the number of tokens per second against next best alternative on artificial analysis benchmark.
CUDA ecosystem extends AI into markets a single chip alone cannot reach, allowing NVIDIA to help sovereigns, neoclouds and enterprises build AI infrastructure.
Frontier AI labs' growth limited by compute availability rather than technology or customer demand, with NVIDIA positioned as essential partner to help power their flywheel.
Macroeconomic Environment
Extreme pricing conditions in memory with magnitude of price increase exceeding prior expectations and headed even higher into next year.
Memory scarcity being driven in large part by AI buildout itself, with tighter memory supply a symptom of same demand surge driving NVIDIA's growth.
Ongoing geopolitical uncertainty with no China data-center compute revenue in forward outlook.
Less than 1% of total data-center revenue in Hopper 200 products shipped to customers based in China in Q2 in accordance with US Government licenses.
Growth Opportunities and Strategies
AWS partnership expansion announced with AWS deploying additional 2 million GPUs starting this quarter through Q2 fiscal 2029 along with Vera CPUs.
AWS adopting full physical AI stack including Omniverse, Cosmos, Isaac and Jetson to power fleet of warehouse robots.
AWS will serve NVIDIA Nemotron family of open models on Amazon Bedrock and SageMaker.
Vera Rubin expected to mark fastest product ramp in NVIDIA's history, having already received purchase orders from every major hyperscaler, AI cloud and system OEM.
Vera CPU expected to be deployed by every major hyperscaler, neocloud, AI lab and system OEM, with shipments already underway to lead partners including OCI, SpaceX AI and AWS.
CPU revenue expected to more than double in fiscal 2028, positioning NVIDIA as one of world's leading server CPU suppliers.
Groq 3 LPX expected to ship in volume later this quarter to early adopters, with Nebius being the first.
Revenue sharing structure introduced for neoclouds where NVIDIA provides take-or-pay commitment on portion of facility capacity and shares in portion of neocloud's revenue earned above floor.
Model can expand addressable market and create recurring usage-linked revenue stream with potential to drive billions in revenue over medium to long term.
Three unique capabilities reinforcing one another as engines of growth: platform running every model, full-stack AI factory platform capturing more of data center TAM, and CUDA ecosystem extending AI into markets no single chip could reach.
Partnership with Noetra announced to build NVIDIA DSX AI factory that will create open models to power AI agents, digital twins, robotics and physical AI applications.
South Korea's LG and Hyundai Motor Group partnering with NVIDIA to build and scale AI.
Record 35 new NVIDIA powered AI supercomputers unveiled in Europe to advance industry and scientific breakthroughs.
Financial Guidance and Outlook
Fiscal 2028 revenue expected to grow approximately 70% year-over-year, representing supply-constrained outlook.
Q3 total revenue expected to be $108 billion, plus or minus 2%.
Sequential growth in Q3 expected to be driven primarily by ACIE, while growth in hyperscale expected to re-accelerate in Q4 and into fiscal 2028 as Vera Rubin supply grows.
Vera Rubin expected to account for about 20% of data-center revenue in Q3.
Supply expected to remain bottleneck at least through end of fiscal 2028, though company will work to close supply/demand gap.
Q3 GAAP and non-GAAP gross margins expected to be 74% plus or minus 50 basis points.
Margins expected to bottom in Q4 in 71% to 72% range before settling at 72% to 73% in fiscal 2028 as executed price increases take effect in Q1.
GAAP and non-GAAP operating expenses expected to be approximately $9.2 billion and $9.0 billion respectively for Q3.
Full-year OpEx expected to grow in low 50s, driven by broadening of product portfolio and further increase in usage of AI tools.
Full-year fiscal 2027 GAAP and non-GAAP tax rates expected to be between 16% and 18%, excluding discrete items and material changes to tax environment.
Demand from AI labs for which company expects to leverage balance sheet expected to contribute toward roughly a quarter of business next year.
OpenAI committed to substantial deployments of NVIDIA AI infrastructure through 2030, with existing and planned commitments representing approximately 12 gigawatts of NVIDIA compute.
AI Infrastructure and Platform Strategy
NVIDIA providing full-stack AI factory platform spanning entire AI life cycle from data preparation and pre-training through post-training to agentic inference.
Revenue opportunity per gigawatt increased from $18 billion with Hopper to $25 billion with Blackwell to $40 billion with Vera Rubin, with productivity representing X factor increase each generation.
Vera Rubin completing agentic tasks 1.8x faster on SPEC benchmark and providing 5 times bandwidth per watt than any other data-center CPU.
Vera Rubin production shipments commenced earlier this month with purchase orders already received from every major hyperscaler, AI cloud and system OEM.
Using NVIDIA DSX reference designs, neocloud partners bringing capacity online faster and at lower token cost.
NVIDIA compute fully utilized across every cloud served, with economic value it generates for hyperscale, neocloud and AI lab partners continuing to rise.
Demand and Supply Dynamics
Unconstrained demand significantly exceeds 70% guidance, with company growing 100% year-over-year this year.
Entire supply chain challenged with everybody running flat out, with more capacity coming online all the time.
Yields being improved and company working hard with every supplier to increase capacity that roadmap requires.
Customers' forecasts point to growth doubling next year, but company expects to grow approximately 70% as supply constrained.
Demand acceleration occurring even at NVIDIA's scale, with NVIDIA compute productive, fungible, rentable and durable.
Agentic AI requiring 15 to 100 times more compute than human using it depending on type of problem being solved.
Large language models larger than ever because they're smarter than ever, with agents going through reasoning, planning and multiple turns of tool use.
