NVIDIA restructured its reporting framework in Q1 2027, signalling its transition from a graphics card company to the structural backbone of the global AI economy. Backed by a large installed base and an extensive ecosystem of CUDA applications, the company is positioned to capture robust demand for AI infrastructure across different industries and applications. By delivering the industry’s lowest token cost, highest throughput, and maximum returns, NVIDIA ensures its platform remains the standard for profitable, enterprise-scale AI deployment.
| Segmental Revenue | Q1 2026 (US$ million) | Q1 2027 (US$ million) | Percentage change |
| Data Centre | 39,112 | 75,246 | +92.4% |
| Hyperscale | 17,599 | 37,869 | +115.2% |
| AI Clouds, Industrial, & Enterprise | 21,513 | 37,377 | +73.7% |
| Edge Computing | 4,950 | 6,369 | +28.7% |
| Total | 44,062 | 81,615 | +85.2% |
Revenue surged 85.2% year-on-year and 19.8% sequentially to US$81.6 billion, driven by surging inference demand despite not generating any revenue from China. The Blackwell architecture is ramping across a broad customer base of hyperscalers, model makers, AI clouds, and sovereign states. In fact, NVIDIA achieved year-on-year revenue acceleration for three consecutive quarters, while also delivering sequential revenue growth for fourteen straight quarters. During the quarter, the company also incurred higher compensation, compute, and infrastructure costs.
Data centre revenue increased 92.4% year-on-year and 20.8% sequentially to US$75.2 billion.
- Computing revenue surged US$60.4 billion 77% year-on-year
- Networking revenue nearly tripled to US$14.8 billion year-on-year.
Management returned US$20 billion to shareholders in Q1 2027 and plans to return approximately 50% of free cash flow to shareholders this year. Quarterly dividend per share rose from US$0.01 to US$0.25. Management announced an additional US$80 billion share buyback programme on top of its existing US$39 billion plan.
Revenue in Q2 2027 is expected to grow sequentially to around US$91 billion, driven primarily by sustained strength in data centre revenue excluding China. The company has in fact secured US$145 billion in combined inventory, purchase commitments, and prepaids to secure and support this revenue growth. Blackwell and Rubin revenue between 2025 and 2027 is expected to reach US$1 trillion. Gross margin is forecast to maintain. The full-year operating expenses are expected to grow between 45% and 49% year-on-year, driven by continued investments in R&D and the accelerated adoption of AI tools.
Data Centre
Data Centre revenue is further categorised into two subsegments, namely hyperscale and AI cloud, industrial, and enterprise (ACIE). Hyperscale includes public cloud and large consumer internet companies, while ACIE targets AI clouds, enterprise, industrial, and sovereign AI deployments through AI data centres and factories.
Hyperscale revenue grew 12.0% sequentially to US$37.9 billion, representing about half of data centre revenue. Revenue from ACIE grew 31.1% sequentially to US$37.4 billion during the quarter as revenue from AI cloud alone more than tripled year-on-year. This suggests growth is broadening beyond traditional hyperscale customers.
The NVIDIA GB300 NVL72 is experiencing one of NVIDIA’s fastest product ramps, supported by demand from hyperscalers and frontier AI companies. Management expects a continued structural migration from CPUs to GPUs.
The number of partner data centres exceeding 10 megawatts nearly doubled year-on-year to over 80 sites. Sovereign revenue grew over 80% year-on-year, and NVIDIA infrastructure is now deployed across nearly 40 countries.
The Vera CPU and Rubin GPU architectures provide NVIDIA with a total addressable market of around US$200 billion beyond legacy graphics chips into broader AI infrastructure. The Vera Rubin platform is expected to begin shipping in 2H 2026. Vera Rubin is expected to outperform its predecessor Grace Blackwell.
Edge Computing
This segment highlights data processing devices for agentic and physical AI, including PCs, gaming consoles, workstations, AI-RAN (Radio Access Network) base stations, robotics, and automotive hardware. The segmental revenue increased 9.6% sequentially and 28.7% year-on-year to US$6.4 billion, driven by Robust Blackwell workstation demand. The growth was partially offset by lower consumer demand as memory and system prices increased. Further, the annualised revenue run rate of physical AI exceeded US$9 billion.
Key analyst questions
Hyperscale cloud remains an important near-term growth driver due to continued AI infrastructure investment. The hyperscalers represent the most direct go-to-market path given their highly concentrated numbers. However, NVIDIA’s overall growth is expected to outpace pure hyperscale cloud capital expenditure.
Management believes AI clouds, enterprise, industrial, and sovereign AI deployments will eventually become the larger opportunity given the larger addressable market. In the future, management expects physical AI and robotics to grow significantly within the next five years.
NVIDIA believes its integrated full-stack, end-to-end, co-designed approach positions it to serve a wider range of AI workloads and customer requirements. Management expects AI factory deployments to accelerate due to faster implementation and broader model compatibility.
Management highlighted inference as a key growth area, supported by demand from frontier AI companies and the recent partnership with Anthropic.
Hyperscalers’ capital expenditures are expected to reach US$1 trillion in 2026 and are projected to range between US$3 trillion and US$4 trillion by 2030. SaaS and AI companies are increasingly using high-performance computing to directly drive revenue and profits. As a result, the company is expected to remain capacity-constrained in the near term.
AI agents are expected to become a major long-term driver of compute demand. Agentic AI workloads require both CPUs for orchestration and tool execution, and GPUs for inference and reasoning. Revenue from CPU is expected to near US$20 billion in 2026.
The fifth perspective
NVIDIA continues to enhance its performance, driven by its versatile platform and software stack. As customers continue to be profitable and generate revenue from compute and GPUs even after they are fully depreciated, there will be sustained demand for its products and services. The company is also tapping into its new Vera CPU architecture for further growth.
