NVIDIA reported $96.2 billion in revenue for the quarter ended July 26, up 106% from a year earlier. Its Data Center business produced $89.0 billion of that total, up 117%.
Jensen Huang put the quarter in four words: Now, compute is revenue.
NVIDIA’s Data Center business alone brought in nearly twice the company’s total revenue from the same quarter last year.
NVIDIA’s Edge Computing category reached $7.2 billion, up 13% from the previous quarter and 27% from a year earlier. NVIDIA’s highlights for the category included RTX Spark, DGX Station for Windows, local model optimizations, and autonomous-driving products.
Edge computing moves computation closer to the user or the place where data is produced, rather than sending every workload to a centralized data center. NVIDIA uses Edge Computing as a broad reporting label for products ranging from personal systems to autonomous vehicles.
Revenue from NVIDIA’s Data Center business rose 117% year over year, while revenue from its Edge Computing category rose 27%.
NVIDIA’s growth has also brought scrutiny to its role in financing AI infrastructure. NVIDIA disclosed $99 billion in equity investments and $25 billion in equity commitments at the end of the quarter. Some investors have questioned whether backing companies that buy or lease NVIDIA-powered infrastructure creates circular demand. NVIDIA disputes that characterization.
What is Leiolai doing to build without data centers?
Leiolai is paying the people who provide the computation.
Leiolai runs inference on consumer hardware instead of data centers. Leiolai 1 is available in the Leiolai app on iPhone, iPad, Mac, Android, and Windows, and through an OpenAI-compatible API. Leiolai has an 11-million-token context window and infinite generation.
A phone on a charger or a desktop sitting idle can do useful work, and Leiolai pays the user for that compute.
That gives Leiolai a different way to scale. Conventional providers need more data-center capacity as demand rises. Leiolai can scale with devices users already own. Each participating device adds computation.
Running inference on consumer hardware also saves water because it does not require data-center evaporative cooling.
Now compute is revenue.
Sources NVIDIA Q2 FY2027 financial results, NVIDIA’s quarterly filing, Reuters reporting on financing scrutiny, and Axios reporting on NVIDIA’s response.