Want to dig deeper on NVIDIA Corp and thousands of companies?
Rapidly search an extensive universe of public and private content — earnings transcripts, company filings, equity research, trade journals, and more.
Request a Free TrialCompanies by industry
Aerospace & DefenseAutomobile ComponentsAutomobilesBanksBeveragesBiotechnologyBroadline RetailCapital MarketsChemicalsCommunications EquipmentConsumer FinanceConsumer Staples Distribution & RetailContainers & PackagingElectric UtilitiesElectronic Equipment, Instruments & ComponentsEnergy Equipment & ServicesEntertainmentFinancial ServicesFood ProductsGround TransportationHealth Care Equipment & SuppliesHealth Care Providers & ServicesHotels, Restaurants & LeisureHousehold DurablesHousehold ProductsIT ServicesIndustrial ConglomeratesIndustrial REITsInsuranceInteractive Media & ServicesLife Sciences Tools & ServicesMachineryMediaMetals & MiningMulti-UtilitiesOil, Gas & Consumable FuelsPersonal Care ProductsPharmaceuticalsSemiconductors & Semiconductor EquipmentSoftwareSpecialized REITsSpecialty RetailTechnology Hardware, Storage & PeripheralsTextiles, Apparel & Luxury GoodsTobaccoWireless Telecommunication Services
NVIDIA Corp Earnings - Analysis & Highlights for Q4 2025
Overview
PositivesNegativesOutlook
- Revenue was a record $570 million in Automotive, up 27% sequentially and up 103% YoY. Full-year revenue of $1.7 billion increased 55% YoY.
- Regional cloud hosting NVIDIA GPUs increased as a percentage of Data Center revenue, reflecting continued AI factory build-outs globally and rapidly rising demand for AI reasoning models and agents.
- Blackwell sales exceeded expectations in Q4.
- The company delivered $11 billion of Blackwell revenue to meet strong demand.
- Gaming revenue of $2.5 billion decreased by 22% sequentially and 11% YoY.
- GAAP operating expenses were up 9% and non-GAAP operating expenses were 11%, reflecting higher engineering development costs and higher compute and infrastructure costs for new product introductions.
- Demand remained strong throughout the holiday, but Q4 shipments were impacted by supply constraints.
- Data Center sales in China remained well below levels seen on the onset of export controls.
- GAAP and non-GAAP operating expenses are expected to be approximately $5.2 billion and $3.6 billion, respectively.
- Networking is expected to return to growth in Q1.
- Full year fiscal year 2026 operating expenses are expected to grow to be in the mid-30s.
- Sequential growth is expected in both data center and gaming. Within data center, sequential growth is expected from both compute and networking.
- The company expects to be in the mid-70s later this year.
Q&A Highlights from NVIDIA Corp Earnings Call Q4 2025
- Analyst asked about the potential future of inference dedicated clusters and the overall impact to NVIDIA and their customers.
- Jensen Huang explained that there are multiple scaling laws, including pre-training, post-training, and inference scaling. He believes that the future of inference dedicated clusters lies in Test-Time compute or reasoning, long thinking, and inference scaling. He also mentioned that NVIDIA's architecture is designed to be flexible and easy to use in all of these different ways, allowing for a unified architecture.
- Jensen Huang explained that there are multiple scaling laws, including pre-training, post-training, and inference scaling. He believes that the future of inference dedicated clusters lies in Test-Time compute or reasoning, long thinking, and inference scaling. He also mentioned that NVIDIA's architecture is designed to be flexible and easy to use in all of these different ways, allowing for a unified architecture.
- Analyst asked about the ramp of the NVL72 platforms and if there are still bottlenecks to consider at a systems level.
- Jensen Huang stated that the company has successfully ramped up Grace Blackwell, delivering some $11 billion of revenues last quarter. He mentioned that the company has a large installation of Grace Blackwells for their own engineering and design teams and software teams, and that many customers are bringing their systems online. He also acknowledged that nothing is easy about what they're doing, but they're doing great and all of their partners are doing great.
- Jensen Huang stated that the company has successfully ramped up Grace Blackwell, delivering some $11 billion of revenues last quarter. He mentioned that the company has a large installation of Grace Blackwells for their own engineering and design teams and software teams, and that many customers are bringing their systems online. He also acknowledged that nothing is easy about what they're doing, but they're doing great and all of their partners are doing great.
- Analyst asked about the company's confidence in sustaining strong demand into next year and how DeepSeek and other innovations have affected that view.
- The company's confidence in sustaining strong demand into next year is based on several factors, including a good line of sight of the amount of capital investment that data centers are building out, the vast majority of software being based on machine learning, and forecasts and plans from top partners. Additionally, the company is aware of the vibrant start-up scene and the need for computing infrastructure by these companies.
- The company's confidence in sustaining strong demand into next year is based on several factors, including a good line of sight of the amount of capital investment that data centers are building out, the vast majority of software being based on machine learning, and forecasts and plans from top partners. Additionally, the company is aware of the vibrant start-up scene and the need for computing infrastructure by these companies.
- Analyst asked about the demand dynamics for Ultra and how the company manages the simultaneous ramps of Blackwell solutions.
- The company is on track to execute Blackwell Ultra in the second half of 2023, with new networking, memories, and processors. The team is working closely with partners and customers to ensure a smooth transition, and the architecture of the system will remain the same. The company is also working on the click after that, called Vera Rubin, and preparing for the transition to that product.
- The company is on track to execute Blackwell Ultra in the second half of 2023, with new networking, memories, and processors. The team is working closely with partners and customers to ensure a smooth transition, and the architecture of the system will remain the same. The company is also working on the click after that, called Vera Rubin, and preparing for the transition to that product.
- Analyst asked about the balance between customer ASIC and merchant GPU and whether customers will build heterogeneous superclusters that use both.
- Jensen Huang, CEO of NVIDIA, explained that NVIDIA's architecture is general, end-to-end, and everywhere, making it more versatile and accessible than ASICs. He also emphasized the importance of performance and the software stack, which is incredibly hard to build, and the ability to deploy quickly. He stated that the company's technology is not only more advanced but also has better software capability and a faster deployment rate.
- Jensen Huang, CEO of NVIDIA, explained that NVIDIA's architecture is general, end-to-end, and everywhere, making it more versatile and accessible than ASICs. He also emphasized the importance of performance and the software stack, which is incredibly hard to build, and the ability to deploy quickly. He stated that the company's technology is not only more advanced but also has better software capability and a faster deployment rate.
- Analyst asked about the company's growth rate and how it can keep growing so fast with the mix shift towards the US.
- Jensen Huang responded that China's percentage is approximately the same as in previous quarters, but it is half of what it was before the export control. He also mentioned that AI has gone mainstream and is being integrated into every application, and that the company is in the beginning of a new transition where all software and services will be based on machine learning. He concluded by stating that no technology has ever had the opportunity to address a larger part of the world's GDP than AI, and that NVIDIA's technology is in the beginnings of this new era.
- Jensen Huang responded that China's percentage is approximately the same as in previous quarters, but it is half of what it was before the export control. He also mentioned that AI has gone mainstream and is being integrated into every application, and that the company is in the beginning of a new transition where all software and services will be based on machine learning. He concluded by stating that no technology has ever had the opportunity to address a larger part of the world's GDP than AI, and that NVIDIA's technology is in the beginnings of this new era.
- Analyst asked about the growth of enterprise within the data center, specifically the growth of enterprise compared to hyperscalers.
- Enterprise within the data center grew 2x year-on-year in the January quarter, and it is growing at a similar pace to the large CSPs. Both enterprises and CSPs are important areas for the company, and they both grow well.
- Enterprise within the data center grew 2x year-on-year in the January quarter, and it is growing at a similar pace to the large CSPs. Both enterprises and CSPs are important areas for the company, and they both grow well.
- Analyst asked about the infrastructure that has been deployed and the replacement cycle perspective, specifically if it's GB300 or Rubin cycle where there may be refresh opportunities.
- The company's GPUs are still being used for data processing, data curation, and machine learning-based search. The less intensive workloads are put onto the installed base of the past, and all of the company's GPUs are well employed.
- The company's GPUs are still being used for data processing, data curation, and machine learning-based search. The less intensive workloads are put onto the installed base of the past, and all of the company's GPUs are well employed.
- Analyst asked about the impact of tariffs on the broader semiconductor industry.
- Colette M. Kress, Executive Vice President and CFO, stated that the company's gross margins are complex and involve many different configurations, making it difficult to assess the impact of tariffs. She also mentioned that the company is awaiting further information from the US government on the plan and timing of the tariffs.
- Colette M. Kress, Executive Vice President and CFO, stated that the company's gross margins are complex and involve many different configurations, making it difficult to assess the impact of tariffs. She also mentioned that the company is awaiting further information from the US government on the plan and timing of the tariffs.
- Analyst asked about the demand for Blackwell, NVIDIA's latest AI platform.
- Jensen Huang, CEO of NVIDIA, stated that the demand for Blackwell is extraordinary, driven by the evolution of AI beyond perception and generative AI into reasoning. He also mentioned that reasoning models can consume 100 times more compute than traditional models, and that future reasoning models may consume even more compute. He stated that Blackwell is designed to address these scaling laws, and that the company is seeing strong growth in AI computing.
- Jensen Huang, CEO of NVIDIA, stated that the demand for Blackwell is extraordinary, driven by the evolution of AI beyond perception and generative AI into reasoning. He also mentioned that reasoning models can consume 100 times more compute than traditional models, and that future reasoning models may consume even more compute. He stated that Blackwell is designed to address these scaling laws, and that the company is seeing strong growth in AI computing.
