Table of Contents
- NVIDIA Q3 Financial Results at a Glance
- NVIDIA Revenue Growth Reflects Expanding AI Infrastructure Demand
- Data Center Revenue Reaches $107.6 Billion
- Vera Rubin NVL72 Enters Production and Shipping
- CUDA Remains Central to NVIDIA’s AI Computing Ecosystem
- NVIDIA DGX Spark Expands Local Artificial Intelligence Computing
- NVIDIA’s Artificial Intelligence Ecosystem Continues to Expand
- Edge Computing, Robotics and Autonomous Driving
- Financial Performance, Profitability and Cash Flow
- NVIDIA Share Repurchases and Dividend Payments
- NVIDIA Q4 Financial Outlook
- What NVIDIA’s Q3 Results Mean for the AI Industry
- Conclusion: NVIDIA’s Q3 Earnings Highlight the Scale of AI Computing Demand
- Editorial Note and Financial Disclaimer
NVIDIA Q3 earnings demonstrate the continuing expansion of the company’s artificial intelligence and accelerated computing business, with revenue reaching $115.8 billion for the third quarter ended October 25, 2026. The result represents a 20% increase from the previous quarter and 103% growth compared with the same period a year earlier. Data Center revenue contributed $107.6 billion, reinforcing the importance of AI infrastructure, advanced computing systems and large-scale deployment by cloud providers and enterprise customers.
The latest financial results provide a detailed view of NVIDIA’s position in a technology market increasingly shaped by Artificial Intelligence (AI), accelerated computing, data processing and increasingly sophisticated software workloads. The company’s performance reflects demand for computing infrastructure capable of supporting AI model training, inference, scientific research, enterprise applications and other data-intensive operations.
Beyond the headline revenue figure, the quarter included several significant developments across NVIDIA’s product and technology portfolio. Production and shipments of the Vera Rubin NVL72 platform began, the company expanded its local AI computing portfolio with a 64GB NVIDIA DGX Spark model, and its CUDA and CUDA-Q ecosystems continued to develop. NVIDIA also reported activity across robotics, autonomous driving, gaming, scientific computing, networking and enterprise AI.
The financial results are important because they show how closely NVIDIA’s business is connected to the wider expansion of AI infrastructure. Organizations developing and deploying AI systems need more than individual processors. They require integrated platforms combining computing hardware, networking, software, systems architecture and developer tools. NVIDIA’s latest announcements reflect this broader approach.
This article examines NVIDIA Q3 FY2027 earnings in detail, including revenue growth, Data Center performance, profitability, cash flow, shareholder returns, Vera Rubin, CUDA, DGX Spark and the company’s fourth-quarter financial outlook.
NVIDIA Q3 Financial Results at a Glance
NVIDIA reported revenue of $115.8 billion for Q3. This compares with $96.2 billion in Q2 FY2027 and $57.0 billion in Q3 FY2026. Revenue increased by 20% sequentially and 103% year over year.
The financial performance extended beyond revenue growth. NVIDIA reported substantial increases in operating income, net income and diluted earnings per share. Both GAAP and non-GAAP financial measures showed strong year-over-year expansion.
The following table summarizes the principal financial results.
| Financial measure | Q3 FY2027 GAAP | Q3 FY2027 non-GAAP |
|---|---|---|
| Revenue | $115.8 billion | $115.8 billion |
| Gross margin | 74.8% | 74.9% |
| Operating expenses | $9.2 billion | $8.8 billion |
| Operating income | $77.5 billion | $77.9 billion |
| Net income | $65.3 billion | $64.9 billion |
| Diluted earnings per share | $2.70 | $2.68 |
Amounts are rounded as presented in the company’s financial summary. GAAP and non-GAAP figures are separate reporting measures and should not be treated as interchangeable.
Revenue Growth Compared With Previous Periods
The year-over-year comparison illustrates the scale of NVIDIA’s recent expansion.
| Reporting period | Revenue |
|---|---|
| Q3 FY2026 | $57.0 billion |
| Q2 FY2027 | $96.2 billion |
| Q3 FY2027 | $115.8 billion |
Revenue increased by $58.8 billion compared with Q3 FY2026. It also increased by $19.6 billion compared with Q2 FY2027.
These comparisons provide two different perspectives. The year-over-year result demonstrates the substantial expansion of NVIDIA’s business over twelve months, while the sequential increase shows that revenue continued to grow even after the company had already reached a significantly higher revenue level in the previous quarter.
The results also highlight the scale of demand being served by NVIDIA’s computing platforms. As AI workloads become more complex and organizations deploy additional infrastructure, demand extends across processors, complete computing systems, networking, software and supporting technologies.
NVIDIA Revenue Growth Reflects Expanding AI Infrastructure Demand
The most important feature of NVIDIA’s Q3 FY2027 financial performance is the contribution made by its Data Center business. Revenue from this segment reached $107.6 billion, representing a 21% increase from the previous quarter and a 110% increase compared with the same quarter a year earlier.
This result demonstrates the central role that AI infrastructure plays in NVIDIA’s current business. Large technology companies, cloud providers, AI developers and other organizations require substantial computing resources to train and operate increasingly capable AI systems.
The infrastructure supporting these workloads is becoming more sophisticated. Modern AI deployments can involve large groups of processors working together, high-speed connections between computing systems, specialized networking, data storage, software frameworks and management tools.
NVIDIA’s business strategy addresses several of these requirements through an integrated computing ecosystem.
Why AI Infrastructure Is Important to NVIDIA’s Growth
Artificial Intelligence depends on computing resources that can process large volumes of information and perform complex calculations efficiently. Training an AI model can require extensive processing across large datasets, while operating a deployed model can involve continuous inference requests from applications and users.
These activities create demand for different types of computing capacity.
AI training infrastructure is designed to support the development and refinement of models. Inference infrastructure is used when trained models generate responses, analyze information, recognize patterns or perform other tasks. Both activities can require substantial computing resources, although their workloads and deployment requirements differ.
The growing use of AI agents adds another dimension. AI agents can perform multi-step tasks, interact with software tools and coordinate operations. These applications can require repeated model inference, data access and execution of additional processes.
NVIDIA’s announcements during the quarter addressed several parts of this expanding environment, including Vera Rubin infrastructure, Vera CPUs, networking, agent-oriented platforms and local AI systems.
Cloud Providers and AI Developers Expand Their Infrastructure
The Q3 business highlights described activity involving major cloud providers and AI-focused organizations.
NVIDIA reported that Vera Rubin NVL72 systems were operating at partners including CoreWeave, Google Cloud, Microsoft Azure, Oracle Cloud Infrastructure and Nebius. The company also described further expansion of its Vera Rubin cloud ecosystem, including the general availability of systems and Vera CPUs through Oracle Cloud Infrastructure and Vera Rubin-powered A5X instances through Google Cloud.
Other organizations identified in the company’s announcements included Meta, OpenAI, Nscale and Verda, among others. These developments illustrate how NVIDIA’s products are being incorporated into a wider network of cloud and AI infrastructure providers.
For customers, cloud access can provide a way to use advanced computing resources without independently building every component of a large-scale AI facility. For cloud providers, offering specialized AI infrastructure can support services aimed at developers, enterprises and organizations with significant computing requirements.
This ecosystem is important to NVIDIA because demand for AI computing is not limited to a single customer type. It can originate from large cloud platforms, AI model developers, enterprise technology departments, research institutions and specialized infrastructure operators.
Data Center Revenue Reaches $107.6 Billion
Data Center revenue of $107.6 billion accounted for the majority of NVIDIA’s reported quarterly revenue. The segment’s 110% year-over-year increase highlights the importance of accelerated computing and AI-related infrastructure to the company’s financial performance.
The Data Center business includes more than individual processors. The infrastructure described in NVIDIA’s quarterly announcements includes complete rack-scale systems, CPUs, GPUs, networking technologies, data processing units and software.

This broader product approach is relevant because large AI systems require components to work together. Computing performance is influenced by processing capacity, memory, communication between systems, power requirements and the software used to manage workloads.
Rack-Scale Computing and Integrated Systems
NVIDIA’s Vera Rubin NVL72 platform represents the company’s continued development of integrated computing infrastructure.
Rack-scale computing brings multiple components together into a coordinated system. Rather than treating each processor or server as an isolated product, a rack-scale platform is designed around the performance of the complete computing environment.
This approach is particularly relevant to large AI workloads. When many processors need to exchange information, the speed and efficiency of communication between components can influence the performance of the overall system.
NVIDIA’s Vera Rubin announcements also included advances to its open architecture and participation from technology companies developing infrastructure around the platform. The company described ecosystem activity involving Cisco, Google, Meta, Microsoft and Visionbay.
The expansion of this ecosystem matters because large infrastructure projects involve multiple suppliers and technology partners. Computing systems need networking, manufacturing capacity, deployment expertise and cloud integration. Collaboration across these areas can support the development and rollout of more complex AI facilities.
AI Factories and Infrastructure Deployment
Another development during the quarter was the introduction of DSX POD, a reference design for complete Vera Rubin AI factories. The concept involves modular containers in which partners can integrate and validate infrastructure before delivery.
The broader idea of an AI factory reflects a shift in how organizations think about computing facilities. Instead of viewing a data center only as a location containing servers, an AI factory is designed around the production of AI computing capacity.
This includes the hardware and networking required to process information, the software used to operate the systems and the supporting infrastructure needed to keep them running.
NVIDIA also described ecosystem adoption of its DSX approach in Europe, with an emphasis on reducing AI factory power requirements and enabling near-zero facility cooling water consumption.
These developments highlight an important consideration for the industry: AI infrastructure must be evaluated not only by computing performance but also by the resources required to deploy and operate it.
Electricity availability, cooling, physical space, networking and supply-chain capacity can all influence the pace of infrastructure expansion.
Vera Rubin NVL72 Enters Production and Shipping
One of the most significant product developments reported during Q3 FY2027 was the production ramp and start of shipments for NVIDIA Vera Rubin NVL72.
NVIDIA described this as the fastest product ramp in the company’s history. The announcement is important because the introduction of a new computing platform involves more than making a product available. It also requires manufacturing coordination, system integration, software readiness and deployment through infrastructure partners.
The Vera Rubin platform is positioned within NVIDIA’s broader strategy of supplying integrated systems for increasingly demanding AI and high-performance computing workloads.
What Makes Vera Rubin Important?
The Vera Rubin platform brings together technologies intended to support advanced computing requirements. NVIDIA’s announcements included the Vera Rubin NVL72 system, Vera CPUs and additional platform developments for scientific computing and AI infrastructure.
The company also introduced NVIDIA Rubin HPC, a GPU designed for simulation and data-processing workloads in science, engineering and analytics.
These developments show that NVIDIA’s infrastructure strategy extends beyond conventional AI model training. Scientific research, industrial simulation and other advanced computing applications can also require large amounts of processing capacity.
The Vera CPU platform was presented as a rack-scale offering for high-performance computing. NVIDIA described it as combining its first fully custom data center CPU with networking designed for demanding scientific modeling, simulation and agentic science workloads.
The company also reported that AI-native organizations and cloud providers were building with Vera CPUs, joining other infrastructure partners.
The Importance of the Vera Rubin Ecosystem
NVIDIA’s Vera Rubin strategy includes partnerships with companies involved in cloud infrastructure, networking and system deployment.
The quarterly announcements described systems operating at several cloud and infrastructure partners, alongside further platform availability and ecosystem expansion.
This is significant because large computing platforms are rarely deployed in isolation. Customers need access to the systems, integration with their existing environments and the ability to support the software and applications they intend to run.
A broad ecosystem can help address these requirements by bringing together infrastructure providers, technology companies and developers.
NVIDIA’s reported progress therefore reflects both product development and the expansion of the network of organizations involved in deploying the platform.
CUDA Remains Central to NVIDIA’s AI Computing Ecosystem
CUDA is an important part of NVIDIA’s technology strategy because it provides a software foundation for developers working with accelerated computing.
Hardware performance is only one part of a computing platform’s value. Developers also need software tools, libraries and programming environments that allow them to use the hardware effectively.
CUDA supports this relationship between NVIDIA’s processors and the applications built to run on them. It is part of the wider ecosystem that helps developers create software for accelerated computing, scientific applications, AI workloads and other demanding tasks.
Why CUDA Matters to Artificial Intelligence
AI development involves a combination of hardware and software. A processor may provide substantial computing capacity, but developers still need tools that allow their applications to access that capacity.
CUDA is relevant because it supports software development for NVIDIA’s accelerated computing architecture. Developers can use the ecosystem to build and optimize workloads that take advantage of compatible NVIDIA hardware.
This matters across different stages of AI development.
During model training, computing systems perform repeated calculations to adjust model parameters. During inference, the system processes inputs and produces outputs using an already trained model. Both stages can benefit from accelerated computing, although their performance requirements may differ.
The CUDA ecosystem also has relevance beyond AI. Scientific simulations, data analytics and engineering workloads can use accelerated computing to process calculations that would otherwise require substantial computing time.
CUDA and the Wider Developer Ecosystem
The value of a software platform is influenced by the range of applications and developers it supports.
A well-established developer ecosystem can make it easier for organizations to build on existing tools, develop specialized applications and integrate new computing capabilities into established workflows.
For NVIDIA, CUDA is part of a broader platform strategy that connects hardware, software and developer adoption.
The company’s quarterly developments also included activity involving CUDA-Q, its open-source platform for quantum computing development.
CUDA-Q and Quantum Computing Development
NVIDIA reported an expansion of its CUDA-Q platform with CUDA-Q Logical, which is intended to support the development of fault-tolerant quantum computing.
The company also reported a simulation of a fault-tolerant quantum processor on the JUPITER system at the Jülich Supercomputing Centre.
Quantum computing remains a specialized and developing area of technology. The relevance of CUDA-Q is that it provides a software environment for work involving quantum processors and classical computing resources.
The announcement adds another dimension to NVIDIA’s computing ecosystem. While Data Center AI infrastructure represents the largest source of current revenue in the reported quarter, the company’s research and development activities also extend into other advanced computing fields.
NVIDIA DGX Spark Expands Local Artificial Intelligence Computing
NVIDIA also expanded its local AI computing portfolio during Q3 FY2027 with a new 64GB model of NVIDIA DGX Spark.
The announcement is relevant because AI computing is not limited to large cloud facilities. Developers, researchers and technical teams may also want to run AI models locally, experiment with applications and build workflows without relying entirely on remote infrastructure.
DGX Spark is part of NVIDIA’s effort to make advanced AI computing capabilities available in a desktop form factor.
What Is NVIDIA DGX Spark?
NVIDIA DGX Spark is a compact AI computing system designed to bring substantial AI processing capabilities to the desktop.
The new 64GB model expands the product lineup and was described by NVIDIA as bringing near-frontier AI capabilities to desktops at a more accessible price point.
Local AI computing can be useful for developers who need to test models, prototype applications, experiment with AI agents or work with supported AI software in a more direct environment.
The ability to run workloads locally can also offer practical advantages in certain development settings. Teams may prefer to test applications close to where they are building them, manage their own development environment or reduce dependence on continuous remote access for particular tasks.
However, local systems and large cloud platforms serve different purposes. A desktop AI system does not automatically replace the scale, capacity or shared infrastructure of a major data center. Its role is to make a different class of development and deployment activity more accessible.
Local AI and the Growth of AI Agents
NVIDIA’s DGX Spark announcement was accompanied by developments involving AI models and local agents.
The company identified support for models from NVIDIA Nemotron, Black Forest Labs, Liquid AI and Qwen, as well as expanded support involving Hermes, OpenClaw and Perplexity.
These developments reflect interest in AI systems that can operate within local computing environments.
AI agents can be designed to carry out multi-step activities, use tools and interact with applications. Running supported models and agents locally may provide developers with additional options for testing and building these systems.
For businesses, the choice between local and cloud-based AI depends on several factors, including the size of the model, processing requirements, security considerations, cost, data handling and the intended application.
DGX Spark adds another option within NVIDIA’s wider computing portfolio.
Why DGX Spark Matters to NVIDIA’s Product Strategy
DGX Spark is strategically relevant because it broadens the range of environments in which developers can work with NVIDIA AI technologies.
The company’s Data Center systems address large-scale infrastructure requirements. DGX Spark addresses a different need: desktop access to local AI development and computing.
This creates a more varied product ecosystem spanning large cloud deployments, enterprise infrastructure and local development systems.
It also reinforces the importance of software compatibility. Hardware is most useful when developers have access to models, tools and frameworks that can run on it.
NVIDIA’s local AI announcements therefore connect its desktop products with the wider AI software ecosystem.
NVIDIA’s Artificial Intelligence Ecosystem Continues to Expand
The Q3 FY2027 announcements covered several areas beyond Data Center systems and DGX Spark. These developments demonstrate the range of applications NVIDIA is targeting through its hardware, software and platform ecosystem.
The company’s activities included enterprise AI, scientific research, healthcare, supply-chain applications, autonomous driving, robotics, telecommunications and gaming.
Enterprise AI and AI Agent Development
NVIDIA reported the launch of its Open Agent Safety Platform and the introduction of the OpenShell Reference Architecture.
The architecture combines OpenShell software, Vera Rubin for model inference, Vera CPUs for agent execution and BlueField-4 DPUs running NVIDIA Sentry to monitor agent behaviour and enforce security policies in silicon.
This development is relevant because AI agents can interact with software tools and carry out tasks that extend beyond generating text or answering questions.
As organizations explore agent-based applications, questions around execution, monitoring and security become increasingly important. An agent that can interact with business systems needs controls that help organizations manage its activity.
NVIDIA’s announcements indicate that the company is addressing this area through a combination of software and infrastructure components.
The company also reported an agreement to acquire Hugging Face, with the stated aim of scaling its platform and making intelligence more open, capable and accessible. This is a development in the wider AI ecosystem and should be considered separately from the quarter’s reported revenue figures.
Scientific Research and Healthcare
NVIDIA introduced Nemotron for Science to help researchers build specialized agents for scientific discovery.
The company also described collaborations with healthcare and life sciences organizations involving protein design, drug discovery and precision patient care.
Scientific research can involve large datasets, complex calculations and simulations. AI tools may assist researchers with tasks such as analyzing information, identifying patterns and exploring potential solutions.
In healthcare and life sciences, these capabilities can support research workflows, but their practical use depends on the requirements of each application and the appropriate scientific and clinical validation.
NVIDIA’s developments show how accelerated computing and AI software are being positioned for specialized research environments.
AI in Supply Chains
The company also reported a partnership with Palantir focused on sovereign AI intelligence and NVIDIA Nemotron open models for critical supply chains.
Supply-chain operations involve multiple stages, organizations and information systems. AI tools can be used to analyze data, support planning and help identify operational patterns.
The reported collaboration reflects an effort to apply AI capabilities to business environments where information management and operational coordination are important.
NVIDIA also described work with Firebird to scale potential AI infrastructure capacity.
Manufacturing and Infrastructure Partnerships
The expansion of AI infrastructure requires more than demand for computing systems. It also depends on manufacturing capacity and the availability of specialized components.
NVIDIA announced plans with global supply-chain partners, including Amkor Technology, IBIDEN, King Yuan Electronics, Samsung Electro-Mechanics, Siliconware Precision Industries and Unimicron Technology Corporation.
The stated objective was to scale critical manufacturing capabilities to support demand for NVIDIA infrastructure.
These partnerships are relevant because advanced computing systems rely on complex manufacturing processes and coordinated supply chains. Increasing production capacity can be an important part of supporting a broader infrastructure rollout.
Networking and AI Factory Connectivity
NVIDIA also announced plans with Cisco for a new optical-module form factor called USFP, designed to increase bandwidth density and power efficiency for next-generation AI factory networks.
Networking is an important part of large-scale AI infrastructure. When computing systems operate together, data must move between processors, servers and other components.
As AI facilities become larger and more complex, network design can influence how efficiently computing resources are used.
NVIDIA’s work in this area reflects the company’s approach of developing a broader infrastructure platform rather than focusing only on processing hardware.
The company also reported a collaboration with Australia’s data center ecosystem supporting a planned buildout of up to two gigawatts by 2027 for NVIDIA DSX AI factories.
Edge Computing, Robotics and Autonomous Driving
NVIDIA’s quarterly announcements also covered its Edge Computing activities. The company reported third-quarter Edge Computing revenue of $8.2 billion, up 14% from the previous quarter and 42% compared with a year earlier.
Edge computing brings processing capabilities closer to where data is generated or where applications are used. Depending on the application, this can involve local devices, vehicles, industrial equipment or other connected systems.
The company’s announcements included developments in personal computing, autonomous driving, robotics and telecommunications.
AI-Powered Personal Computing
NVIDIA reported that RTX Spark laptops were shipping from ASUS, Dell, HP, Lenovo, Microsoft Surface and MSI.
The company also introduced DLSS 5 in NBA 2K27 and Hogwarts Legacy, bringing AI-powered lighting and materials capabilities to GeForce RTX 50 Series GPUs, GeForce NOW and RTX Spark.
These developments show how AI technologies are being incorporated into consumer computing and gaming experiences, as well as professional and enterprise applications.
The broader significance is that AI computing is increasingly present across different types of devices and workloads.
Autonomous Driving Technology
NVIDIA introduced Alpamayo development tools and the Halos SAFE framework to help developers build and evaluate autonomous-driving systems.
The company also reported expanded partnerships with Mercedes-Benz and Volvo Cars to advance safe automated driving.
Additional activity involved European mobility organizations demonstrating NVIDIA Hyperion use cases for automated vehicles and future robotaxi fleets.
Autonomous driving requires systems to process information from their surroundings, interpret road conditions and support decisions in complex environments.
AI and accelerated computing can contribute to the development and evaluation of these systems. However, the practical deployment of automated driving technology also requires extensive testing, safety engineering and compliance with applicable requirements.
Robotics and Industrial Applications
NVIDIA advanced its Isaac robotics platform with new development tools, datasets and runtime capabilities.
The company also described an Agile Robots reference humanoid powered by NVIDIA Jetson Thor and built on Isaac GR00T.
NVIDIA Cosmos was expanded with tools for training, evaluating and deploying world models for robotics, autonomous vehicles and industrial applications.
These developments reflect growing interest in AI systems that interact with physical environments.
Robotics applications can require a combination of perception, planning, simulation and real-time processing. AI models can help robots interpret information and perform increasingly complex tasks, while development platforms provide tools for building and evaluating those systems.
AI-RAN and Telecommunications
NVIDIA also reported progress in AI-RAN through work with Nokia and global telecommunications operators. The announcement included T-Mobile’s first AI-RAN call on a commercial network.
AI-RAN refers to the integration of AI and accelerated computing into radio access network environments.
Telecommunications networks generate and process large volumes of data. The use of AI in network operations can create opportunities for more advanced analysis and distributed computing.
NVIDIA’s reported activity suggests that its accelerated computing technologies are being explored across additional infrastructure markets beyond traditional data centers.
Financial Performance, Profitability and Cash Flow
Revenue growth is an important measure of NVIDIA’s performance, but profitability and cash generation provide additional insight into the company’s financial position.
For Q3 FY2027, NVIDIA reported GAAP gross margin of 74.8% and non-GAAP gross margin of 74.9%.
GAAP operating income was $77.5 billion, compared with non-GAAP operating income of $77.9 billion. GAAP net income was $65.3 billion, while non-GAAP net income was $64.9 billion.
Diluted earnings per share were $2.70 on a GAAP basis and $2.68 on a non-GAAP basis.
Understanding GAAP and Non-GAAP Results
GAAP refers to generally accepted accounting principles. It provides a standardized framework for financial reporting.
Non-GAAP measures are adjusted financial measures that exclude or modify certain items under the company’s stated methodology. They can provide an additional perspective on operating performance, but they are not substitutes for GAAP reporting.
Investors reviewing NVIDIA’s results should compare the two sets of measures carefully and understand the adjustments used in the non-GAAP presentation.
The differences in Q3 FY2027 figures are relatively small in some categories, but the distinction remains important. For example, gross margin was 74.8% under GAAP and 74.9% under non-GAAP reporting.
Operating Expenses and Operating Income
NVIDIA reported GAAP operating expenses of $9.2 billion and non-GAAP operating expenses of $8.8 billion.
GAAP operating income reached $77.5 billion, compared with $36.0 billion in Q3 FY2026. Non-GAAP operating income was $77.9 billion, compared with $36.1 billion in the year-earlier quarter.
The year-over-year increase in operating income reflects the expansion of revenue and the company’s reported profitability during the quarter.
Operating expenses also increased compared with the previous year. This is an important part of evaluating the results because a growing technology business may need to invest in research and development, sales, infrastructure and other activities to support its operations.
The relationship between revenue, expenses and operating income provides a more complete view than revenue growth alone.
Net Income and Earnings Per Share
GAAP net income increased to $65.3 billion from $31.9 billion in Q3 FY2026.
Non-GAAP net income reached $64.9 billion, compared with $31.0 billion in the corresponding period a year earlier.
Diluted GAAP earnings per share increased to $2.70 from $1.30. Non-GAAP diluted earnings per share increased to $2.68 from $1.27.
Earnings per share provide a measure of reported profit relative to the weighted average number of shares used in the calculation.
For investors, EPS is one component of financial analysis. It is generally considered alongside revenue growth, profitability, cash flow, valuation and the company’s forward outlook.
Operating Cash Flow and Free Cash Flow
NVIDIA generated $48.253 billion in GAAP net cash from operating activities during Q3 FY2027.
Free cash flow was $43.777 billion.
Cash flow is particularly relevant for a technology company operating in a market where infrastructure development, manufacturing capacity and product innovation can require significant financial resources.
Operating cash flow measures the cash generated by business operations during a reporting period. Free cash flow, as presented in the company’s financial statements, reflects operating cash flow after specified capital expenditure and related adjustments.
These measures provide another perspective on NVIDIA’s financial performance. Strong reported earnings and substantial cash generation can support investment in the business, shareholder returns and other financial priorities.
It is also important to distinguish cash flow from accounting profit. The two measures are related, but they are calculated differently and can be affected by different factors.
NVIDIA Share Repurchases and Dividend Payments
NVIDIA continued returning capital to shareholders during Q3 FY2027.
The company reported that it returned $28.1 billion through share repurchases and cash dividends during the quarter.
In September 2026, NVIDIA’s Board of Directors approved an additional $150 billion increase to the company’s share repurchase authorization, without an expiration date.
At the end of Q3 FY2027, approximately $227.3 billion remained available under the share repurchase authorization.
Share Repurchases
Share repurchases involve a company buying back its own shares. Depending on the number of shares repurchased and other factors, buybacks can affect the number of shares outstanding and future earnings-per-share calculations.
NVIDIA’s reported quarterly financial statements included $22.122 billion in payments related to repurchases of common stock.
The company also reported dividends paid of $6.026 billion in its cash flow statement.
The combined shareholder-return figure reported by NVIDIA was $28.1 billion.
Quarterly Dividend
NVIDIA announced a quarterly cash dividend of $0.25 per share.
The payment was scheduled for December 23, 2026, to shareholders of record on December 2, 2026.
Dividend payments are one way companies distribute capital to shareholders. Investors evaluating a dividend should consider the amount, payment schedule and the company’s broader financial position.
Share repurchases and dividends are separate mechanisms, and neither should be interpreted in isolation as a complete measure of investment value.
NVIDIA Q4 Financial Outlook
NVIDIA’s outlook for Q4 provides an indication of management’s expectations for the next reporting period.
The company projected revenue of $130.0 billion, plus or minus 3%. The fourth quarter includes an additional week.
This guidance is a forward-looking estimate rather than a reported financial result. Actual performance may differ from the outlook.
Revenue Guidance
NVIDIA’s Q4 revenue outlook is $130.0 billion, plus or minus 3%.
The company also stated that it was not assuming any Data Center compute revenue from China in its outlook.
This qualification is important when interpreting the guidance. It describes an assumption used in preparing the company’s forecast and should not be interpreted as a statement that no other China-related activity exists.
The revenue estimate indicates that NVIDIA expects another substantial quarter, although the final result will depend on business conditions and the factors affecting customer demand and infrastructure deployment.
Gross Margin Expectations
NVIDIA expects GAAP and non-GAAP gross margins to be between 71.0% and 72.0%.
This range is below the Q3 FY2027 reported gross margins of 74.8% GAAP and 74.9% non-GAAP.
Gross margin can be influenced by product mix, manufacturing costs, system configurations and other business factors. The company’s guidance provides a range rather than a guaranteed outcome.
Operating Expense Expectations
The company expects Q4 FY2027 operating expenses of approximately:
- GAAP operating expenses: $11.0 billion.
- Non-GAAP operating expenses: $10.7 billion.
These estimates provide an indication of the expense base NVIDIA expects to support during the quarter.
Operating expenses are relevant because they reflect the resources used to run and develop the business. As NVIDIA expands its product portfolio and supports a larger infrastructure ecosystem, expense levels are an important part of assessing operating performance.
Fiscal 2027 Tax Rate Guidance
For the full fiscal year 2027, NVIDIA expects GAAP and non-GAAP tax rates to be between 16.0% and 18.0%.
The guidance excludes discrete items and material changes to NVIDIA’s tax environment.
Investors should keep this qualification in mind when comparing the projected tax rate with reported results in individual quarters.
What NVIDIA’s Q3 Results Mean for the AI Industry
NVIDIA’s third-quarter results provide a snapshot of an AI industry in which demand for computing infrastructure continues to expand.
The company’s $115.8 billion in revenue and $107.6 billion in Data Center revenue show the scale of its current business. However, the broader significance of the quarter can be seen in the range of technologies and applications covered by its announcements.
The company is developing products for large AI facilities, local AI computing, scientific research, autonomous vehicles, robotics, gaming and telecommunications.
These markets have different requirements, but they share a need for computing systems capable of processing increasingly complex workloads.
AI Infrastructure Is Becoming More Integrated
The Vera Rubin announcements illustrate the movement toward integrated computing platforms.
Large AI workloads depend on more than the performance of a single processor. They require systems that combine computing, networking and software.
This creates opportunities for suppliers that can provide several connected parts of the infrastructure stack.
NVIDIA’s strategy involves CPUs, GPUs, networking, software and complete systems. The company’s reported partnerships also show how important external infrastructure providers and manufacturing partners are to the deployment process.
Software and Developer Adoption Remain Important
CUDA and the wider NVIDIA software ecosystem are relevant because hardware needs compatible software and developer tools to support practical applications.
A platform’s long-term usefulness is influenced by the tools available to developers, the applications that can run on it and the ease with which customers can integrate it into their environments.
The continued development of CUDA, CUDA-Q and local AI tools reflects the importance of this software dimension.
For organizations selecting AI infrastructure, the decision can involve more than comparing processor specifications. Software compatibility, developer experience, deployment requirements and the ability to support existing workflows can all influence the choice.
AI Computing Is Expanding Beyond Large Data Centers
DGX Spark and the company’s Edge Computing developments show that AI computing is not confined to large cloud facilities.
Local systems can support development and experimentation, while edge platforms can bring computing closer to vehicles, industrial systems and other applications.
This does not mean that desktop or edge systems replace large AI data centers. Instead, the different types of infrastructure serve different needs.
A company might use large cloud systems for extensive model training, local hardware for development and testing, and edge computing for applications that need processing near the point where data is generated.
The expansion of these environments could create demand for a wider range of computing products and software tools.
Infrastructure Constraints Will Remain Important
The growth of AI infrastructure also creates practical challenges.
Large computing facilities require electricity, cooling, physical space, networking and access to specialized components. These requirements can influence how quickly new systems are deployed and how efficiently they operate.
NVIDIA’s DSX and DSX POD announcements address some of these infrastructure considerations through reference designs and an emphasis on power and cooling efficiency.
However, the wider deployment of AI facilities depends on multiple factors, including energy infrastructure, supply-chain capacity, construction timelines and customer investment decisions.
These factors are relevant when assessing the sustainability of current growth rates. Strong demand is an important part of the picture, but the ability to deliver and operate infrastructure at scale also matters.
Conclusion: NVIDIA’s Q3 Earnings Highlight the Scale of AI Computing Demand
NVIDIA Q3 earnings show a company continuing to expand its financial performance alongside a broad range of developments across AI infrastructure and accelerated computing.
Revenue reached $115.8 billion, increasing 20% from the previous quarter and 103% year over year. Data Center revenue contributed $107.6 billion, up 110% from the same quarter a year earlier.
The company also reported GAAP net income of $65.3 billion, non-GAAP net income of $64.9 billion, GAAP diluted earnings per share of $2.70 and non-GAAP diluted earnings per share of $2.68.
Cash generation was another significant part of the quarter, with $48.253 billion in operating cash flow and $43.777 billion in free cash flow.
On the product and technology side, NVIDIA began shipping Vera Rubin NVL72, expanded the Vera Rubin ecosystem, introduced a 64GB DGX Spark model and continued developing CUDA and CUDA-Q. Its announcements also covered enterprise AI, AI agents, scientific research, healthcare, robotics, autonomous driving, networking and edge computing.
The Q4 FY2027 revenue outlook of $130.0 billion, plus or minus 3%, indicates that the company expects another substantial quarter. The guidance also includes specific assumptions around gross margins, operating expenses and Data Center compute revenue from China.
Taken together, the results illustrate how NVIDIA’s business is connected to the expanding market for AI infrastructure. Its financial performance, product roadmap and partner ecosystem provide important context for understanding the company’s current position in accelerated computing.
The next stage of NVIDIA’s development will depend on several factors, including customer demand, product execution, infrastructure deployment, manufacturing capacity and the continued adoption of AI across different industries.
For investors, technology professionals and businesses evaluating the AI market, the Q3 FY2027 results offer a detailed view of the scale of NVIDIA’s operations and the breadth of its current technology strategy.
Editorial Note and Financial Disclaimer
This article is intended for general informational and educational purposes. It discusses NVIDIA’s reported Q3 FY2027 financial results, business developments and forward-looking guidance. It does not constitute financial, investment, legal, tax or trading advice, and it should not be interpreted as a recommendation to buy, sell or hold NVIDIA shares or any other security.
Financial figures and company guidance are presented for informational purposes. Forward-looking statements, estimates and projections are subject to risks, uncertainties and assumptions, and actual results may differ materially from expectations. Readers should review relevant company disclosures and conduct their own independent research before making financial or investment decisions.










