Jensen Huang believes Nvidia can grow revenue roughly 70% next year despite already operating at extraordinary global scale. His confidence comes from Nvidia’s position across nearly every major layer of the expanding artificial intelligence infrastructure market.
Speaking at Goldman Sachs’ Communacopia and Technology conference, Jensen Huang argued that Nvidia has unusual visibility into future demand. The company works directly with cloud providers, AI laboratories, manufacturers, startups, data centers, and semiconductor suppliers worldwide.
That broad network gives Nvidia insight into infrastructure projects long before customers actually begin deploying new computing systems. Jensen Huang says this visibility helps the company understand how much AI capacity companies are preparing to build globally.
Nvidia’s growth story has moved far beyond selling individual processors to companies developing artificial intelligence products and services. Its largest systems now combine GPUs, CPUs, networking equipment, memory, software, and supporting infrastructure into complete computing platforms.
Jensen Huang said people still misunderstand Nvidia when they think about the company primarily as a chip manufacturer. Modern Nvidia systems can contain millions of components while consuming enormous amounts of electricity inside specialized data center environments.
One example is Nvidia’s GB200 NVL72 system, which combines 72 Blackwell GPUs with 36 Grace processors. Huang said orders for that platform are currently increasing approximately 27% every month across Nvidia’s expanding customer base.
Those numbers help explain why Jensen Huang remains confident despite concerns that Nvidia’s extraordinary growth cannot continue indefinitely. The company is already generating revenue at levels rarely seen across the global semiconductor industry.
Analysts expect Nvidia to finish its current fiscal year with approximately $400 billion in total annual revenue. Another 70% increase would push that figure toward roughly $680 billion during the following fiscal year.
Achieving that growth becomes increasingly difficult as Nvidia’s existing business grows larger with every successful quarter. However, Jensen Huang believes artificial intelligence infrastructure spending is expanding quickly enough to overcome that mathematical challenge.
Much of his confidence comes from Nvidia supporting models produced by nearly every major artificial intelligence laboratory. OpenAI, Anthropic, Google, independent startups, and open-model developers can all run their workloads using Nvidia technology.
That gives Nvidia an important advantage because the company does not need one particular AI laboratory to dominate. Jensen Huang argues that Nvidia benefits whenever successful models require increasingly large amounts of computing infrastructure.
The company also receives demand information from major cloud providers and specialist companies building dedicated artificial intelligence computing capacity. Those relationships provide early visibility into planned deployments before many new data centers begin serving commercial workloads.
Jensen Huang said Nvidia tracks global power availability, land, construction, equipment, and expected computing requirements across major projects. That effectively gives management an unusually detailed picture of where future artificial intelligence infrastructure spending may emerge.
The company’s influence also stretches into newer cloud providers built specifically around artificial intelligence workloads and accelerated computing. These businesses depend heavily on Nvidia systems when expanding capacity for developers, model builders, and enterprise customers.
Every new AI data center requires processors, networking, storage, cooling, electricity, and sophisticated software before customers generate useful workloads. Jensen Huang believes Nvidia’s position across several of those layers strengthens its ability to capture continued infrastructure spending.
However, Nvidia’s growing reach has also attracted questions about investments in companies that later become major hardware customers. Critics argue some arrangements could appear circular when Nvidia’s capital eventually returns through purchases of Nvidia-powered infrastructure.
Jensen Huang rejected that criticism during his Goldman Sachs appearance, arguing that Nvidia invests only after seeing genuine customer demand. He said the company examines contracts and revenue before committing capital to businesses across its wider ecosystem.
His argument is straightforward: Nvidia may invest relatively small amounts while those companies generate much larger hardware purchases afterward. Huang believes that dynamic reflects commercial demand rather than artificial spending created solely through supplier financing.
Still, the debate matters because many young AI companies remain heavily dependent on outside funding to finance infrastructure. Some raise billions before spending substantial portions of that capital on Nvidia chips and Nvidia-powered cloud capacity.
That creates a possible vulnerability if artificial intelligence companies eventually become more efficient or struggle to monetize expensive services. Improved models could use fewer resources, while financial pressure might force some startups to reduce infrastructure spending dramatically.
Competition represents another challenge because nearly every major technology company wants to reduce dependence on Nvidia’s dominant hardware ecosystem. Amazon, Google, Microsoft, OpenAI, and other companies are developing custom processors for increasingly important artificial intelligence workloads.
Specialized chip startups are also targeting inference, where companies run already-trained models instead of developing them from scratch. These businesses believe custom architectures can provide lower costs or better efficiency for specific types of AI applications.
Jensen Huang remains confident because replacing Nvidia requires considerably more than building a competitive processor with impressive benchmark performance. Customers also depend on networking, software libraries, development tools, engineering support, and Nvidia’s enormous developer ecosystem.
That ecosystem creates a defensive advantage while artificial intelligence workloads continue changing faster than conventional infrastructure replacement cycles. Companies need systems supporting existing models while remaining flexible enough for new architectures arriving only months later.
Jensen Huang also expects AI demand to expand far beyond chatbots into coding, cybersecurity, scientific research, robotics, and autonomous agents. Each new category could create additional computing workloads instead of simply dividing existing demand among more providers.
Agentic AI could become particularly important because autonomous systems perform longer tasks than conventional chatbots answering individual questions. Those workflows require repeated reasoning, tool calls, context processing, and inference throughout every stage of an assignment.
Physical AI represents another opportunity as robots, industrial systems, vehicles, and factories begin using increasingly capable artificial intelligence. Nvidia already provides technology covering model development, simulation, edge computing, and deployment across many physical environments.
Jensen Huang therefore sees artificial intelligence demand broadening while Nvidia remains deeply embedded within the infrastructure supporting that expansion. His 70% forecast assumes companies will continue increasing investment as AI moves into more commercial and industrial workloads.
Still, no technology company can maintain extraordinary growth indefinitely once competitors become stronger and customers demand better economics. Nvidia will eventually face tougher pricing pressure, improving alternatives, changing architectures, and businesses seeking to reduce infrastructure spending.
That makes Jensen Huang’s forecast both ambitious and revealing about how strongly Nvidia views the current AI investment cycle. Delivering another 70% requires global infrastructure spending to remain enormous while Nvidia protects its position against increasingly determined competitors.
For now, Nvidia’s financial performance gives Jensen Huang considerable evidence supporting his confidence in continued artificial intelligence growth. The harder question is whether AI demand can keep expanding fast enough to make another extraordinary year appear almost ordinary.
