Nvidia reported 20,000 AI startups building on its platform.

In its Q1 2025 earnings call, Nvidia CEO Jensen Huang highlighted the explosive adoption of generative AI across industries, noting there are "some 15,000, 20,000 startups in all different fields from multimedia to digital characters, design to application productivity, digital biology," all leveraging Nvidia's accelerated computing platform. He emphasized that this wave is driving a "foundational, full-stack computing platform shift" as computing moves from information retrieval to generating intelligent outputs. "The computer is now generating contextually relevant, intelligent answers," Huang explained. "That's going to change computing stacks all over the world. Even the PC computing stack is going to get revolutionized." Demand from consumer internet companies, enterprises, cloud providers, automotive companies and healthcare organizations is "incredible" as they invest in "AI factories" built on thousands of Nvidia GPUs.

To address this demand, Nvidia began shipping its H100 "Hopper" architecture GPUs in Q1 and announced its next-generation "Blackwell" platform, which delivers 4-30X faster AI training and inference than Hopper, with over 100 Blackwell systems from major computer makers launching this year. However, despite posting record-breaking $26 billion in revenue in Q1, customer demand is significantly outpacing Nvidia's ability to supply. "We're racing every single day," Huang said regarding fulfillment efforts. The company expects demand for the current H100 flagship GPU to exceed supply for some time even as it transitions to the H200 and Blackwell platforms. This urgency stems from competitive pressures: "The next company who reaches the next major plateau gets to announce a groundbreaking AI, and the second one after that gets to announce something that's 0.3% better," Huang explained. "Time to train matters a great deal. The difference between time to train that is three months earlier is everything." Huang predicted this supply crunch will persist well into next year, stating "Blackwell is well ahead of supply and we expect demand may exceed supply well into next year."

The economics driving this demand are compelling for cloud providers and other companies. Huang stated that "for every $1 spent on Nvidia AI infrastructure, cloud providers have an opportunity to earn $5 in GPU instance hosting revenue over four years." He provided a specific example: a language model with 70 billion parameters using Nvidia's latest H200 GPUs could generate 24,000 tokens per second and support 2,400 concurrent users per server, meaning "for every $1 spent on Nvidia H200 servers at current prices per token, an API provider can generate $7 in revenue over four years." These economics are further improved by Nvidia's ongoing software optimizations—in the latest quarter, improvements delivered a 3X speedup on the H100, enabling a 3X cost reduction for customers. This strong return on investment is fueling demand from cloud giants like Amazon, Google, Meta, Microsoft and Oracle as they race to provision AI capacity.

Beyond GPUs, Nvidia is leveraging its major position in datacenter networking to expand its AI platform. While Infiniband has driven strong year-over-year growth in networking, Huang emphasized that Ethernet represents a major new opportunity. In Q1, the company began shipping its Spectrum-X platform, optimized for AI workloads over Ethernet. "Spectrum-X opens a brand new market to Nvidia networking and enables Ethernet-only datacenters to accommodate large-scale AI," Huang said. "We expect Spectrum-X to jump to a multi-billion dollar product line within a year." Huang stated Nvidia is "all-in on Ethernet" and will deliver a major roadmap of Spectrum switches to complement its Infiniband and NVLink interconnects, allowing the company to target everything from single-node AI systems to massive clusters. The company also began sampling its 51.2 terabit per second Spectrum-4 Ethernet switch during the quarter.