© HelloRF Zcool / Shutterstock.com
Nvidia (NASDAQ:NVDA) has become the face of the AI revolution. Its graphics processing units have powered the breakthrough advances in generative AI, enabling applications from chatbots to image generation and beyond. As a result, the company has experienced enormous growth, reaching a market capitalization of $4.6 trillion.
Analysts and investors remain focused on growing competition in the AI chip market, where Nvidia faces pressure to continue advancing new, more powerful chips to stay ahead. The belief is that it must innovate rapidly with architectures like Blackwell and Rubin to maintain its lead, or risk ceding the field to the competition.
Yet one key number has emerged that underscores why Nvidia remains the undisputed king of AI — and is likely to stay that way for the foreseeable future.
Growing Competition in AI Accelerators
Despite Nvidia’s dominance, concerns persist about potential market share losses. Competitors such as Advanced Micro Devices (NASDAQ:AMD), Broadcom (NASDAQ:AVGO), and Marvell Technology (NASDAQ:MRVL) are gaining traction in AI chips. Major cloud providers and tech giants, including Amazon (NASDAQ:AMZN) with its Trainium chips and Alphabet (NASDAQ:GOOG)(NASDAQ:GOOGL) with Tensor Processing Units, are developing custom accelerators to reduce or replace their reliance on Nvidia’s products.
These in-house efforts aim to lower costs and optimize specific workloads, particularly inference. Analysts project that custom silicon from hyperscalers could account for a larger portion of the AI accelerator market in coming years. Big Tech companies continue heavy investments in their own chips, partnering with designers like Broadcom to challenge Nvidia’s position.
The Power of Nvidia’s Installed Base
It is true that Nvidia can’t rest on its laurels amid this intense competition. It must sustain rapid innovation cycles and expand its software ecosystem, including CUDA, to keep developers tied to its platform as the data center buildout continues. However, the company’s existing strengths provide a real and durable substantial moat.
Using data from company reports and UBS estimates, analysts at Creative Strategies compared Nvidia’s data center compute revenue from all of its GPU generations — Ampere, Hopper, Blackwell, Blackwell Ultra, Rubin, and beyond — and compared them to peers including AMD, Broadcom, and Marvell. The results are staggering: Hopper revenue alone exceeds the combined compute revenue of all its rivals combined.
From the beginning in early 2023 when the AI revolution really sprung into the public consciousness until today, Hopper GPUs remain a dominant force for growth. Even as Blackwell and Rubin ramp up, Hopper’s ongoing sales — driven by the massive installed base for training, inference, and other workloads — outpace competitors’ collective offerings.
While the Blackwell and Blackwell Ultra GPUs now surpass Hopper and will accelerate their growth in the quarters to come, Hopper’s sustained cash flows provide a revenue foundation for Nvidia’s aggressive R&D for future architectures. This has funded the development and production of Blackwell and Rubin, allowing an annual release cadence that competitors struggle to match. The prior-generation chips reinforce Nvidia’s dominance while enabling advances that widen the performance gap.
Key Takeaway
This single metric of Hopper revenue surpassing all peers combined illustrates Nvidia’s structural lead in AI. The installed base generates enormous ongoing revenue, supporting innovation and locking in customers to its ecosystem, which deters share erosion.
While competition intensifies, Nvidia’s position remains as strong as ever. Investors can confidently view any significant price weakness as a buying opportunity for Nvidia stock. Given the chipmaker’s history of leadership in powering the AI revolution, and no clear threat to its dominance, investors can be assured Nvidia’s leadership atop AI will remain well into the future.