Nvidia expands its competitive advantage from GPUs to data orchestration in the Vera Rubin architecture
According to observers following the earnings release, Nvidia is expanding its advantage beyond GPUs into data orchestration in data centers. According to the company, the Vera Rubin architecture with Vera CPUs speeds up data flow between memory and GPUs by roughly 3×, which is becoming more important as gigawatt-scale data centers are built.
According to statements made after the release of earnings this week, Nvidia is expanding its competitive advantage beyond GPU chips themselves into the orchestration layer of data centers. Nvidia shares rose roughly tenfold between the beginning of 2023 and mid-2025, but growth slowed over the past year due to investor concerns about increasing competition in the GPU chip market (Amazon and Google, for example, are developing their own chips). However, a new interpretation that emerged after the latest earnings, according to the source, indicates that Nvidia also has a lead in the systems surrounding GPUs.
This centers on the Vera Rubin architecture, which combines Rubin GPUs with other components — CPUs named Vera, the Groq 3 LPX inference accelerator, and corresponding racks for storage and networking. According to Jason Hardy, vice president of storage technologies at Nvidia, Vera CPUs primarily address the flow of data between memory and GPUs, because memory in any single server is always limited, and getting data to the right place at the right time is an increasingly difficult task. According to Hardy, Nvidia measured a speedup of roughly 3× in these operations, allowing flash storage capacity to be used more fully without limiting performance (a so-called bottleneck).
Competitors are addressing a similar problem differently — according to its own statements, OpenAI aims to minimize data movement with its Jalapeño chip by keeping the entire task within a single interconnected system. The principle is the same — efficiency improves through smarter management of data flow as well as more computing power — but this opens up a new layer of infrastructure over which companies will compete.
The source states that this shift does not automatically mean victory for Nvidia, because it will have to compete with rival chip manufacturers and hyperscalers at this new level as well. According to the source, however, Nvidia has a substantial lead in this area, at least in the early stages.
Why it matters
According to the source, competition among AI chip manufacturers is shifting from raw GPU performance to the ability to efficiently manage data flow across the entire system — through memory, networking, and storage. For operators of gigawatt-scale data centers, this means that the choice of infrastructure supplier will depend not only on the performance of GPU chips themselves, but also on how well the entire system can maintain high efficiency (fewer watts per token) as the volume of computation grows.
Relevant practical impact
What this means
For a business
For companies operating or purchasing AI infrastructure at scale, the competitive landscape is expanding from raw GPU performance to the ability to efficiently orchestrate data across the entire system; this changes which suppliers and architectures need to be compared when making investment decisions about gigawatt-scale data centers.
StrategyCheck the original
Event sources
only one source so far · 1 publisher, 1 independent. We count feeds from the same owner only once.