Nvidia's new 'Vera' central processors have been selected by cloud-computing unicorn Modal to power its AI agent infrastructure. This move marks a strategic expansion for Nvidia into the CPU market, challenging long-time leaders Intel and AMD. Investors are monitoring how this shift toward a multi-vendor hardware strategy affects long-term competition in the data center sector.
Nvidia is challenging the long-standing dominance of Intel and AMD in the server chip market with its new 'Vera' central processors. In a significant move, cloud-computing startup Modal has selected Nvidia’s Vera chips to power its AI agent infrastructure, shifting away from the traditional x86 architecture provided by Intel and AMD. This is a critical development for Nvidia, which is trying to build a major position in the data center chip space.
Nvidia’s Push Into CPUs
For years, Nvidia has been best known for its graphics processing units, or GPUs. By entering the central processor unit (CPU) market, the company is aiming for a much larger piece of the data center industry, which it estimates is worth $200 billion. Nvidia has set an internal revenue target of $20 billion from Vera chip sales for the current fiscal year. This indicates the company is looking to move beyond its core business and compete directly with established incumbents in server infrastructure.
The decision by Modal to adopt Nvidia’s new hardware was driven by performance needs. According to company reports, internal testing showed that the Vera processors could double the workload capacity for its AI agents compared to the older Intel and AMD configurations. However, this is not a simple swap; Modal’s engineering team has had to undertake significant software changes to make the integration work effectively, highlighting that performance gains often come with complex operational requirements.
Managing Hardware Competition
Despite this win for Nvidia, the market remains highly competitive. Modal has clarified that it is not fully moving away from its previous partners. Instead, it is adopting a mix-and-match approach, where workloads are routed to whichever hardware performs the specific task best. This strategy allows companies like Modal to avoid 'hardware lock-in'—the risk of being tied to a single supplier for all their needs.
For investors, this development highlights the changing nature of AI infrastructure spending. Data center operators are increasingly prioritizing 'performance-per-watt'—essentially, how much work they can get done for every unit of electricity they pay for. While Nvidia has secured a major foothold with a high-growth startup like Modal, the company will face continued pressure as Intel and AMD work to improve their own chip offerings. The success of this move for Nvidia will depend on whether it can prove its performance advantage is sustainable and convince more large-scale data center operators to follow Modal’s lead.