SAN FRANCISCO, Oct 19: Nvidia has strengthened its position in the artificial intelligence hardware market after cloud computing startup Modal announced plans to adopt the company’s new Vera central processing units, marking a significant step in Nvidia’s efforts to compete with established processor manufacturers Intel and AMD.
Modal, which provides computing resources to developers building and deploying artificial intelligence applications, said it would begin using Nvidia’s Vera processors to run AI agents. The decision represents an expansion of Nvidia’s influence beyond its established graphics processing unit business, which has become a key component of modern AI training and computing infrastructure.
According to the announcement, early testing showed that the Vera processors could double the amount of work completed by AI agents compared with the Intel and AMD processors currently used by Modal. The reported improvement encouraged the startup to incorporate Nvidia’s technology into its platform, although the company will continue operating processors from its existing suppliers.
The development comes as technology companies invest heavily in infrastructure designed to support increasingly sophisticated AI applications. Unlike conventional chatbots that primarily respond to individual prompts, AI agents can undertake extended tasks, coordinate multiple operations and work through sequences of instructions with varying degrees of independence.
Such systems require computing infrastructure capable of handling continuous workloads efficiently. Processor performance, software compatibility and operating costs are therefore becoming increasingly important considerations for cloud providers and AI developers.
Nvidia targets a broader processor market
Nvidia has built its reputation around graphics processors that are widely used to train and operate AI models. Its latest push into central processing units reflects a strategy to capture a larger share of the computing infrastructure required by the expanding AI industry.
Central processing units, or CPUs, perform general-purpose computing tasks and coordinate different operations within a computer system. Graphics processing units, or GPUs, are particularly effective at handling the parallel calculations needed for many AI workloads.
Both types of processors play important roles in modern data centres. As AI applications become more complex, technology companies are increasingly examining how different processors can be combined to improve performance and reduce the cost of running large-scale services.
Nvidia’s Vera chips are intended to compete in a market historically dominated by Intel and AMD. Winning customers in this segment could give the company an opportunity to provide a more comprehensive range of computing hardware to businesses building AI infrastructure.
Modal’s decision is significant because the startup had previously relied exclusively on Intel and AMD processors for the workloads covered by the announcement. Its willingness to introduce a new supplier suggests that Nvidia’s CPU offering is attracting interest beyond its traditional GPU customer base.
However, the move does not mean Modal has decided to abandon its existing hardware partners. Chief Executive Erik Bernhardsson said the company would operate a mixed fleet of processors and allocate workloads according to which hardware delivers the best performance.
That approach could provide a useful test of Nvidia’s ability to compete on practical computing performance rather than relying solely on its established reputation in AI acceleration.
AI agents create new computing demands
The growing use of AI agents is creating new opportunities for companies supplying processors, cloud services and software tools.
Businesses are exploring agents for tasks such as software development, research, information processing and workflow automation. These applications may require systems to perform multiple steps, interact with different services and continue working over extended periods.
The underlying computing requirements can differ from those of a single AI model responding to a short request. An agent may need to process instructions, retrieve information, execute code, evaluate results and repeat operations before completing a task.
Improving the speed at which these activities are completed can help developers deliver services more quickly. It may also allow cloud providers to serve more customers using their available computing resources.
Modal’s early tests indicated that Nvidia’s Vera processors could increase the amount of agent-related work completed compared with the Intel and AMD chips already available to the startup. However, the results should be viewed in the context of the specific workloads tested rather than as proof that Nvidia’s processors outperform competing products in every application.
Bernhardsson also acknowledged that Intel and AMD are developing newer processors that could narrow the performance difference. Modal intends to evaluate those products when they become available.
This suggests that competition in the CPU market will depend on continuing improvements rather than a single set of benchmark results.
Modal’s rapid expansion adds commercial significance
The processor decision comes as Modal expands its cloud-computing business. The company raised $750 million in a funding round led by Accel last week, reaching a valuation of $15.75 billion.
Modal said its service usage had tripled and its revenue had increased more than 15-fold year over year. It also indicated that it was on track to generate $1 billion in sales over the following year.
These figures highlight the growing demand for specialised computing services as companies develop AI products and seek infrastructure capable of supporting them.
For Nvidia, securing a customer with ambitious expansion plans could provide an opportunity to demonstrate the commercial value of its new processors. For Modal, access to another hardware supplier may provide additional flexibility when selecting the most suitable computing resources for customers.
Nevertheless, incorporating Vera processors will require significant software changes on Modal’s side. Hardware performance alone does not determine whether a processor can be deployed successfully in a production environment.
Cloud providers must also consider software compatibility, reliability, availability, power consumption and the time required to adapt existing systems. These factors can influence the total cost of operating a service.
A mixed hardware strategy could help Modal avoid dependence on a single supplier while allowing it to adopt new technology where measurable benefits justify the investment.
A potential multibillion-dollar opportunity
Nvidia has told investors that it expects Vera processor sales to generate at least $20 billion during its current fiscal year. The company has estimated the broader market opportunity at approximately $200 billion.
It has also indicated that OpenAI, Anthropic and Oracle, alongside AI startup Perplexity, plan to use the processors. Such prospective deployments could help Nvidia establish a stronger presence in the central processing market.
The opportunity, however, will depend on whether the company can translate early customer interest into sustained demand. Intel and AMD have established positions in the CPU industry, supported by extensive software ecosystems and longstanding relationships with hardware manufacturers and enterprise customers.
Nvidia will need to demonstrate that its processors provide sufficient advantages in relevant workloads to persuade more businesses to adopt them.
The emergence of AI agents could be an important factor in that competition. If businesses increasingly rely on systems that carry out complex tasks independently, demand for efficient computing infrastructure may grow across multiple hardware categories.
For customers, competition between processor manufacturers could create more choices and encourage improvements in performance and cost efficiency. For cloud providers, access to different types of hardware could make it easier to match specific workloads with the most suitable technology.
Nvidia’s agreement with Modal therefore represents more than an additional customer win. It offers an early indication of how the company is attempting to expand its role across the computing systems underpinning the AI economy.
The longer-term outcome will depend on real-world performance, software integration and the ability of competing manufacturers to respond with their own products.