Nvidia has secured fast-growing AI cloud startup Modal as a customer for its new Vera central processor, marking an early breakthrough in a market long dominated by Intel and AMD.
Modal, which provides computing infrastructure to AI developers, said it will begin using Vera CPUs to run AI agents capable of performing long-running tasks.
Until now, the startup had relied entirely on Intel and AMD processors for this type of work.
The move is significant because Nvidia is already dominant in AI graphics processors, or GPUs. Vera gives the company a new route into the equally important market for general-purpose processors inside AI data centres.
FCRF’s Chief AI Officer Certification Returns on November 14 to Prepare Future AI Leaders
Modal Says Vera Doubled AI-Agent Performance
Modal said its early testing found Nvidia’s Vera processor could complete roughly twice as much AI-agent work as the Intel and AMD chips currently available in its infrastructure.
That performance persuaded the company to add Nvidia CPUs despite the software changes required to support a new processor architecture.
Modal Chief Executive Erik Bernhardsson stressed that the company is not abandoning Intel or AMD.
Instead, it plans to operate Vera alongside processors from its existing suppliers.
Modal’s system can then decide which type of hardware is best suited to a particular workload.
That approach matters because AI infrastructure is becoming increasingly specialised.
Different chips can perform better depending on whether a customer is training an AI model, serving responses, running software code or operating autonomous AI agents over long periods.
Modal also noted that its comparison involved Intel and AMD processors already on the market.
Both companies are developing newer chips that could narrow Nvidia’s advantage once they become available.
What Makes a CPU Different From Nvidia’s AI GPUs
Nvidia built its AI dominance primarily through GPUs.
A GPU is designed to perform large numbers of calculations in parallel, making it especially effective for training and running large AI models.
A CPU plays a different role.
It acts as the more general-purpose processor inside a server, handling operating systems, application logic, scheduling and many tasks that do not require massive parallel processing.
Most AI data centres therefore use both.
Historically, Intel and AMD have supplied much of the CPU side, while Nvidia has dominated AI accelerators.
Vera is Nvidia’s attempt to capture more of the entire computing system.
The processor is designed to work closely with Nvidia’s GPUs and networking technology, allowing the company to offer customers a more complete AI infrastructure platform.
AI agents may provide an especially attractive market because they can run for hours, execute software and continuously interact with other systems.
That places demands not only on GPUs but also on the CPUs coordinating those tasks.
Nvidia Targets $20 Billion From Vera
Nvidia has told investors it expects at least $20 billion in Vera CPU sales during the current fiscal year.
The company estimates the broader market opportunity at around $200 billion.
Modal joins several major technology companies that Nvidia says plan to use Vera.
OpenAI, Anthropic and Oracle have already been named as customers, alongside AI startups including Perplexity.
Winning Modal provides another important signal because the startup had previously built its CPU infrastructure entirely around Nvidia’s established rivals.
Modal itself is expanding quickly.
The company raised $750 million last week in a funding round led by Accel that valued it at $15.75 billion.
It expects to record around $1 billion in sales over the next year.
Usage of its computing services has tripled, while revenue has increased more than fifteen-fold from a year earlier.
That rapid growth makes its hardware choices particularly valuable for chipmakers competing to supply the next generation of AI data centres.
Intel and AMD Face a New Competitive Threat
Nvidia’s entry into CPUs changes the competitive picture for Intel and AMD.
The two companies have spent decades competing against each other for the processor market used in servers and personal computers.
AI has already weakened that traditional balance by making GPUs far more important inside data centres.
If Nvidia can successfully combine its GPUs with its own CPUs, networking products and software, customers may increasingly buy larger portions of their infrastructure from a single supplier.
That could strengthen Nvidia’s control over the AI computing stack.
Modal’s decision does not mean Intel and AMD are being pushed out.
The startup is deliberately keeping multiple suppliers and says future processors will be tested based on performance.
But Nvidia no longer has to compete only for the GPU portion of an AI system.
Vera gives it a direct route into workloads that Intel and AMD previously controlled.
What this means for you:
Nvidia is expanding beyond the GPUs that made it dominant in AI. If Vera gains wider adoption, the company could control a larger share of the hardware powering AI agents and data centres, increasing competitive pressure on Intel and AMD.
Follow for daily updates on cybercrime, corporate fraud, DFIR, hacking, investigations, and digital forensics
