Why is Nvidia investing beyond the chip business?
Nvidia’s business does not end when it sells a chip. The company’s chips power AI systems, but the demand for those chips depends on whether AI companies can build products people will pay for. Investing in those companies gives Nvidia a direct stake in both sides of that equation.
Perplexity is a clear example. Nvidia is reportedly negotiating an investment at a valuation above $30 billion, more than 50 percent higher than Perplexity’s valuation in its funding round a year ago. Perplexity’s annualized revenue has also grown from $250 million to more than $750 million. One source of that growth is Perplexity Computer, an AI agent that performs tasks automatically.
That detail matters because AI agents generally use more tokens than simple question-and-answer tools. More token use means more computation, which can increase demand for the chips that run these systems. The same company that benefits financially when Perplexity grows may also supply the hardware Perplexity needs to operate.
This is not limited to one investment. Nvidia has made similar deals involving Poolside, Groq, and Enfabrica. The reported Groq deal was valued at $20 billion, while Enfabrica raised $900 million. Nvidia had also considered an acqui-hire of Perplexity, bringing its technology and staff into Nvidia directly.
There is a strategic layer too. Perplexity joined Nvidia’s Nemotron Coalition in March, an effort that promotes open AI models as a counterweight to Chinese development efforts. Nvidia is therefore supporting companies that can expand AI usage, strengthen its ecosystem, and potentially buy more Nvidia hardware.
The takeaway is simple: Nvidia is investing in the customers and applications that help create future chip demand.
What changed Perplexity's valuation so quickly?
Perplexity’s valuation appears to have risen quickly because investors are valuing it as more than an AI answer service. Nvidia is reportedly discussing an investment at a valuation above $30 billion, linking Perplexity’s software and traffic directly to Nvidia’s larger plan for AI infrastructure.
The unusual part is that the investment can support both sides of the market. Nvidia gains a relationship with a high-traffic AI product, while Perplexity gains access to capital for computing. If Perplexity uses that money to buy Nvidia GPUs, part of the investment can return to Nvidia as chip revenue. That is the circular financing concern raised around the proposed deal.
This is not proof that Perplexity’s underlying business suddenly became worth $30 billion. The reported valuation may reflect strategic value: securing a software partner, creating demand for GPU computing, and gaining an entry point into a widely used AI service. Those benefits can matter to Nvidia even if Perplexity’s direct profits are still uncertain.
Nvidia has recently made similar deals involving Poolside, Groq, valued at $20 billion, and Enfabrica, valued at $900 million. The pattern suggests that Nvidia is not only selling hardware. It is also putting money into companies whose growth may create future demand for that hardware.
That changes the valuation logic. A company can attract a higher price when its investors expect it to become both a useful product and a major customer. The key takeaway is that Perplexity’s higher valuation may reflect its position inside Nvidia’s AI ecosystem as much as its standalone financial performance.
How does Perplexity Computer drive revenue and compute demand?
Perplexity Computer matters because it places the company in the AI application layer, where revenue must come from software use rather than from selling chips. The supplied research does not provide Computer’s pricing, user numbers, or exact product features, so there is no reliable basis for claiming a specific revenue model. The central question is simpler: can Perplexity turn an AI application into a durable source of downstream revenue?
That question also explains the link to compute demand. A successful application creates repeated requests for model responses, and repeated requests require computing capacity. More usage can therefore support more revenue while also increasing demand for the infrastructure behind the service. For Nvidia, that creates a potential software partnership alongside its hardware business.
The unusual risk is that this connection can run in reverse. Perplexity has reportedly raised more than $1.7 billion, including investment from Nvidia, and its valuation could rise from about $20 billion to more than $30 billion after the new round. Critics describe the broader pattern as “circular financing” when invested capital allegedly flows back through chip orders. In that view, hardware purchases can make application growth look stronger without proving that customers will generate sustainable revenue.
Supporters see a different possibility: long-term software partnerships that help Nvidia participate in the application layer. Skeptics argue that the valuation only makes sense if Perplexity can build durable downstream AI revenues and avoid excessive dependence on financing or Nvidia’s ecosystem.
The takeaway is that Computer’s value is not measured by compute consumption alone. It must convert that consumption into lasting customer revenue. Otherwise, rising chip demand may reflect funding flows rather than a self-sustaining software business.
Why do AI agents consume more tokens than chatbots?
A chatbot usually responds to one request at a time. An AI agent is built to carry out a task, which can require several rounds of searching, reading, deciding, and producing an answer. Each round sends more text through the model, so the agent consumes more tokens than a simple chat exchange.
The important difference is not just the length of the final response. An agent may need to process the user’s request, review information, decide what to do next, and keep track of the work already completed. The model handles these steps through inference, the process that runs an AI system after it has been trained. More steps mean more inference, even when the user sees only one final answer.
That distinction matters for Nvidia. Its business depends not only on large companies buying chips for training models, but also on sustained demand for inference across enterprise and personal workflows. If AI agents become part of search and other daily tasks, they could create a broader stream of computing demand than occasional chatbot questions.
This also explains why Perplexity is relevant to Nvidia’s reported investment talks. Perplexity operates in AI search, a category closely connected to agents that gather and organize information for users. If these systems handle more involved requests, they may require more model execution behind each interaction.
The practical takeaway is simple: agents turn one visible question into a chain of hidden machine actions. That can make them more useful, but it also makes every successful interaction more expensive in computing resources. For Nvidia, wider agent adoption could help connect AI applications directly to continued demand for its inference chips.
Could Nvidia's investment money return as chip revenue?
Nvidia’s reported investment in Perplexity could create a financial loop: money moves into an AI company, and some of that company’s growth may later require more computing infrastructure. But that outcome is possible, not guaranteed.
The reported deal would value Perplexity at more than $30 billion, up from about $20 billion in its 2025 financing. Perplexity operates in AI search and agents, areas that require models to process requests, retrieve information, and produce answers. As those services expand, their operators need more computing capacity and supporting infrastructure.
That creates a potential path back to Nvidia. If Perplexity uses part of its new funding to expand its systems, some of the resulting infrastructure spending could eventually benefit Nvidia through demand for its chips. In that case, Nvidia would gain in two ways: it would own a stake in a fast-growing AI company, and it could also supply technology used to run that company.
The important limitation is that the available reporting does not confirm that Perplexity would spend the investment on Nvidia hardware. The financing itself has not been officially confirmed either. Perplexity declined to comment to Reuters, and Nvidia did not immediately respond to its request for comment.
So the investment should not be described as guaranteed chip sales. It is better understood as a bet on the expanding AI ecosystem, including search, agents, infrastructure, and enterprise workflows. If Perplexity grows, its computing needs may grow with it. The takeaway is simple: Nvidia’s investment could return as chip revenue, but only if Perplexity turns fresh capital into larger AI operations that require more computing power.
What does the Nemotron Coalition reveal about the partnership?
Nvidia and Perplexity already work together, so a possible investment would deepen an existing relationship rather than create one from scratch. That distinction matters. The two companies are connected through both technical collaboration and Nvidia’s broader plans for open AI models.
In March 2026, Nvidia announced the NVIDIA Nemotron Coalition. The group includes Perplexity, Mistral AI, Cursor, LangChain, Black Forest Labs, Reflection AI, and Thinking Machines Lab. Perplexity joined as an inaugural member, placing it among the companies helping shape Nvidia’s Nemotron model ecosystem.
Perplexity’s role is not simply to provide feedback from the sidelines. It is contributing expertise in model development and real-world AI deployment. That gives Nvidia access to lessons from an AI company building products for users, while Perplexity gains a direct connection to Nvidia’s model work and infrastructure.
Nvidia’s own documentation also indicates that Perplexity already relies heavily on Nvidia infrastructure. The exact terms of that dependence are not provided here, but the broader point is clear: Nvidia is both a technology supplier and a potential financial partner.
That combination can make the relationship more valuable for both sides. Nvidia can support a fast-growing AI application company while learning how its models and computing systems perform in production. Perplexity can strengthen its access to the hardware and model ecosystem behind its services.
The coalition therefore reveals a partnership with several layers: infrastructure, model development, deployment expertise, and possible financing. The important takeaway is that any investment would not be an isolated bet. It would build on an active technical relationship already tied to Nvidia’s Nemotron strategy.
Why might Nvidia prefer strategic investments to acquisitions?
Nvidia could gain exposure to Perplexity’s growth without taking on the cost and responsibility of owning the entire company. The reported deal is an investment in a funding round, not an acquisition. That distinction matters because Perplexity is already building a business with its own products, revenue, cloud arrangements, and plans to go public in 2028.
The potential upside is substantial. Perplexity’s annualised revenue reportedly rose from less than $250 million at the start of the year to more than $750 million. Its AI-powered search engine is also expanding through Perplexity Computer, a cloud-based agent that professionals use to automate computer tasks. An investment would let Nvidia participate if that growth continues, while Perplexity remains focused on developing and operating its products.
There is also a possible connection to Nvidia’s core business. Perplexity’s services require cloud computing, and the company signed a $750 million cloud agreement with Microsoft earlier this year. Nvidia would not need to own Perplexity to benefit from greater demand for the computing systems that support AI services. As Perplexity grows, more activity may flow through the infrastructure that uses Nvidia’s chips, although the available reporting does not specify how much Nvidia hardware the startup uses.
The price also changes the calculation. A funding round valuing Perplexity above $30 billion would give Nvidia a way to buy a stake in a rapidly appreciating company, rather than commit to purchasing the whole business at that valuation. Perplexity’s value was reported at $20 billion last September, and the new figure would be more than 50% higher than its previous financing.
The takeaway is simple: a strategic investment can give Nvidia financial upside and a closer relationship with an important AI customer, while leaving Perplexity independent.
What risks come with a $30 billion-plus AI application valuation?
A valuation above $30 billion leaves little room for disappointment. Perplexity’s reported valuation would be 50% higher than it was less than a year ago, while its annual recurring revenue has passed $750 million, reportedly three times its level at the start of 2026. That growth is impressive, but the price also assumes that rapid expansion will continue.
The risk is not simply that revenue stops growing. A company can grow quickly and still struggle to justify a much higher valuation if growth slows, costs rise, or investors demand stronger evidence of durable profits. The available figures also need careful interpretation. Perplexity’s reported $750 million in annual recurring revenue is not the same as the $750 million cloud agreement it signed with Microsoft earlier this year. One describes recurring business revenue, while the other is a cloud commitment. Treating them as the same would make the company look stronger than the evidence supports.
The planned 2028 public offering adds another pressure point. A private valuation can rise during intense funding interest, but public-market investors will eventually examine revenue quality, spending, and dependence on expensive computing resources in much greater detail. Nvidia’s reported interest may provide capital and strategic support, but it can also reinforce expectations that Perplexity must expand quickly.
There is also a circular risk in the structure: AI companies need vast computing resources, and chip companies benefit when those resources are purchased. If growth fails to match the valuation, the same spending that helped build the business can become a burden.
The takeaway is simple: $30 billion prices Perplexity for continued acceleration, not merely success. A strong product and rising revenue still have to catch up with that expectation.
How could this affect Perplexity's path to an IPO?
Perplexity’s reported valuation could rise above $30 billion before the company reaches the public market. That would be a major change from the $20 billion valuation reportedly finalized in September last year, and it would give the company a much higher private-market benchmark to defend in an IPO.
The unusual part is that the valuation increase is arriving alongside a sharp rise in revenue. Perplexity’s annualized revenue has climbed above $750 million, compared with less than $250 million at the beginning of the year. That growth has been partly linked to Perplexity Computer, its cloud-based AI agent for automating computer tasks for professional users.
Nvidia’s possible participation could add more than capital. A major technology company investing in the round may give Perplexity additional credibility as it builds toward a public offering. It could also help investors view the company as more than an AI search product, especially as Perplexity expands into professional task automation.
But a higher private valuation also creates a higher bar. If public-market investors do not accept a valuation above $30 billion, an IPO could require the company to price below its latest private round. That would make the offering harder to present as a success, even with strong revenue growth.
The key takeaway is that Nvidia’s potential investment could strengthen Perplexity’s IPO story, but it would not settle the question of public-market value. The company would still need to show that its rapid revenue growth can support the expectations attached to a $30 billion-plus valuation.
What should AI teams learn from Nvidia's ecosystem strategy?
AI teams should pay attention when an infrastructure company considers investing in an application company. Nvidia is reportedly in talks to back Perplexity at a valuation above $30 billion, while Perplexity is seeking funding and its revenue has reportedly reached $750 million. The important point is not simply the size of the possible investment. It is the connection between the systems that run AI and the products people use.
Perplexity has also reportedly signed a $750 million agreement with Microsoft to use Azure cloud infrastructure. That means its growth depends on more than model quality or user demand. It also depends on access to large-scale computing infrastructure. Nvidia supplies a critical part of the hardware used for AI workloads, while Microsoft supplies cloud capacity. A possible Nvidia investment would place another major infrastructure company closer to a fast-growing AI application.
For AI teams, the practical lesson is to treat infrastructure choices as part of product strategy. Cloud contracts, hardware availability, and model-serving costs can shape how quickly a product grows and how much revenue becomes profit. Partnerships may create advantages, but they can also make a company more dependent on a small number of powerful suppliers.
Perplexity CEO Aravind Srinivas said in June that the company plans to pursue an IPO in 2028, regardless of how planned OpenAI and Anthropic listings perform. That signals an ambition to build an independent business, even while relying on major infrastructure partners.
The takeaway is simple: AI companies should plan for both sides of the stack. Build a product users want, but also understand who provides the computing needed to deliver it. In AI, the infrastructure behind the product can become part of the company’s strategic future.