Nvidia’s Perplexity Talks Show AI Search Is Becoming Chip Strategy

Nvidia is reportedly discussing an investment in Perplexity at a valuation above $30 billion, and the story says more about Nvidia’s strategy than one startup’s fundraising.

Reuters-linked reports, citing The Information, say Nvidia is in talks to invest in Perplexity’s new equity round. The report also says Nvidia had earlier considered a technology licensing arrangement and hiring some of Perplexity’s staff before the discussions shifted toward a direct equity investment.

That pattern should sound familiar. Nvidia is no longer acting only like a chip supplier waiting for AI companies to buy GPUs. It is increasingly placing itself around the companies that generate AI demand, shape developer behaviour and create the workloads that keep data centres full.

Perplexity is a useful target for that kind of investment because AI search sits close to everyday user behaviour. A chatbot can answer a prompt. An AI search engine can become the place where people research products, compare companies, summarize news, look up documents and make decisions. That means heavy inference demand, constant model use and a direct challenge to Google’s search model.

For Nvidia, that demand matters. The more AI search grows, the more compute is needed to serve queries, retrieve sources, summarize results and personalize answers. Investing in a company like Perplexity gives Nvidia exposure to the application layer that helps justify the hardware layer.

This is not Nvidia’s first move in that direction. The company has also been linked to licensing and talent arrangements in AI, including the reported Poolside deal, and it has backed infrastructure plays such as Cloverleaf’s AI data-centre power-site strategy.

The Perplexity valuation is also notable. A $30 billion-plus valuation would be a huge leap for a company competing in a market dominated by Google, OpenAI, Microsoft and Meta. It shows how strongly investors believe search is being rewritten by AI, even before the economics are fully proven.

There are real questions. AI search is expensive to run, publishers are worried about traffic losses, and hallucination risk remains a problem when users rely on answers instead of clicking through to original sources. Perplexity also has to show that it can turn usage into durable revenue without becoming too dependent on one investor, one model provider or one infrastructure path.

Still, the strategic logic is clear. Nvidia wants to be part of the AI economy wherever compute demand is created: chips, data centres, model factories, developer tools and now potentially search interfaces.

If the Perplexity investment happens, it will reinforce a simple point. Nvidia’s AI power no longer comes only from selling silicon. It is starting to look like a company building influence across the entire demand chain for AI.