Interesting News

Nvidia Just Poured Billions Into Perplexity and Poolside — Here’s the AI Power Play Behind It

September 1, 2026 by BenchMark Website Design No Comments

Nvidia doesn’t just make the chips the AI industry runs on anymore. Increasingly, it’s the one deciding who gets to run at all.

In the space of a single week, the world’s most valuable company has been reported to be closing in on two enormous bets that have nothing to do with silicon and everything to do with who controls the next generation of AI models. First, Bloomberg reported that Nvidia is committing roughly $6 billion to AI startup Poolside — a deal that includes licensing its technology and absorbing more than 100 of its engineers. Then, just days later, The Information reported that Nvidia is in talks to invest in AI search company Perplexity at a valuation north of $30 billion. Together, the two deals — first flagged in Tech Startups’ August 24 roundup — mark a striking escalation in Nvidia’s transformation from chip vendor to full-stack AI kingmaker, and they say a lot about where the AI arms race with China is headed next.

The Nvidia Perplexity investment: buying into the search wars

Perplexity has spent the past two years positioning itself as the AI-native alternative to Google search — a chatbot-style answer engine backed by heavyweights including Jeff Bezos and SoftBank. Business is booming: according to Benzinga, the startup’s annualized revenue has rocketed from under $250 million at the start of 2026 to more than $750 million now, fueled by growing demand for its search tools and its new “Perplexity Computer” AI agent product aimed at automating computer-based tasks for professionals.

A Nvidia-backed round at a $30 billion-plus valuation — up sharply from roughly $20 billion just months earlier — would cement Perplexity’s status as one of the hottest names in consumer AI, and give Nvidia a direct stake in one of the most GPU-hungry applications on the market: real-time AI search and agents. Perplexity already has a $750 million infrastructure agreement with Microsoft Azure and reportedly plans to pursue an IPO as early as 2028, according to Tech Startups.

The Nvidia Poolside deal: buying its way into building models

The Poolside deal is even more telling. Poolside is a lesser-known AI startup that’s been quietly building foundation models and coding-focused AI systems. Under the new arrangement, Nvidia isn’t just writing a check — it’s absorbing capability. Roughly $1 billion of the $6 billion package is a direct equity investment, according to Tech Startups, with the rest structured as technology licensing — and more than 100 Poolside engineers are moving over to work on Nvidia’s own “Nemotron” model family.

The goal, per multiple reports including Forbes and The Daily Caller, is building one of the most capable U.S.-made open-weight AI models yet — a direct answer to the wave of powerful, freely downloadable Chinese models like DeepSeek, Alibaba’s Qwen and Moonshot AI’s 2.8-trillion-parameter Kimi K3, which have been winning over developers worldwide with lower costs and fewer restrictions.

Why an “unremarkable” chipmaker became AI’s biggest kingmaker

It’s easy to forget that Nvidia’s core business is still selling graphics processors. But 2026 has made clear the company sees its future as something much bigger: the connective tissue of the entire AI economy. According to TechCrunch’s rundown of Nvidia’s portfolio, the company took part in 67 venture deals in 2025 alone, backing OpenAI, Anthropic, xAI, Mistral, Cursor and dozens of infrastructure and robotics startups — often in the same “circular” pattern critics have flagged, where Nvidia invests cash into a company that then spends much of it buying Nvidia chips.

The Perplexity and Poolside deals extend that playbook into new territory. Investing in Perplexity buys Nvidia influence over one of the fastest-growing AI application-layer companies — and a hedge against Google’s dominance in search. The Poolside deal goes further still, turning Nvidia into an actual model builder, not just an infrastructure supplier, arguably competing with some of its own customers like OpenAI and Anthropic even as it also backs them.

What it means for the AI arms race in 2026

Taken together, the two deals show Nvidia hedging every layer of the AI stack — chips, cloud infrastructure, consumer applications, and now the models themselves — while explicitly positioning U.S. open-weight AI as a counterweight to China’s fast-advancing open model ecosystem. With Beijing-linked labs giving away increasingly powerful models for free, Washington and Silicon Valley have grown anxious about ceding the “default” AI stack to Chinese technology. Nvidia’s Poolside gambit is a bet that America needs its own credible open-weight flagship — and that whoever builds it will need Nvidia’s chips to run it.

The bigger picture is a company using its unprecedented market value — and its equally unprecedented centrality to every AI company’s cost structure — to make sure that no matter who wins the AI race, Nvidia already owns a piece of the winner. Whether that concentration of influence is good for the broader AI ecosystem is a separate question. But for now, the AI arms race of 2026 increasingly runs through one company’s investment committee.