As dark clouds gather over the artificial intelligence construction boom, NVIDIA's CEO Jensen Huang shows no signs of resting on his laurels or halting his expansion efforts.
Just this past Sunday, reports emerged that NVIDIA is in talks to invest in Perplexity, a company that began as a search engine provider but has pivoted into an AI agent services firm with remarkably strong revenue growth this year. The potential investment could push Perplexity's valuation beyond $30 billion, and the two parties have also explored technology licensing agreements.
Only days before that development, Huang completed another rapid-fire investment round, marking the third in a series of NVIDIA's strategic plays. This latest target specializes in securing land and power supplies for AI data centers. As reported, the move aims to lock in hardware package orders during the early stages of data center projects while proactively mitigating potential chip supply glut risks.
Cloud service providers and other data center developers are grappling with delays in power infrastructure and unexpected cost overruns, which could easily lead to inventory backlogs of chip supplies. Speaking of unforeseen cost increases, server manufacturers have already notified clients that systems bundled with next year's flagship Blackwell and Rubin NVIDIA chips will carry price tags 17% higher. For a 1-gigawatt data center, that translates to at least $5 billion in additional costs.
Huang's flurry of activity extends equally into the AI model development arena. Last week, NVIDIA finalized a deal to pay $6 billion by the end of next year for partial software licensing from startup Poolside, along with acquiring 100 of its R&D personnel. Poolside initially developed code-generation AI agents before expanding into data center and open-weight model territories. This transaction underscores Huang's determination to propel Nemotron into the upper echelon of world-class open-source models.
If you want deeper insight into that initiative, our previous coverage on Nemotron provides comprehensive details. But here's the question lingering in my mind: given the nearly one-sided public sentiment against AI and its supporting data centers, does Huang have a viable strategy to turn the tide? Activists opposing data center construction harbor numerous economically misguided beliefs, and dispelling those misconceptions proves extraordinarily challenging.
As recent reporting highlights, even the relentless messaging from Sam Altman, Dario Amodei, and Demis Hassabis about AI's potential to cure cancer has lost its persuasive power with the general public. The same reporting notes that even pharmaceutical industry professionals who once championed AI are beginning to develop resistance toward frontier AI laboratories.
Of course, none of these concerns will surface in NVIDIA's upcoming July-quarter earnings report due this week. Instead, revenue figures will likely continue their positive trajectory. Huang and his astute CFO Colette Kress will probably release multiple favorable operating metrics as a preemptive response to the concerns the market will inevitably raise: the increasingly challenging economic and political environment facing NVIDIA's major chip customers and partner data center operators.