The boundary between infrastructure provider and application layer is blurring. On August 23, 2026, The Information reported that Nvidia is in discussions to invest in Perplexity AI at a valuation exceeding $30 billion. While these discussions remain unconfirmed by either party, the move represents a significant escalation in Nvidia’s capital deployment strategy, signaling a transition from merely supplying the hardware that powers AI search to securing a permanent stake in the inference layer itself.
This potential investment follows a distinct pattern observed in Nvidia’s recent capital flows. Before pivoting toward a traditional equity stake, Nvidia reportedly explored paying billions to license Perplexity’s technology and acquire specific talent. This licensing-to-equity pivot mirrors the playbook Nvidia executed with Poolside in August 2026 and Groq in December 2025. By shifting from transactional licensing to equity ownership, Nvidia is effectively locking in long-term demand for its compute resources while gaining influence over the product roadmaps of the companies most likely to consume them.
Perplexity’s valuation trajectory has been nothing short of aggressive. From a $121 million valuation in April 2023 to the current $30 billion-plus discussions, the company has seen roughly 175x growth in just 30 months. This rapid ascent is underpinned by a scaling revenue base, though market figures require careful scrutiny. According to Sacra and Tracxn, Perplexity’s ARR has climbed from approximately $63 million at the end of 2024 to a range of $450 million to $500 million as of April 2026. While some external reporting has cited a $750 million ARR figure, this appears to be a conflation with the $750 million Microsoft Azure infrastructure commitment signed by the company in January 2026.
For Nvidia, this investment is a logical extension of the compute landlord thesis. As the primary supplier of the silicon required for high-frequency AI inference, Nvidia has a vested interest in ensuring that the search layer – a massive consumer of compute – remains robust and aligned with its hardware ecosystem. By deepening its stake in Perplexity, which is already an existing portfolio company following Nvidia’s lead in the July 2025 Series E extension, the chipmaker is insulating its future revenue streams against potential shifts in search architecture. The strategic importance of this positioning is amplified by Perplexity’s distribution efforts, including its integration with Samsung’s Bixby across approximately 800 million devices. Furthermore, the company’s operational focus remains fixed on a 2028 IPO target, as confirmed by CEO Aravind Srinivas and CBO Dmitry Shevelenko. This timeline provides a clear horizon for Nvidia to realize returns on its capital, provided the company can navigate the evolving legal landscape surrounding AI agents, such as the recent 9th Circuit ruling in Amazon v. Perplexity AI, which clarified the status of AI agents as tools rather than persons under the CFAA.
Nvidia’s broader portfolio – which includes significant exposure to OpenAI, Anthropic, xAI, Poolside, and SSI – demonstrates a systematic approach to market dominance. By acting as both the supplier of the underlying infrastructure and a key investor in the application layer, Nvidia is creating a self-reinforcing cycle of demand. These reported discussions underscore a deliberate strategy of vertical integration, where Nvidia secures control over both the hardware supply chain and the downstream software ecosystems that dictate compute consumption.