Descriptions:
Perplexity CEO Aravind Srinivas sits down with Bloomberg Technology to discuss the company’s newly announced chip-agnostic hybrid compute orchestration strategy. Demonstrated alongside Intel at a recent event, the system dynamically routes AI workloads between local on-device processing and cloud servers based on task complexity, cost, privacy requirements, and accuracy needs. Srinivas confirmed the solution works with both Intel and NVIDIA RTX hardware, extending Perplexity’s established model-agnostic philosophy into the chip layer.
Srinivas also shared several notable financial disclosures: Perplexity’s revenue tripled in the first five months of 2026, with the company crossing $500 million ARR around mid-April. Subscriber mix has shifted sharply toward power users, with the $200/month Max plan growing from roughly 9% to 30% of subscriptions—signaling strong demand for the platform’s deep research and orchestration capabilities among users willing to spend $2,000 annually.
On competitive threats from Anthropic, OpenAI, and xAI, Srinivas expressed confidence rather than concern: every improvement from any frontier lab strengthens Perplexity’s routing layer, which aggregates the best available models into a single unified interface. He also clarified the company’s engagement philosophy—Perplexity optimizes for user retention across multiple research sessions rather than maximizing time-on-platform per query, a deliberate departure from traditional engagement-driven product design.
📺 Source: Bloomberg Technology · Published June 02, 2026
🏷️ Format: Interview
