Perplexity just made the most ambitious bet of the AI race so far. The company launched Computer, a multi-model agent orchestration platform that coordinates 19 different AI models to complete complex workflows entirely in the background. The product is available only to Perplexity Max subscribers at $200 per month.
This is not a feature launch. This is a thesis.
The Core Argument
For over a year, Perplexity has been quietly collecting enterprise usage data that tells a different story than what most AI companies are selling. In January 2025, more than 90% of enterprise tasks on the Perplexity platform were handled by just two models. By December 2025, no single model commanded more than 25% of usage.
The implication: models are not converging toward a single general-purpose intelligence. They are specializing. And the company that figures out how to coordinate all of them will capture more value than any individual model maker.
“Treat models as interchangeable tools rather than core products,” CEO Aravind Srinivas wrote on X. “When models specialize, they just become tools similar to the file system, CLI tools, connectors, browser, search.”
Inside the Machine
Computer uses Claude Opus 4.6 as its central reasoning engine for orchestration and coding. Gemini powers deep research queries. Nano Banana generates images and Veo 3.1 handles video. Grok handles lightweight, speed-sensitive tasks. GPT-5.2 manages long-context recall and expansive web search.
That roster is not fixed. New models get added as they demonstrate strength in specific domains. Users can also step in manually and assign subtasks to particular models if they prefer.
The product sits somewhere between OpenClaw and Claude Cowork as a general-purpose digital worker. A user can describe a desired outcome — “Plan a weeklong trip to Japan, find flights under $1,200, and build a full itinerary with restaurant reservations” — and Computer will autonomously break that project into components, assign each to the right model, and work in the background.
Why This Matters
The AI industry is grappling with a fundamental question: now that foundation models have become extraordinarily capable, who captures the value? The model makers — OpenAI, Anthropic, Google — or the companies that sit above them and turn raw intelligence into reliable products?
Perplexity is answering with a $20 billion valuation bet on the latter.
This matters for a few reasons. First, the enterprise data is compelling. The fragmentation of model usage across tasks suggests that no single model is winning on all fronts. Second, the orchestration approach addresses a real pain point: choosing the right model for the right task is becoming its own expertise. Third, it signals a maturation of the AI market away from “best single model” competitions toward “best system” arguments.
What This Means for the Industry
If Perplexity’s thesis holds, we should expect more companies to pursue multi-model strategies rather than betting entirely on a single foundation model. The value chain is shifting upward — from model capabilities to orchestration and product design.
For enterprises, this raises practical questions. Do you train your team on one model or become fluent in many? Do you build orchestration capabilities in-house or buy them?
The answer likely depends on your workflow complexity. Simple, repetitive tasks still work fine with a single capable model. But for multi-step workflows requiring different cognitive strengths, orchestration is becoming the competitive advantage.
Perplexity has made its bet. The market will decide if specialization beats generality.