Perplexity AI's race to compete with Google just got a lot more expensive. The company has committed $750 million over three years to Microsoft Azure for the computing infrastructure needed to run its AI-powered search engine at scale, according to recent reports. This massive investment reveals a hidden truth about the AI search boom: building a better answer engine is only half the battle. The other half is paying for the 24/7 computational power to run it.

Perplexity, founded by Aravind Srinivas, has grown into a $20 billion company by offering something fundamentally different from traditional search. Instead of returning a list of links, Perplexity gathers information from multiple sources across the web, synthesizes it, and delivers a direct answer with citations. This approach has resonated with users tired of clicking through blue links, but it comes with a steep price tag that most people never see.

Why Does AI Search Cost So Much More Than Traditional Search?

Every search query on Perplexity requires multiple computational steps that traditional search engines avoid. The system must retrieve relevant information, run language models to synthesize that information, rank results, and generate a natural-language answer, all in real time. Each of these steps demands significant computing power, and the costs scale directly with usage. As more users ask questions, the infrastructure bill grows proportionally.

By contrast, Google's traditional search model relies on pre-indexed links and simpler ranking algorithms that are far less computationally expensive to run. Perplexity's approach is more helpful to users but dramatically more costly to operate. This is why the company needed to secure such a substantial infrastructure partnership with Microsoft Azure.

How Is Perplexity Building a Sustainable Business Model?

With infrastructure costs eating into margins, Perplexity has deliberately chosen not to rely on advertising as its primary revenue source. Instead, the company is pursuing three main strategies to generate revenue and offset its massive computational expenses:

• Subscription Services: Perplexity Pro offers subscribers access to better AI models, advanced research capabilities, and additional productivity tools designed to help users work more efficiently.
• Enterprise Solutions: Many organizations have information scattered across multiple systems and departments, forcing employees to waste time searching for data that already exists internally. Perplexity's business tools act as a knowledge layer for organizations, helping retrieve information, analyze documents, and streamline research workflows.
• AI-Powered Browser: Perplexity launched Comet, its own AI-focused browser, to gain greater control over how users research and interact with AI tools. This mirrors strategies used by Google with Chrome and Microsoft with Windows.

The enterprise opportunity may ultimately prove larger than consumer search. Business customers typically generate more revenue than individual users, which is why many AI companies are expanding into enterprise products and services. Perplexity even attempted to acquire Google's Chrome browser for $34.5 billion in 2025, though the deal never materialized due to regulatory and competitive concerns.

What's Next for Perplexity Beyond Search?

Perplexity is preparing for the next evolution of AI, moving beyond simply providing answers toward taking action. Features such as Computer, Comet, and advanced research tools show that the company wants AI to play a larger role in helping users complete work rather than simply providing information. This shift from search to AI agents represents a much larger business opportunity, since businesses are often willing to pay more for automation than they are for information access.

"Google remains one of the best search engines in the world," said Aravind Srinivas, founder of Perplexity AI.

Aravind Srinivas, Founder of Perplexity AI

Even Srinivas has acknowledged Google's dominance in search. However, Perplexity does not necessarily need to outperform Google to succeed. If the company can capture a meaningful share of AI search, enterprise intelligence, and agent-driven products, the $750 million infrastructure investment could pay for itself many times over.

The broader lesson here extends beyond Perplexity. The AI search race is not just about creating better systems; it is fundamentally about being able to afford the infrastructure to run them at scale. Competing with Google is extremely difficult because the company benefits from massive infrastructure, global distribution, strong brand recognition, and one of the largest advertising businesses in the world. Perplexity's advantage is that it was built specifically for the AI era, allowing it to move faster and experiment more freely without protecting a legacy search business.

As the AI search market matures, infrastructure costs will likely become the defining competitive advantage. Companies that can secure reliable, affordable computing power will survive. Those that cannot will struggle to scale, no matter how good their AI models are.