The battle for artificial intelligence supremacy is entering a new phase, and according to Perplexity CEO Aravind Srinivas, the ultimate winner may not be determined by who builds the most powerful AI model. Instead, success will come down to a much more practical metric: how efficiently companies can convert computing power and energy into real economic value for users.

Speaking in an interview with CNBC on Wednesday, Srinivas argued that the companies capable of generating the highest value from the energy their AI systems consume will eventually command the strongest market positions and highest valuations.

The Metric That Could Define the AI Industry

Srinivas believes the future of AI will be shaped by a balance between intelligence, speed, cost efficiency, privacy, and accuracy. Rather than focusing solely on raw model performance, he emphasized the importance of maximizing what he described as the economic value generated per unit of energy consumed.

In simple terms, AI companies that can provide better results while using fewer computing resources will gain a significant competitive advantage. As AI adoption expands across industries, the cost of running increasingly sophisticated models has become one of the sector’s biggest challenges.

Large language models require enormous amounts of computing power, advanced graphics processors, and substantial electricity consumption. As a result, energy efficiency is rapidly becoming just as important as innovation.

Why Energy Matters in Artificial Intelligence

Every interaction with an AI system involves processing data through units known as tokens. When users submit a question or request, AI models break the task into thousands of smaller data components that must be analyzed and processed.

The more complex the task, the more tokens are generated and the more computing power is required. This directly impacts operational costs because every token consumes processing resources and electricity.

Srinivas suggested that the strongest long-term AI businesses will be those that deliver the greatest economic output while minimizing these costs. While some AI providers currently generate substantial revenue through premium-priced models, he noted that high prices alone do not guarantee lasting leadership.

As competition intensifies, efficiency is expected to become a critical differentiator.

The Shift Toward Agentic AI

Perplexity is increasingly focusing on one of the fastest-growing segments of artificial intelligence: agentic AI.

Unlike traditional AI chatbots that primarily answer questions, agentic AI systems are designed to complete multi-step tasks, make decisions, and execute workflows with limited human intervention.

The company introduced Perplexity Computer earlier this year, describing it as an AI agent capable of handling complex assignments over extended periods. These systems are intended to move beyond simple conversations and function more like digital assistants that can actively perform work on behalf of users.

Industry analysts view AI agents as one of the next major growth opportunities in the sector, with businesses seeking tools that can automate research, scheduling, document creation, customer support, and other operational tasks.

Perplexity Expands Across Microsoft and Apple Ecosystems

In a major product expansion, Perplexity announced that its Personal Computer platform will now be available on Microsoft's Windows operating system.

The move significantly broadens the company’s reach, allowing users to connect the AI assistant directly with popular workplace applications such as Microsoft Word and Outlook. The system can also access and interact with files stored on a user's device, enabling more advanced productivity features.

The Windows rollout follows Perplexity’s earlier launch on Apple’s Mac platform, giving the company access to two of the world's largest desktop operating system ecosystems.

The expansion reflects a broader industry trend in which AI providers are embedding assistants directly into everyday workflows rather than limiting them to standalone chatbot experiences.

Competition in the AI Race Continues to Intensify

Perplexity faces formidable competition from some of the most influential technology companies in the world.

Major players including OpenAI, Anthropic, and Google are aggressively investing in AI agents, advanced reasoning models, and enterprise-focused solutions. Billions of dollars continue to flow into AI infrastructure, cloud computing capacity, and next-generation model development.

The competitive landscape has evolved rapidly over the past two years, with companies racing to secure users, enterprise contracts, and strategic partnerships. As a result, innovation cycles are becoming shorter and expectations for AI performance continue to rise.

Many industry experts believe that the next phase of competition will be less about launching bigger models and more about creating AI systems that are affordable, scalable, and capable of delivering measurable business outcomes.

Valuations Highlight Investor Confidence

Investor enthusiasm for artificial intelligence remains exceptionally strong.

Perplexity was most recently reported to be valued at approximately $20 billion, reflecting its rapid growth and increasing influence within the AI ecosystem.

However, the company remains considerably smaller than some of its largest rivals. OpenAI's valuation has surged beyond $850 billion, while Anthropic's valuation has approached the $1 trillion mark as demand for AI-related investments continues to accelerate.

Anthropic recently confidentially filed for a U.S. initial public offering, underscoring the strong appetite among investors for companies positioned at the center of the AI revolution.

The continued rise in valuations highlights how investors increasingly view artificial intelligence as one of the most transformative technology opportunities of the decade.

The Long-Term Winner May Not Be the Biggest

Srinivas’ comments reflect a growing belief across the technology sector that sustainable success in AI will require more than powerful models and rapid revenue growth.

As computing costs rise and demand for AI services expands globally, companies will face increasing pressure to deliver greater value with fewer resources. Those that can balance intelligence, speed, privacy, affordability, and energy efficiency may ultimately emerge as the dominant players.

In an industry often focused on breakthrough capabilities and headline-grabbing valuations, Perplexity’s CEO argues that the most important measure of success may be surprisingly simple: generating the highest economic value from every watt of power consumed.