Artificial intelligence is advancing at an incredible pace, but one of the industry's biggest leaders believes the technology is still far too expensive for widespread adoption. Palo Alto Networks CEO Nikesh Arora recently argued that AI costs need to fall dramatically—by as much as 90%—before businesses can fully unlock its potential.
According to Arora, while companies are eager to integrate AI into their operations, the high cost of computing power, infrastructure, and large language models remains a significant barrier. Many organizations are experimenting with AI, but scaling those deployments across an entire business can quickly become cost-prohibitive.
Arora believes the industry will eventually reach a point where AI becomes as affordable and accessible as cloud computing is today. Lower hardware costs, more efficient AI models, and increased competition among chipmakers and cloud providers are expected to help drive prices down over the coming years.
His comments come as technology giants continue investing billions of dollars into AI infrastructure, including new data centers, custom chips, and next-generation AI services. While these investments are accelerating innovation, they have also contributed to soaring operational costs that are often passed on to enterprise customers.
For businesses considering AI adoption, Arora's message is clear: the technology has enormous potential, but reaching mass adoption will require a significant reduction in costs. As AI becomes more efficient and less expensive, organizations of all sizes may finally be able to leverage advanced automation, analytics, and intelligent assistants without breaking their budgets.