Palo Alto’s CEO says AI prices must fall 90% before the tech really takes off

What happened

Palo Alto Networks CEO Nikesh Arora said AI model usage costs need to drop by up to 90% before businesses can adopt AI widely. He voiced this opinion during an interview with CNBC, responding to OpenAI’s recent announcement that its GPT-5.6 model is 54% cheaper to run compared to earlier versions. Arora’s comment underscores that even with incremental efficiency gains, current AI token pricing remains a major barrier to scalable enterprise deployment.

Why it matters

AI adoption at scale is cost-sensitive. Most businesses won’t justify integrating expensive AI models into products or workflows until running the models becomes dramatically cheaper. Token pricing reflects the cost of compute per interaction, and while OpenAI’s new model reduces this cost, Arora’s estimate shows a much steeper price drop is still needed. This price barrier affects everything from customer-facing AI tools to internal automation and data processing, slowing AI’s practical rollout despite technical advances.

High model costs also shape vendor economics. Cloud providers and AI service vendors face pressure to innovate on infrastructure and pricing models to unlock mass adoption. Meanwhile, developers and founders aiming to build AI-powered tools must factor in these pricing dynamics when designing scalable solutions or pitching to investors focused on unit economics.

What to watch next

Keep an eye on AI infrastructure developments aimed at slashing compute costs, like hardware acceleration and model optimization. Also watch for pricing moves from major AI model providers, including OpenAI and competitors, to see if they aggressively cut token prices to speed adoption.

Investors and businesses should track partnerships between AI companies and cloud providers that might bundle cheaper AI access. Finally, monitor if any startups focus on disruptive, low-cost AI model deployment platforms designed specifically to tackle this pricing challenge for enterprises.

AI Quick Briefs Editorial Desk