Nikesh Arora, CEO of Palo Alto Networks, emphasized that current AI token pricing is a significant barrier for businesses looking to integrate artificial intelligence widely. He forecasts that costs must fall by up to 90% within two years to support broader adoption.
- AI token costs must decrease up to 90% within two years for widespread business use.
- High token expenses strain AI budgets and hinder enterprise adoption.
- Cheaper alternatives and open-weight models gain interest amid rising costs.
Market signal
Palo Alto Networks CEO Nikesh Arora highlighted escalating token costs as a critical barrier for scaling AI adoption across businesses. He noted that although improvements, such as OpenAI's 54% increase in token efficiency, are promising, the current pricing model still requires substantial improvement. Arora expects token prices to drop to 20% of current levels within the next year and down to 10% the year after.
This bench-marking signals a broader market expectation for AI price reforms, revealing enterprise sensitivity to per-token charges. Industry voices including Palantir's CEO express frustration with current token-based pricing, underscoring an urgent need for innovation in cost structure that supports mass enterprise deployment without inflating budgets.
Operator impact
High token costs are prompting many businesses to reassess their AI adoption strategies, with some choosing open-weight or alternative AI models that promise more cost-effective scaling. This shift presents operational challenges and opportunities for AI platform providers who must balance model sophistication with affordability to attract and retain enterprise clients.
Meanwhile, the rising expenditure on AI infrastructure is pushing companies to secure substantial funding, exemplified by recent multi-billion-dollar bond and debt raises from tech giants. This inflow aims to sustain growth and innovation, but operators must also navigate a market where infinite demand pressures technological and economic optimization.
What to watch next
Monitor developments in token pricing strategies and the adoption of alternative AI pricing models like open-weight frameworks, which could reshape how enterprises pay for AI services. These models may lower barriers and expand AI accessibility globally, especially as US and Chinese providers compete on cost and capability.
Additionally, observe how enterprise budgets for AI evolve alongside these pricing changes, and whether expected efficiency gains will indeed rationalize spending while supporting demand. The balance between growing AI capabilities and affordability will be a critical factor in technology providers’ competitive positioning and enterprise procurement decisions over the coming years.