Full Breakdown
Palo Alto Networks CEO Urges Massive Cut in AI Token Costs
7/10/2026, 6:42:36 PM
Arora Calls for Drastic Token Cost Reductions
Palo Alto Networks chief executive Nikesh Arora told CNBC that token pricing must fall dramatically to enable large-scale artificial-intelligence deployment in enterprises. He said the current cost structure “makes AI tools increasingly difficult for businesses to implement” and warned that without a steep decline, many firms will be unable to adopt advanced models.
Context: Token Pricing as an Enterprise Barrier
Rising token expenses have emerged as a major pain point for corporate AI budgets. The concern follows OpenAI CEO Sam Altman’s recent claim that the company’s newest model is 54 % more token-efficient for agentic coding, a benchmark Arora described as “a good start.” Industry observers view token efficiency as a key lever for controlling operational spend on generative-AI services.
Official Statements from Palo Alto Networks Leadership
Arora emphasized that token efficiency should be reduced to roughly 20 % of today’s levels within the next twelve months and to 90 % of current costs by the following year. He framed the pricing issue as a strategic obstacle, stating that “we need to see the pricing for AI come down” to broaden adoption across the corporate sector. The CEO also suggested that further adjustments may be required, noting “I think we probably need another turn at it.”
Implications for AI Adoption in Business
If token costs are trimmed as Arora proposes, enterprises could allocate more of their AI budgets to development and deployment rather than consumption fees, potentially accelerating the integration of generative-AI tools into products and services. Conversely, sustained high pricing may limit AI projects to well-funded tech giants, widening the gap between large and mid-size firms.
Verbatim Quotes
- “I think 54% is a good start,” — Nikesh Arora, CEO, Palo Alto Networks
- “I think we probably need another turn at it.” — Nikesh Arora, CEO, Palo Alto Networks
- “We need to see the pricing for AI come down,” — Nikesh Arora, CEO, Palo Alto Networks
- “ Arora said token efficiency needs to drop to as much as 20% over the next 12 months, and 90% by the following year.” — Nikesh Arora, CEO, Palo Alto Networks
