AI Pricing Crisis: CEOs Demand 90% Cost Cuts by 2028 as LLM Economics Unravel

Summary

Palo Alto Networks CEO Nikesh Arora publicly called for dramatic reductions in AI large language model (LLM) pricing, demanding costs drop 20% by 2027 and 90% by 2028 for enterprise adoption to be economically viable. His plea reflects a critical disconnect between what AI companies charge for LLMs and the actual economic value delivered—decades of automation promises have largely failed, and LLMs instead function primarily as tools to intensify work from existing employees while keeping payroll stagnant. Industry critic Ed Zitron independently corroborated this analysis on the same CNBC panel, arguing the AI LLM market is fundamentally oversized: industry valuations assume a $1 trillion addressable market while the realistic addressable market for LLMs is merely $10–30 billion. Together, Arora and Zitron's comments signal an industry-wide reckoning: the AI sector's pricing model is deeply disconnected from economic reality, forcing even the biggest enterprises to demand structural correction.

• Major enterprise customers are now publicly rebelling against unsustainable AI costs, forcing industry reckoning

Editorial Opinion

The AI industry faces an inevitable reckoning. Billions have poured into companies promising labor automation and revolutionary enterprise economics—yet here we are with major corporate buyers pleading for 90% price cuts just to justify the cost. When even a CEO sympathetic to AI adoption has to publicly beg for lower prices, it's clear the business model is broken. AI hasn't delivered automation; it's extracted rents from enterprises desperate not to appear technologically backward.