AI is developing "model-year" pricing, where older capabilities become cheaper while frontier models stay expensive due to higher compute demands, according to Michael Watkins. Commoditized AI cannot create durable competitive advantage, so executives must distinguish between capabilities available to everyone and those that differentiate the business. Speaking at BlackRock’s US Infrastructure Summit in March 2026, OpenAI CEO Sam Altman described a future where "intelligence is a utility, like electricity or water, and people buy it from us on a meter." However, the bandwidth comparison shows corporate spending did not collapse as prices fell; volume consumed grew faster than price fell. When OpenAI first sold access to GPT-3 in late 2021, processing a million tokens of text – roughly 750,000 words – cost $60. Three years later, models matching GPT-3 performance cost six cents for the same volume – a thousand-fold reduction. Yet the bill for the most capable current model remained high. Uber’s CTO revealed the company exhausted its 2026 AI budget in just four months after 5,000 engineers adopted an agentic coding tool costing $500-$2,000 per user per month. Epoch AI tracked six benchmarks and found the price to reach a fixed performance level fell between nine-fold and 900-fold per year depending on the task. The price of GPT-4-level performance on PhD-level science questions fell roughly 40-fold per year.