Urgent.News

What's breaking now, across thousands of outlets.

AI

The Collapsing Cost of Running AI

Alexander Kopylkov on why AI got dramatically cheaper to run this year, and why that made record spending and smarter investing happen at the same time.

The Collapsing Cost of Running AI

The price of running AI has fallen significantly in recent years, while the total amount spent on AI infrastructure has reached a record high. This disparity has changed the way investors view AI companies. Each time a user interacts with an AI-powered app, there is a real cost involved, known as inference. Over the past two years, the cost of inference has dropped sharply.

Despite the lower per-question price, total spending on AI has surged. Major players like Amazon, Microsoft, Alphabet, and Meta have all increased their AI-related spending substantially. The combination of falling AI costs and record spending has reshaped the landscape for AI investments.

Alexander Kopylkov, a seasoned investor in technology, initially found this situation confusing. However, he realized that this is a familiar pattern in economics, similar to how improvements in fuel efficiency did not lead to decreased coal consumption. Instead, cheaper and more efficient power created new uses for coal that were not viable before. Similarly, cheaper AI has led to increased usage, turning it into a commodity that is in high demand.

As AI becomes more affordable and accessible, companies are leveraging this technology in new ways. Rather than just asking one question, businesses are now firing off thousands of follow-up questions automatically. This rapid increase in usage is outpacing the decline in per-question costs, leaving investors with the bulk of the bill. The era of having an edge due to access to a powerful AI model is fading, as the same models are now available to companies of all sizes at the same price.

In 2026, the companies that investors are counting on for capital are those with unique assets that cannot be easily replicated by rivals. These assets include proprietary customer data, complex workflows that would take months to recreate, and relationships built over years rather than instant sign-ups. Companies that possess these valuable elements are better positioned to succeed in the AI-driven market.

The high spending on AI infrastructure is a bet on the rapid growth of AI usage, and those who can effectively utilize this technology stand to benefit, while those relying solely on the underlying tool may struggle to create sustainable businesses.

Written by urgent.news from HackerNoon's reporting — not their text. Machine-written — may contain errors; check the original before relying on it.

Read the original at hackernoon.com →

More in AI

More from Monday 24 August →