AI Industry Faces Potential Shift Toward Cheaper, Smaller Models as Costs Rise

The artificial intelligence industry's foundational assumption that larger models are superior is being tested by mounting operational costs that are forcing companies to reconsider their model choices. Recent tests, including one by legal AI tool Harvey, demonstrate that smaller models can deliver comparable quality at significantly lower costs—Harvey achieved a 3x reduction in inference costs without sacrificing quality by strategically combining Claude Opus with smaller alternatives. Industry figures like Coinbase co-founder Brian Armstrong predict that within 12-18 months, 80% of AI workloads will shift to models that are 99% cheaper, while only 20% requiring maximum intelligence will use cutting-edge models. This represents a fundamental departure from the scaling-first approach that has dominated the field, where investor subsidies eliminated cost pressures and made the most advanced option the default choice. The real divide emerging is not between proprietary and open-source models, but between large and small models regardless of their origin. If this prediction materializes, it could substantially impact the financial projections of major AI companies heading toward IPOs.