AI industry confronts power‑stack race and looming token‑price surge

Analysts warn that the competition to dominate artificial‑intelligence development increasingly hinges on the ability to assemble a complete "AI Power Stack"—reliable electricity, large‑scale compute, domestically produced chips and a broad application ecosystem. The United States is promoting export bundles that combine hardware, models, software and standards, while China pursues a longer‑term platform strategy centred on its own semiconductor and model initiatives.

At the same time, the seemingly cheap AI‑as‑a‑service pricing model is unsustainable. OpenAI and Anthropic posted multi‑billion‑dollar operating losses in 2025, and their token prices are subsidised by investor capital. As investors demand returns, token costs could rise sharply, risking a costly lock‑in for midsized firms that have built their operations around a single provider. Experts suggest a hybrid architecture that incorporates open‑source models to shield businesses from future price shocks.

The convergence of strategic technology rivalry and volatile pricing underscores the urgency for firms and governments to consider the entire AI ecosystem rather than isolated model performance.