Most AI projects fail — 3 priorities to improve your odds

Spending on AI continues to rise — Gartner forecasts fast growth through 2026 — but generative AI has slipped into the Trough of Disillusionment. That cooling of hype has prompted boards to question the returns from AI experiments, a concern reinforced by MIT research suggesting 95% of generative AI projects fail to deliver value.

Gartner analyst John‑David Lovelock says the trough is an opportunity to refocus investments and recommends three priorities through 2026. First, concentrate on capacity building: technology providers are buying AI‑optimized servers and building data centers, and organizations must decide whether to run their own infrastructure, work with major cloud providers, use platform operators, or call large language models via APIs.

Second, build strong partnerships. For most companies, the sensible path this year is to lean on an established partner stack rather than pursue bespoke moonshot projects.

generative ai, ai hype, mit research, project failure, capacity building, ai servers, data centers, cloud providers, llms, partnerships