Tldr: The growing interest in AI from governments is a welcome development, but we believe that excitement should be tempered with discipline. We look at how government technology investment often fails to accomplish its goals. This usually stems from a lack of clear rationale for government action at all, inadequate funding, and the inherent limitations of top-down approaches to technological development. As a result, we see a combination of small, low value grants at one end and grandiose “grands projets” on the other. We see this pattern across EU-wide efforts and increasingly on a smaller scale in the UK. We propose tests that any serious government investment in technology should pass and provide two examples that meet the bar.