Share with your CIO
The European Commission is committing €1.3 billion to AI procurement between 2025 and 2027, covering generative AI applications in health, digital identity infrastructure, and public service modernization. Amber Sinha’s analysis at Tech Policy Press argues the Commission is repeating a well-documented failure mode: treating AI as a policy goal rather than a tool, which hands agenda-setting power to a small cluster of vendors and retrofits public needs around commercial product roadmaps instead of the other way around.
What this means for your business
If your organization sells into government or depends on public digital infrastructure, the vendor concentration Sinha describes is the structural condition you’re already operating inside. A handful of identity and AI vendors, including Thales, IDEMIA, and MOSIP, effectively define what “digital government” looks like because procurement processes are written to fit their capabilities. Enterprise CIOs who rely on that infrastructure inherit its constraints, including surveillance-enabling architectures and exclusionary design choices that were never actually decided by democratic deliberation.
The deeper problem Sinha identifies is what might be called agenda capture: the moment when a vendor’s product vision becomes the government’s policy vision, not because a minister chose it, but because no minister was competent or incentivized to choose otherwise. This isn’t unique to government. Enterprise AI procurement follows the same gravity. When internal teams lack the technical fluency to write a meaningful scope of work, vendors fill the vacuum with their own. The RFP becomes a lightly edited version of the vendor’s sales deck, and the organization ends up with a solution to a problem the vendor defined.
The EU’s AI-first spending push gives large platform vendors, including the hyperscalers already embedded in European public sector contracts, enormous leverage to position themselves as the default infrastructure layer before national governments have done the needs assessment to know what they actually require. Any enterprise CIO watching this dynamic should treat it as a leading indicator: when a government anchor customer moves fast on AI spend without rigorous procurement discipline, it compresses the evaluation window for everyone downstream and normalizes vendor-led scoping as the industry standard. The budget cycle you’re defending right now is where that discipline either holds or doesn’t.
Based on reporting from AI Procurement and the Capture of Public Purpose, originally published 2025-11-17 03:00:00.
