Governments Are Buying AI Capability, Not Just Writing Rules for It

Greece tests agentic AI on city services

The Ministry of Digital Governance and Artificial Intelligence of Greece signed a Memorandum of Understanding with Presight focused on investigating agentic AI for municipal waste management and water-network monitoring, with the collaboration also aiming to enhance safety on public transport networks. G42 and four group companies signed additional MOUs with Greek partners covering AI infrastructure, government services, life sciences and smart cities, and the initiatives are subject to definitive agreements and regulatory approvals, which means these are explorations, not signed deployments. m:…2533

The structure matters more than any single use case: a national ministry is procuring investigation into agentic systems for operational services, alongside infrastructure and services agreements from a Gulf AI group. For vendors, the buyer is a government testing whether agents can run city infrastructure, and the gating items are definitive agreements and approvals rather than model capability alone. m:…2533

Qatar builds the workforce alongside the office

Scale AI signed an agreement with United Development Company to establish a long-term office at The Pearl in Doha, covering nearly 2,000 square meters, building on a partnership with the State of Qatar and its communications and IT ministry originally announced at Web Qatar 2025. The Doha team has grown to nearly 100 employees, about half of them technical engineers, and Scale provides data and evaluation benchmarks for AI models in Qatar across healthcare, justice, education, tourism and government sectors. m:…7660

The workforce number is the more durable signal: Scale has trained nearly 6,000 government employees through the ALIF program, delivered through the Qatar Digital Academy. A vendor that trains the buyer workforce is selling institutional capability, not just a tool, and Scale reports similar partnerships with government AI evaluation bodies in the United States, the United Kingdom, Korea and Singapore, which suggests evaluation itself is becoming a government procurement category. m:…7660

Singapore and Washington set the health terms

Google is partnering with the Singapore Ministry of Health to deploy frontier AI across the healthcare system, establishing a new research hub focused on foundational AI models, while Google.org and Google DeepMind contribute 7 million dollars to the Philanthropy Asia Alliance Health for Human Potential Coalition, including 1 million dollars for PathGen, an AI platform for disease detection. Google DeepMind is working with the National Health System to evaluate an AI co-clinician tool, Google Cloud is partnering with NHG Health on a population health program using AI-driven predictive insights, and Google Research is bringing open models such as MedGemma to frontline care settings. m:…6015

In the United States, a serious new competitor to current dominant electronic health record companies is expected within one to two years. The ONC is scaling back its certification program in the HTI-6 rule to lower barriers for new participants, proposing to remove AI model card requirements that were not fit for purpose and rarely used, and working with the OIG, FTC and DOJ to enforce information blocking rules, while noting AI applications increasingly migrate outside traditional EHRs to avoid certification requirements. TEFCA, the national exchange framework, has reached approximately 1.7 billion transactions at 200 million per month, which gives the interoperability bet a measurable base. The ONC also aims to ensure AI-generated data can be incorporated into traditional EHRs and that information exchange runs two ways. m:…6574