The US Department of Defense's Chief Digital and Artificial Intelligence Office (CDAO) awarded contracts worth up to $200 million each to four AI vendors - Anthropic, Google, OpenAI, and xAI - according to a Forrester blog analysis. The multi-vendor structure lets the DoD draw on different providers' capabilities rather than locking into a single supplier, a portfolio approach Forrester compared to the Joint Warfighting Cloud Capability program.
The contracts signal how large government buyers are building flexibility into their AI sourcing strategies from the start. Instead of betting on one company's platform, the DoD can match specific workloads to specific vendors, a model that raises practical questions about interoperability, cost attribution, and data governance across providers.
The UK parallel and data-use terms
Forrester's analysis also referenced a separate agreement between the UK government and Google Cloud. The UK government confirmed terms that prohibit Google from training its AI models on government data. Forrester presented this as one example of the kind of data-use language emerging in large government technology contracts. The DoD contracts and the UK deal are separate agreements with separate terms and no stated connection.
Procurement teams watching these deals should note that data-handling provisions are becoming a central negotiating point, not an afterthought. The UK's explicit prohibition on model training reflects a growing insistence that government data stay walled off from commercial AI development pipelines.
Marketplace upgrades and procurement activity
On August 19, 2026, government procurement marketplace company Glass launched AI-powered upgrades to its G-Commerce marketplace. The update includes AI-driven product search and discovery, redesigned administrator dashboards and analytics, updated navigation, order tracking, a product-review and quote-request feature, and an expanded supplier network. Glass said it has processed more than $15.2 million in government purchases since its 2020 launch and reported more than 125 public agency customers and 60,000 government users.
Glass cited a $40 billion annual market in federal Government Purchase Card purchases and pointed to public schools as a large potential customer base, given their combined technology and supply spending. The company's AI search and discovery features aim at the kind of fragmented purchasing workflows that slow down procurement staff across agencies.
Separately, gov tech investment activity totaled $2.8 billion in the second quarter of 2026, up 45% from $1.9 billion in the first quarter but roughly half of the $6.1 billion recorded in the second quarter of 2025, according to market research reported by Government Technology. Deal tracker Jeff Cook said the quarter's volume was carried by a small number of large transactions and that many deals this year involve smaller, faster-growing companies with AI-focused products - a pattern distinct from the larger private-equity-backed roll-ups common a year earlier.
What connects these developments
The DoD contracts and the Glass product update are separate stories involving different organizations. But both show how AI is entering government purchasing at different levels: one at the scale of enterprise contracting for AI platforms themselves, the other at the level of a purchasing marketplace adding AI features to search, sourcing, and compliance-related workflows. The investment data adds a third dimension - smaller AI-focused gov tech firms attracting capital even as total deal volume drops from 2025 peaks.
MarketScale analysis notes that agencies and vendors should treat these as separate data points rather than evidence of one unified trend. Contract structures like the DoD's multi-vendor awards raise practical questions about how logging, data-retention rules, and cost attribution are standardized across providers when several vendors operate side by side. Procurement teams evaluating marketplace tools like Glass's update should look for verifiable detail on how AI-assisted search or recommendation features integrate with existing compliance and audit requirements, rather than assuming integration is automatic.
Why this matters for government professionals
For contracting officers and procurement leads, the DoD's multi-vendor model is worth studying - not as a template to copy, but as a signal that AI sourcing strategies are shifting away from single-vendor lock-in. The operational burden of managing multiple AI providers falls on the teams that handle logging, cost tracking, and compliance verification. Those teams need clear standards before contracts are signed, not after. On the marketplace side, AI-powered search and sourcing tools can speed up purchasing workflows, but only if the audit trail and compliance checks are built into the tooling from day one. The question to ask vendors is not "does it use AI" but "show me the log."
For professionals looking to build these evaluation skills, resources like AI for Government and the AI for Procurement Specialists learning path offer structured guidance on assessing AI tools in public-sector contexts.
Your membership also unlocks:
