GAO Warns AI Pricing and Cyber Risks Could Slow Federal Adoption
The Government Accountability Office warns that the end of promotional AI pricing could increase agencies’ costs and slow adoption. Contractors should be prepared to demonstrate measurable outcomes, control usage costs, support human oversight, provide cybersecurity documentation, and avoid locking agencies into a single model or vendor.
Key Signals
- GAO warns that expiring promotional AI pricing could increase federal agencies’ operating costs
- Compare consumption-based AI pricing with performance- or outcome-based contract structures
- Require contractors to connect AI spending to measurable mission results and cost controls
- Assess human oversight, cybersecurity documentation, and vendor or model flexibility in deployment reviews
- No specific solicitation, contract value, or compliance deadline is identified
"If your contracts themselves are structured merely on consumption, guess what? You’re guaranteed to consume more. If you’re working with AI companies and setting firm fixed contracts that are actually performance and outcome-based, forget how many tokens are consumed, are you getting the problems you have solved, are you reaching the goals that you have and doing it in such a way that’s agnostic and not getting tied into any particular model or vendor?"
The Government Accountability Office (GAO) has warned that federal agencies could face higher costs as promotional pricing for artificial intelligence tools ends. The pricing shift may affect both agencies already deploying AI and those still assessing whether to adopt it. For acquisition and program teams, introductory rates may not provide a reliable basis for forecasting long-term operating expenses, particularly when costs vary with use. The warning puts total cost of ownership, pricing transparency, and the structure of vendor agreements in sharper focus for federal AI procurements.
Agencies evaluating AI offerings may need to look beyond launch discounts and compare consumption-based pricing with performance- or outcome-based terms. A contract that charges primarily for usage can create incentives to consume more without establishing whether the tool is solving a mission problem or delivering the intended result. Conversely, fixed or outcome-oriented arrangements may help agencies connect spending to defined objectives, though acquisition teams still need to specify how results will be measured and how performance will be validated. The signal identifies no specific solicitation, contract value, or compliance deadline, so it should be read as a strategic procurement consideration rather than an announced opportunity.
The commentary also highlights vendor flexibility and oversight as important elements of deployment reviews. Justin Herman, Federal Civilian Director at Cognition, urged agencies and contractors to focus on results rather than token consumption, asking: “If your contracts themselves are structured merely on consumption, guess what? You’re guaranteed to consume more.” He contrasted consumption-based agreements with “firm fixed contracts that are actually performance and outcome-based,” emphasizing whether agencies are solving problems and reaching goals in a manner that is “agnostic and not getting tied into any particular model or vendor.” The procurement implication is that contractors should explain how their solutions can adapt as models, prices, and agency needs change, rather than relying on a single model or provider.
Cybersecurity documentation, human oversight, and agency AI inventories are also relevant to federal deployment decisions. Contractors should be ready to describe security controls and provide documentation that supports agency review, while explaining where people remain involved in managing or validating AI outputs. Agencies may also need current inventories to understand what systems are in use and how they are deployed. The source names the National Institute of Standards and Technology and the Office of the National Cyber Director among the relevant agencies, alongside GAO; it does not specify new requirements or attribute a particular deadline to them. For GovCon teams, the practical response is to prepare clear, evidence-based materials that address cost, performance, security, oversight, and portability together.
- GAO warns that the end of promotional pricing for AI tools could raise federal agency usage costs and slow adoption.
- Acquisition teams should assess costs beyond introductory rates and model expected spending under realistic usage scenarios.
- Compare consumption-based pricing with performance- or outcome-based terms tied to defined mission results.
- Contractors should substantiate value with measurable outcomes, cost controls, and clear explanations of how results will be assessed.
- Justin Herman of Cognition cautions that consumption-only contracts can incentivize greater usage without ensuring mission problems are solved.
- Explain how solutions support human oversight and reduce unnecessary dependence on a single model or vendor.
- Prepare cybersecurity documentation and information that can support agency deployment reviews and AI inventories.
- The signal names GAO, the National Institute of Standards and Technology, and the Office of the National Cyber Director; it identifies no specific solicitation, award amount, or compliance deadline.
Agencies
- Government Accountability Office
- National Institute of Standards and Technology
- Office of the National Cyber Director
Vendors
- Cognition
Sources
- GAO's Take on AI Costs, Cyber Risks in Federal Tech Strategies | GovCIO Media & ResearchGovCIO Media & Research · Nov 10