
The Pentagon's Next Readiness Test Should Begin When the Cloud Goes Dark
Published Nov. 20, 2026 • Read time: 5m
The Pentagon spends enormous amounts of time determining whether a weapon works, a contractor can deliver and a network can withstand a cyberattack. But there's another question it should be asking: Can the companies supplying America's national defense still operate when a technology provider they depend on suddenly stops working?
Last week's global Microsoft outage exposed how quickly those dependencies can become liabilities. Businesses and government offices lost access to critical services, while Microsoft's preliminary findings revealed a software defect that had gone undetected for months. When increased network traffic collided with reduced server capacity, systems designed to redirect users away from trouble instead helped spread the failure. Within two hours, more than 80 percent of affected gateway servers were unresponsive.
For the Pentagon, the implications extend well beyond unanswered emails. Modern defense manufacturing increasingly relies on commercial software for everything from production planning and quality control to procurement, inventory management and communications with suppliers. When those systems become unavailable, the consequences can extend directly to production schedules and the delivery of critical equipment.
Imagine a smaller subcontractor making a critical component for a missile system. Its drawings are stored in one service, inspection records in another, purchasing approvals in a third and coordination with its prime contractor somewhere else. The company may have functioning machinery, available materials and employees ready to work, yet find itself unable to retrieve approved designs, document completed inspections or coordinate deliveries. Even its backup systems may offer little protection if accessing them requires the same outside provider.
This is the vulnerability Washington needs to address. America's defense industrial base increasingly depends on commercial technology infrastructure that individual contractors neither own nor control. Yet having multiple vendors or backup systems doesn't necessarily provide the independence needed to keep essential operations running during a major disruption.
The Pentagon's Joint Warfighting Cloud Capability program offers a useful example of both the progress Washington has made and the challenges that remain. By awarding contracts to Amazon, Google, Microsoft and Oracle, the department expanded its access to commercial cloud providers. But purchasing services from four companies doesn't automatically mean systems can move between them when one goes down. That requires deliberate planning, compatible technology and the ability to maintain operations without relying on the same underlying services.
A June report from the Government Accountability Office reinforced that concern. Eleven of 24 federal agencies reported difficulties developing multi-vendor cloud strategies. Pentagon officials specifically identified the additional expertise, tools, staffing and resources necessary to make systems operate across different providers. Those challenges are manageable, but they demonstrate why buying redundancy and achieving resilience are two very different things.
The problem becomes particularly consequential when cybersecurity compliance is mistaken for operational readiness. A contractor may satisfy extensive security requirements designed to protect sensitive information without demonstrating that it can continue manufacturing, approving shipments, or communicating with suppliers during a major technology outage. Protecting information from unauthorized access is essential, but so is ensuring authorized employees can access what they need when ordinary systems fail.
The Pentagon should make operational continuity a measurable requirement for its most consequential contracts, particularly those involving specialized components, limited suppliers or production delays that could affect larger defense programs. That means identifying which functions must remain operational, which outside providers support them and how long a disruption could continue before it affects delivery.
And rather than accepting another questionnaire or written contingency plan, procurement officials should require practical demonstrations. Contractors could be asked to maintain essential operations during a simulated 24-hour loss of their primary cloud or communications provider. Those exercises should include critical subcontractors and evaluate whether employees can access approved drawings, document inspections, coordinate shipments and maintain appropriate security protections.
The requirements should reflect the actual risks. A small machine shop shouldn't need two expensive cloud environments to manufacture a relatively ordinary part. Securely accessible production instructions, alternative communications and procedures for documenting work offline may be sufficient. Companies supporting more sensitive or time-critical programs may require independent infrastructure, separate authentication capabilities or more substantial recovery arrangements.
Congress should reinforce these expectations by ensuring that operational resilience becomes part of how federal agencies evaluate critical systems and contractors. Contracting officers need the authority and resources to assess whether continuity plans actually work, rather than relying on promises of 99.9 percent uptime that offer little reassurance when services become unavailable.
The Pentagon already recognizes the importance of operating without reliable connectivity. Its own cloud strategy addresses environments where communications are denied, disconnected, intermittent or limited, acknowledging that forces must remain effective even when their normal connections are unavailable. Applying that same principle to the companies producing the equipment those forces depend on should be an equally important priority.
Because if a future national security crisis requires the United States to accelerate weapons production, the last thing Washington should discover is that a critical supplier has the workers, machinery and materials to meet demand but cannot access the systems needed to complete the job.
Microsoft's outage offered a warning that the Pentagon should take seriously. America's defense industrial base cannot afford to become so dependent on commercial technology that an ordinary software failure can threaten its ability to deliver. And the time to determine which contractors can continue operating when the cloud goes dark is before the country needs them most.
Colton Overcash is the founder and managing principal of Vertex Strategies and a former presidential appointee at the U.S. Department of Homeland Security. He writes on national security, foreign policy, critical infrastructure and emerging technology, and advises companies and organizations on government strategy and public-sector markets.

The Pressure Is the Message: What Xi and Putin Faced in Shenzhen
By Emzari Gelashvili • RealClearDefense
China’s Revised Mobilization Law and the Expanding Reach of National Defense
By Muhammad Shoaib • RealClearDefense
U.S. Putting Weapons in Space: A Strategic Blunder?
By Ajey Lele • RealClearDefense
The Clearance Myth Is a Readiness Problem
By Elijah Fernandez • RealClearDefense
The Pentagon's Next Readiness Test Should Begin When the Cloud Goes Dark
By Colton Overcash • RealClearDefense

U.S. Putting Weapons in Space: A Strategic Blunder?
By Ajey Lele • RealClearDefense
U.S. Senate Seeks to Exchange Shipbuilding Funds for More Munitions
By Naval News
F-35C Completes First Launch and Recovery Trials on a Ford-Class Carrier
By Navy T.
USS George Washington Leaves Mideast As Trump Mulls Ending Iran War
By S&S
ULA’s Latest Launch Failure Spells Trouble for Space Force
By Greg Hadley, A&SF

