Is the government losing $500M a year in cloud savings?

BlackJack3D/Getty Images

COMMENTARY | Multiyear discounts enjoyed by commercial cloud customers aren't reaching the government because of federal spending laws.

A quirk in government fiscal law may be costing the government as much as $500 million a year in savings for cloud computing. 

Cloud service providers typically offer significant discounts for contractual commitments of three years or more. The cloud providers are willing to discount for longer term commitments because it locks in revenue, and it assists in their need to forecast and provision capacity in their data centers.  The discounts are significant. It has been estimated that a typical three-year contract saves users at least 25% over a single-year contract. 

Although government procurement data isn’t perfect, a recent analysis by Deltek indicated that the government spent about $2 billion on infrastructure as a service in fiscal year 2022. An additional 25% discount applied to this estimated IaaS spend would total $500 million in savings or cost avoidance for the government.

Unfortunately, it appears these multiyear discounts and associated savings have not been effectuated for the government. The Anti-Deficiency Act mandates that the government cannot pre-commit to out-year expenditures prior to Congress allocating funds. In most cases, with some exceptions, this means that government buyers of cloud can only commit annually to procuring cloud services, one year at a time.  

Cloud providers typically provide the government with single-year pricing, for each year, meaning the government may be paying 25% more for cloud than similarly situated commercial entities.

In practice, the government tends to purchase all IT and cloud using three-year contracts but with a unique government twist. The contracts are single base-year commitments with two optional years. Optional years are executed annually, once Congress has authorized a budget. The majority of the government’s discretionary spending for commercial goods and services occurs exactly this way, one incremental option year at a time, within overarching multiyear contracts, often for three or five years. To do anything less, would force the government to run thousands of full procurement competitions for essential services, every year, an absurd impracticality. The multiyear contract, with the later years being optional, has been the standard approach to procuring goods and services for decades.

Unfortunately, for discounting purposes, because of the option years, the market dominant CSPs have not considered three-year government contracts to be the equivalent of three-year commercial contracts. As a result, cloud providers typically provide the government with single year pricing, for each year, meaning the government may be paying 25% more for cloud than similarly situated commercial entities. As noted, this could mean leaving as much as $500 million on the table right now – and more as federal expenditures on cloud services rise.

The single-year money limitation isn’t a new procurement issue and other industries selling to the government have learned to accommodate it. For instance, leases of capital equipment (like copiers or cars) are entirely dependent upon amortization schedules over a designated period of years.  With minor exceptions, the government generally structures capital equipment leases as base year/optional year contracts. The government rarely provides multiyear guarantees for leases even though an early year cancellation would destroy the amortization schedule and upsets the leasing company’s business model. Yet, finance companies have accurately assessed the government exercises option years so predictably, that the risk is negligible. Similarly, when the government leases office space, commercial realtors understand that the government will almost always execute option years and that a three-year deal with option years is virtually the same as a fully committed three-year deal. Most telling, research has indicated there have been no significant incidents of the government failing to exercise option years on cloud contracts. 

Underscoring this dubious risk aversion is the simple fact that every government contract, for any good or service, contains a provision that allows the government to cancel at any time “for convenience.”  In a sense, every commercial entity that sells to the government assumes the risk of cancellation every day. Yet, this never stops companies from selling to the government.  Government contractors know that risk of cancellation, especially for core mission services like cloud infrastructure, is miniscule. Cloud is especially protected given the technical obstacles to changing providers.

Given the reality of the negligible risk, commercial cloud companies should treat multiyear government contracts with option years as firm commitments for the entire duration of the term and discount accordingly. The option years are generally only optional on paper and the risk of mid-term cancellation is almost non-existent.  The first cloud company to accurately assess the nominal risk and provide the same discounting that a commercial entity receives for multiyear commitments will receive a significant first-mover advantage and heartfelt thanks from American taxpayers.