Washington State Runs the Cloud That Runs AI

Washington State's role in artificial intelligence is foundational and often overlooked: a very large share of the world's AI workloads run on infrastructure operated by companies headquartered there.

The cloud concentration:

Seattle and its suburbs host the headquarters of two of the three dominant global cloud platforms. Cloud infrastructure is where nearly all AI training and inference actually happens for companies that do not own data centres, which is almost all of them.

This makes decisions made in Washington, about pricing, capacity allocation, chip procurement and regional expansion, consequential for the entire industry. When cloud capacity is constrained, AI companies worldwide feel it.

Both companies also conduct substantial AI research and product development locally, and one has a deep partnership with a leading model laboratory that has shaped how frontier AI reaches enterprise customers.

The engineering base:

The region's technology employment is enormous and long-established, dating to the personal computer era. That produced a deep pool of experienced software and systems engineers, plus a culture of building large-scale distributed systems, precisely the expertise AI infrastructure requires.

The University of Washington's computer science programme is among the strongest in the country and a significant source of machine learning researchers, and the Allen Institute for AI, established as a non-profit research organisation, has produced influential open research and models.

Energy advantage:

The Pacific Northwest has substantial hydroelectric generation, providing low-carbon electricity at historically low prices. Central Washington, in particular, attracted major data centre development on that basis, with facilities clustered near dams.

That advantage is narrowing. Hydropower output varies with snowpack and rainfall, drought years reduce availability, and demand growth has consumed much of the surplus that made the region cheap. New capacity increasingly requires transmission investment or alternative generation.

Other sectors:

Aerospace remains economically significant, applying AI to design simulation, manufacturing quality and predictive maintenance. Global health organisations based in Seattle apply machine learning to disease modelling and health systems research. Agriculture in eastern Washington uses precision techniques for orchards and vineyards.

Constraints:

Housing and living costs are high, though below California's, and the region competes for the same engineering talent as the Bay Area.

Water and environmental review affect data centre siting, and the state's climate commitments create tension with the electricity demands of AI infrastructure.

Economic concentration is a risk: the region's fortunes are closely tied to a small number of very large employers, and their restructuring decisions have outsized local effects.

The outlook:

Washington's position rests on operating the infrastructure layer of the AI economy. That is less visible than model development and arguably more durable: cloud platforms accumulate switching costs, and rebuilding equivalent global infrastructure would take a decade and enormous capital.

As AI workloads grow, the companies running the underlying compute capture value regardless of which models win, and that is the position Washington occupies.