As technology giants make trillions of dollars in long-term commitments to build artificial intelligence infrastructure, Michigan is making its own bet on the AI boom — offering tax incentives, approving massive data centers and preparing an electric grid to serve them. The payoff could mean billions in investment and thousands of construction jobs. But what happens if AI demand falls short of expectations?
ANN ARBOR – Michigan’s rapidly expanding artificial intelligence economy is increasingly tied to one of the biggest corporate spending bets in history.
Nine major technology companies have accumulated roughly $3 trillion in future financial commitments, much of it connected to the enormous infrastructure required to develop and operate artificial intelligence, according to an analysis by The Wall Street Journal.
The commitments include about $1.2 trillion in leases that haven’t started yet and roughly $1.9 trillion in purchase agreements and other contractual obligations involving data centers, computer chips, electricity and other infrastructure.
That is far larger than the roughly $600 billion those companies reported spending on capital projects during the past year.
And Michigan is increasingly along for the ride.
From the massive Oracle/OpenAI Stargate project under construction in Saline Township to other proposed hyperscale data centers across the state, Michigan communities and utilities are making long-term decisions based on essentially the same assumption as Big Tech:
Demand for artificial intelligence computing will continue growing rapidly for years — perhaps decades.
If they’re right, Michigan could capture billions of dollars in investment, thousands of construction jobs and hundreds of permanent technology and infrastructure jobs.
If they’re wrong, the question becomes who absorbs the cost.
$3 Trillion Isn’t Debt — But It’s A Huge Commitment
The $3 trillion figure needs an important qualification.
It is not $3 trillion of hidden corporate debt.
Much of the money represents leases, purchase contracts, energy agreements and other obligations that accounting rules do not yet require companies to record as liabilities on their balance sheets.
But the commitments are real.
Alphabet, Google’s parent company, reported $811 billion in purchase commitments and other contractual obligations as of June 30, compared with $332.4 billion only three months earlier.
Alphabet said those commitments primarily involve technical infrastructure and inventory, along with agreements to secure energy for its data centers. The company also reported $85.2 billion of leases that haven’t yet begun, primarily involving data centers.
Meta reported $349.3 billion in non-cancelable contractual commitments plus another $279 billion in leases that haven’t started, primarily for data centers, colocation facilities and network infrastructure.
Amazon reported roughly $650 billion in total commitments as of June 30, although not all are related to AI because of Amazon’s enormous logistics and retail operations.
The numbers illustrate how quickly the AI infrastructure race is accelerating.
Michigan’s Biggest AI Bet Is In Saline Township
For Michigan, perhaps no company illustrates the opportunity — and potential risk — better than Oracle.
Oracle is a major participant in the OpenAI Stargate data center being developed in Saline Township.
Oracle spent $55.7 billion on capital expenditures during its 2026 fiscal year, more than double the $21.2 billion it spent a year earlier as it rapidly expanded data-center capacity.
The company also has roughly $260 billion in future data-center lease commitments.
Oracle’s expansion has increasingly required outside financing. S&P has downgraded the company’s credit rating to BBB-, one level above junk status, as debt and AI infrastructure spending increased.
That doesn’t mean Oracle is in financial trouble.
Its cloud infrastructure revenue increased 93% in its most recent quarter, while remaining performance obligations — essentially contracted future business — reached $638 billion.
The company is betting that exploding AI demand will justify today’s enormous infrastructure investments.
Michigan is making the same bet.
A Potential $50 Billion Michigan Project
The Citizens Research Council of Michigan estimates the Saline Township Stargate development involves nearly 600 acres and approximately 2 million square feet of buildings.
Land, buildings and other physical infrastructure could represent roughly $5 billion of investment.
But that doesn’t include the extraordinarily expensive GPUs and other computing equipment inside the buildings.
Using chip costs at a comparable Stargate facility in Texas, CRC estimates total investment in the Michigan project could eventually exceed $50 billion.
Oracle says the project will create more than 2,500 union construction jobs, 450 permanent onsite jobs and another 1,500 supporting jobs across Washtenaw County.
Those construction jobs represent a significant economic boost.
But the permanent employment numbers also highlight one of the central debates surrounding data-center incentives.
Unlike automobile plants and other traditional manufacturing projects, data centers can require tens of billions of dollars of investment while producing comparatively few permanent jobs.
Then There’s Michigan’s Electric Grid
The larger long-term issue may be electricity.
The Saline Township project alone is expected to require roughly 1.4 gigawatts of power — electricity demand comparable to a major industrial complex.
Serving facilities of that size can require new transmission lines, substations and potentially new generating capacity.
Those investments can remain on the electric system for decades.
That creates an uncomfortable question:
What happens if utilities build infrastructure around projected data-center demand that never materializes?
Michigan Attorney General Dana Nessel has already raised that concern in state utility proceedings.
In testimony involving Google’s proposed Michigan data center, Nessel’s office argued that special contracts must protect existing utility customers from paying for “stranded assets, bad bets, and failed investments” if anticipated data-center demand disappears.
Gov. Gretchen Whitmer has similarly said data-center developers should pay the full costs they impose on Michigan’s electric system rather than shifting those costs onto existing customers.
That matters because Michigan households are already dealing with rising electricity prices while utilities face billions of dollars in spending to replace aging grid infrastructure.
Michigan’s AI Gamble
Michigan isn’t alone in betting on artificial intelligence.
Technology companies, utilities, private-equity firms, banks and governments around the world are racing to build the infrastructure needed to support it.
And there are powerful reasons to believe demand will continue growing.
But the sheer scale of the financial commitments now being made is creating another question for Michigan policymakers.
The Citizens Research Council warned in a June report that growing financial interdependencies within the AI industry could magnify losses if demand fails to meet current expectations.
CRC didn’t argue Michigan should reject data centers. It concluded they can produce construction jobs, investment and tax revenue.
But it also warned that data centers generate relatively modest sustained economic activity compared with many traditional industrial investments and questioned whether aggressive incentives are always justified.
That’s the balancing act Michigan now faces.
The state has an opportunity to participate in what could become one of the defining industries of the next several decades.
But Michigan is no longer merely watching Silicon Valley make an enormous bet on AI.
With billions of dollars of data centers, power infrastructure and tax incentives now on the table, Michigan is betting right alongside it.
COMING SOON: Who is financing Michigan’s multibillion-dollar AI data center boom? MITechNews will examine the private capital, debt and long-term leases behind these massive projects — and who ultimately carries the financial risk if the AI boom falls short of expectations.
