Why Applied Digital's Data Center Project Stalled

A 430MW data center. Two buildings totaling nearly 1.8 million square feet. Up to $16 billion in investment. For Toronto, South Dakota—a small town in Deuel County near the Minnesota border—that kind of project doesn't just change the local economy; it rewrites it.

It's not happening.

Applied Digital confirmed to local news outlet Keloland that its proposed Toronto data center is "unlikely to go forward," effectively withdrawing from a project the company had publicly floated less than a year ago. The statement was carefully worded—diplomatic, even—but the message was clear: South Dakota didn't make the math work.

Understanding why reveals something important about how data center investment actually flows and what states either do or don't do to attract it.

A Project That Made Headlines, Then Quietly Died

Applied Digital proposed the Toronto campus in May 2025. The specs were serious: 430 megawatts of capacity across two 907,000 square-foot buildings. For context, 430MW is roughly equivalent to the power draw of a small city—or, increasingly, what a single hyperscale AI training cluster demands. This wasn't a modest edge facility or a colocation retrofit; it was a purpose-built, large-scale infrastructure commitment.

The state's own Executive Vice President of External Affairs, Nick Phillips, acknowledged that if approved, Applied Digital would have become one of the largest property taxpayers in South Dakota. That's not a rounding error; that's generational tax base.

Yet despite the scale of the opportunity, the project collapsed before a shovel touched dirt. Applied Digital's public statement nodded toward the underlying issue without naming it directly: the company cited the importance of "a competitive policy environment that enables our customers to operate effectively"—a phrase that sounds diplomatic but functions as a clear diagnosis.

North Dakota Did It Right. South Dakota Didn't.

Applied Digital already operates campuses in Ellendale and Harwood, North Dakota. That's not a coincidence; it's a data point.

North Dakota offers sales tax exemptions on the hardware that makes data centers function: computer equipment, cooling systems, batteries. These exemptions matter more than they might seem. At the scale Applied Digital operates—and certainly at the scale of a 430MW buildout—equipment costs run into the hundreds of millions of dollars. A sales tax exemption on that equipment can represent tens of millions saved before a single server goes online.

South Dakota offers no equivalent. Phillips confirmed this directly when speaking to Keloland last year, noting the cost differential between the two states. Attempts to change that—specifically, efforts to pass legislation creating comparable tax exemptions—failed in the state legislature.

That legislative failure didn't just kill one project; it sent a clear signal to every data center developer watching: South Dakota isn't competing for this capital.

This is where infrastructure development diverges from other industries. A manufacturer might absorb higher input costs if the labor market or logistics are favorable enough. Data center operators have fewer such offsets. Power costs, land, and equipment taxation are the primary levers. When one state exempts equipment and another taxes it, the calculus tips fast—especially when the competing state already has proven infrastructure and an established relationship with the developer.

What South Dakota Was Walking Away From

It's worth sitting with the economic loss here because "unlikely to go forward" is easy to skim past.

Property tax revenue at that scale would have funded schools, roads, and local services in Deuel County for decades. Data centers also create a smaller but highly skilled permanent workforce—operations engineers, facilities technicians, security staff—alongside the much larger construction employment spike during the build phase. A project of this magnitude typically generates thousands of construction jobs over a multi-year buildout.

Beyond direct employment, the multiplier effects matter. Workers spend money locally. Contractors establish local vendor relationships. Utility infrastructure gets upgraded to handle data center loads, which benefits surrounding communities and sometimes opens up capacity for other industrial users.

None of that arrives in Toronto, South Dakota now.

The harder truth is that rural communities don't get many chances at economic anchors of this size. Deuel County isn't a major metropolitan market. Toronto isn't competing with Phoenix or Northern Virginia for data center investment on normal terms. A project like this was, realistically, a once-in-a-generation opportunity—contingent on policy conditions that the state legislature declined to create.

The Data Center Bill of Rights: Protection or Obstacle?

There's a second layer to this story, and it complicates the narrative in useful ways.

Even as Applied Digital's project was stalling over tax policy, South Dakota's legislature was advancing Senate Bill 135—dubbed the "Data Center Bill of Rights." The bill, delivered to Governor Larry Rhoden on March 12, is designed to protect residents from increased utility costs and potential power shortages that large-scale data center operations can impose on regional grids.

The concern is legitimate. Data centers at the 100MW-plus scale draw enormous amounts of power, and in states with less robust grid infrastructure, that demand can stress capacity, raise rates for residential customers, and create reliability issues. Legislators pushing SB 135 aren't wrong to think about those risks.

But here's the tension: the same legislative session that failed to create tax incentives for data centers also moved to create protective restrictions around them. From a developer's perspective, that's a particularly difficult combination—higher costs with additional regulatory constraints attached. It's not that either policy is unreasonable in isolation; it's that together, they signal a state that hasn't decided whether it wants data center investment or not.

That ambivalence is itself a deterrent.

What Comes Next — For Applied Digital and for South Dakota

Applied Digital said it will "continue to monitor opportunities where the right conditions exist for long-term investment." That's not an idle phrase. The company has real projects operating in North Dakota, real customers with real power demands, and real capital to deploy. If South Dakota's policy environment shifts—if a future legislature passes equipment tax exemptions, if the regulatory framework clarifies—the company will notice.

But markets don't wait. The data center construction boom is accelerating, driven by AI infrastructure demand that shows no sign of plateauing. Applied Digital and developers like it are making siting decisions now. Every quarter that South Dakota spends without a competitive incentive framework is a quarter where that capital flows somewhere else—to Iowa, Nebraska, Wyoming, or back north to the Dakotas where the policy environment is already proven.

For South Dakota, the path forward isn't complicated, even if the politics are. The state has genuine assets: available land, a cold climate (a natural advantage for cooling efficiency), proximity to Midwest power generation, and lower land costs than saturated markets. Those assets don't disappear. But they don't do the work alone, either.

The question Governor Rhoden and the legislature will eventually have to answer is a straightforward one: does South Dakota want to compete for this investment or not? Applied Digital already answered that question for them—just not in the way anyone in Deuel County was hoping.

Ready to explore opportunities in the data center market? Visit [InfraSale Marketplace](https://infrasale.com/marketplace) today!