Why Connecticut's Tax Breaks Aren't Attracting Data Centers

Tax breaks are supposed to be the universal language of economic development. Cut the bill, and they will come. Connecticut tried that playbook — and almost nobody showed up.

The state rolled out significant data center tax incentives designed to position itself as a competitive destination for the infrastructure driving cloud computing, AI workloads, and enterprise IT. The pitch made sense on paper. Data centers create jobs, generate tax revenue, consume enormous amounts of power that utilities love, and signal to the broader business community that a state is serious about the digital economy. Connecticut wanted a piece of that.

So far, only one developer has actually taken the state up on its offer.

That's not a rounding error. It's a signal worth taking seriously — because understanding *why* the incentives aren't working reveals something important about what actually drives data center site selection decisions and what Connecticut would need to change to get real traction.

What Connecticut Actually Offered

The state's data center tax incentives were structured to reduce the cost burden on developers and operators — the kind of breaks that, in theory, can meaningfully shift the economics of a capital-intensive project. Data centers are expensive to build. A hyperscale facility can run $1 billion or more in construction costs alone, and ongoing operational expenses — particularly power — are relentless. Any reduction in tax liability matters.

The logic behind the incentive was sound: data centers don't just appear where land is cheap; they appear where the total cost of ownership over a 20-to-30-year asset life is lowest. Connecticut's offer was an attempt to compete on that long-term calculus.

The objective was straightforward — attract developers, generate construction activity, create permanent technical jobs, and establish the state as a node in the broader Northeast data center corridor. Given that New York and New Jersey already have substantial data center footprints, and Northern Virginia dominates the national market, Connecticut was positioning itself as an alternative with better incentives if not better fundamentals.

The problem is that incentives can only do so much when the fundamentals aren't there.

One Taker. That's It.

The data center development boom of the last several years has been extraordinary by almost any measure. Global data center investment has accelerated sharply, driven by AI infrastructure buildout, cloud provider expansion, and enterprise colocation demand. Northern Virginia alone added gigawatts of capacity. Phoenix, Dallas, Chicago, and Atlanta have all seen significant development activity. Even secondary markets like Columbus, Ohio, and Reno, Nevada, have attracted major investment.

Connecticut, meanwhile, can count its data center tax break beneficiaries on one finger.

That's striking when you consider how aggressively developers have been scouting new markets. The constraints in established hubs — power availability, land scarcity, community opposition — have pushed developers to look harder at alternative locations. Connecticut should theoretically be in the conversation. It's in the Northeast, near major financial and media markets, with existing fiber infrastructure and a reasonably educated workforce.

Yet the tax incentive program sits largely unclaimed, which tells you the barriers to entry aren't primarily financial.

What's Actually Holding Developers Back

Site selection for a data center is a multi-variable optimization problem, and tax incentives are just one variable — and rarely the decisive one.

Power is usually the first conversation. Data centers are among the most power-hungry facilities ever built; a single hyperscale campus can draw 100 to 500 megawatts continuously. Developers need not just available power, but *reliable, scalable, competitively priced* power with a utility willing to build transmission infrastructure to serve the site. Connecticut's electricity rates are among the highest in the nation — consistently in the top five states by cost per kilowatt-hour. A tax break that saves a developer millions on property taxes can be wiped out by elevated power costs over a decade of operations.

Then there's the regulatory environment. Connecticut has a reputation — fair or not — for being a difficult state to move quickly in. Data center developers are accustomed to markets where they can go from site control to energization in 18 to 36 months. Permitting complexity, environmental review timelines, and local opposition can stretch that window considerably. Speed to market is itself a form of return on investment, and any state that can't promise a predictable approval pathway is starting at a disadvantage.

Land availability and configuration matter too. Data centers need large, flat, contiguous parcels with access to fiber, water for cooling, and transmission lines nearby. Connecticut's geography — dense development in the southwest, hilly terrain elsewhere — limits the number of sites that check all those boxes simultaneously. The fact that a developer has proposed building a condo complex on at least one relevant property suggests that some of the land Connecticut might have hoped to attract data centers to is being redirected toward residential use, where market dynamics are equally compelling.

How Other States Won the Competition

The contrast with successful data center markets is instructive. Virginia's dominance in Northern Virginia wasn't built on tax breaks alone — it was built on a combination of low power costs, shovel-ready sites, a streamlined permitting process, and decades of relationship-building between the state, utilities, and the developer community. By the time Virginia formalized its data center tax exemptions, the ecosystem was already self-reinforcing.

Georgia took a different approach. Atlanta emerged as a major market partly because Georgia Power and state economic development officials worked together to create a frictionless experience for large power customers. Developers knew what they were getting: competitive rates, a cooperative utility, and a state government that viewed data center investment as genuinely strategic rather than incidental.

Nevada and Arizona built their cases on cheap land, abundant (if increasingly stressed) water, low power costs, and business-friendly regulatory environments. They didn't just offer tax breaks — they removed friction across the entire development process.

The common thread isn't generous incentives. It's *certainty*. Developers making billion-dollar, multi-decade commitments want to know the rules won't change, the power will flow, the permits will come, and the community won't turn hostile. Tax breaks are a sweetener; certainty is the meal.

What Connecticut Would Need to Change

Connecticut isn't without assets. Its location in the Northeast corridor has genuine value for latency-sensitive applications. Its existing fiber infrastructure is real. And if the state is serious about attracting data center investment, there are concrete steps that would matter more than adjusting the tax incentive structure.

The most important would be working directly with utilities to develop a competitive large-power-customer program — something that acknowledges data centers' unique load profile and offers rate structures that make the long-term math work. That's a heavier lift than passing a tax bill, but it would move the needle in ways that incremental tax adjustments won't.

Second, Connecticut could identify and pre-permit specific sites — essentially creating a menu of ready-to-develop parcels where environmental review is complete, zoning is aligned, and utility infrastructure is mapped. Some states call these "megasites." They're expensive to develop upfront, but they dramatically reduce the time and risk for developers, which translates directly into project feasibility.

Third — and perhaps most importantly — the state needs to make a genuine, sustained effort to engage the developer community before the incentives are finalized, not after. The fact that only one developer used the program suggests the design may not reflect what the market actually needs. That's a solvable problem, but it requires listening.

Connecticut still has time to compete for the next wave of data center investment, which is substantial. AI infrastructure buildout is driving demand that the current supply pipeline can't fully absorb. New markets *will* emerge. Whether Connecticut is among them depends on whether the state is willing to do the harder work that goes beyond the tax break — because the tax break, it turns out, wasn't enough.

Explore more about InfraSale Marketplace

Internal Link Suggestions

• [INTERNAL LINK: Connecticut's Economic Development Strategies]
• [INTERNAL LINK: Data Center Trends in the Northeast]
• [INTERNAL LINK: The Future of AI Infrastructure]