Carstensen: CT should invest up to $150M to build applied-AI economy

Hartford Business Journal – Friday, September 25, 2026
By David Krechevsky

Connecticut should invest up to $150 million to help turn the state’s insurance, defense manufacturing and bioscience industries into hubs for applied artificial intelligence, according to a new UConn policy brief released Thursday.

The 14-page report, produced by the University of Connecticut’s Connecticut Center for Economic Analysis (CCEA), argues AI offers an opportunity to reverse decades of sluggish economic growth. It also suggests that Connecticut should not try to compete directly with technology centers such as Boston and New York in developing large AI models and a massive computing infrastructure.

Instead, the state should focus on what the report calls the “application layer” of AI — utilizing the technology in industries where Connecticut already has significant expertise, proprietary data and physical assets.

“Connecticut does not need to build the AI brain; it needs to become the best place in the country to apply it, and the preferred place to found the companies that do,” the report states.

Authored by CCEA Director Fred Carstensen, a UConn professor of finance and economics, the report identifies insurance and actuarial AI, defense and advanced-manufacturing AI, and bioscience and genomic AI as three areas where the state holds competitive advantages.

The recommendations come amid longstanding concerns about the state’s economic growth.

Since 1989, U.S. employment has increased nearly 50%, while Connecticut employment has grown roughly 2%, according to the report. The state also took 18 years to regain its March 2008 employment peak of about 1.72 million jobs, it states.

CCEA argues that Connecticut businesses will adopt AI regardless of state policy. The larger economic question, it says, is whether the companies, intellectual property, revenue and high-paying jobs created around that adoption remain in Connecticut or flow to technology vendors elsewhere.

To encourage that activity to remain in the state, the report proposes a one-time appropriation of between $100 million and $150 million from Connecticut’s current budget surplus.

Under the plan, Connecticut Innovations, the state’s quasi-public venture capital arm, would deploy state capital, but only alongside at least $2 in private or institutional investment for every $1 of public money. CCEA says that structure could turn a $100 million-to-$150 million state commitment into $300 million to $450 million in combined investment.

Rather than creating new agencies, the report recommends using existing entities such as Connecticut Innovations and the Manufacturing Innovation Fund to deploy the funding.

Among its proposals is an “Applied AI Data Commons” that would give qualified startups and researchers controlled access to valuable industry data without requiring companies to relinquish control of sensitive information.

CCEA also recommends creating an “AI Anchor Customer Compact,” under which major Connecticut insurers, manufacturers, health systems and universities would identify business problems that local startups and university researchers could attempt to solve. Successful projects could give startups their first major customers and a path toward commercial growth.

Currently in downtown Hartford, efforts are underway to establish a 100,000-square-foot artificial intelligence and workforce development center that will feature collaborative labs, classrooms and project spaces aimed at helping workers, students, businesses and entrepreneurs learn how to use AI and other emerging technologies. The project is being funded with $20 million from the state’s Innovation Clusters Program.

The city of Stamford is also considering the launch of an AI institute with UConn with the help of Innovation Clusters Program funds.

Another proposal recommended in the CCEA report would provide businesses with $35,000-to-$75,000 “AI innovation vouchers” that could be used with Connecticut-based AI companies, universities or colleges for AI-readiness assessments, pilot projects and cybersecurity preparation.

The report also recommends workforce programs, including six-month, industry-specific AI credentials offered through UConn, the Connecticut State Colleges and Universities system and community colleges, as well as programs intended to keep more college graduates working in Connecticut.

Some colleges in Connecticut have already begun offering AI credentials and courses.

CCEA does not advocate building massive data centers. It does, however, recommend studying whether Connecticut eventually needs a moderate-sized, secure data center capable of handling sensitive defense, insurance and healthcare workloads.

A preliminary assessment in the report identifies the Millstone nuclear complex and Killingly natural-gas plant as two potential locations, though Carstensen says both face constraints and recommends a formal state siting and feasibility study.

The strategy also ties AI development to QuantumCT, which in July received an initial $15 million National Science Foundation award and could receive up to about $145 million more over a decade if it meets commercialization milestones. CCEA argues that developing an applied-AI ecosystem would help Connecticut commercialize quantum technologies in defense, biotechnology and financial services.

CCEA also proposes publishing an annual assessment that would measure the strategy against indicators, including productivity, new company formation, wage growth, graduate retention and the amount of technology-vendor revenue retained in Connecticut.

Access this article at its original source.