The Economic Development Administration (EDA) comes up whenever businesses search for “government tech grants,” but the honest answer for almost every company is that EDA money isn’t built to reach you directly. Here’s how EDA funding actually works in 2026, cited to official sources, and what to pursue instead. (dgm implements AI software; EDA is an economic-development funder, not an AI-adoption one — see the end.)

Who EDA actually funds

EDA is a US Department of Commerce agency, and its grant programs are designed to flow to the organizations that build regional economies — not to individual companies. Its eligible applicants are generally economic-development planning organizations, Tribal governments, states and their political subdivisions, institutions of higher education, and nonprofits working in cooperation with local government. A private for-profit business is not a typical EDA grantee.

That doesn’t mean businesses see no benefit. They do — but indirectly. EDA dollars fund infrastructure, planning, and Revolving Loan Funds (RLFs) that local intermediaries then lend out to businesses in their region. So the path from an EDA grant to your company usually runs through a city, a district, a university, or a nonprofit lender — not a direct application you file yourself.

The closure context you should know about

There’s an important backdrop in 2026: EDA’s own FY2026 Congressional Budget Submission states that the Administration is not requesting funding for EDA’s grant programs and describes an orderly closure and grant close-out (an estimated multi-year effort to wind down roughly 2,800 open grants). Translation: this is not a moment when new, business-facing EDA opportunities are expanding. Before you build a funding plan around any EDA program, confirm it’s currently open — many are not.

The AI Upskill Accelerator: real, but not for a single business

One genuinely new and AI-specific EDA program launched in 2026 deserves a precise description, because it’s easy to misread. The EDA AI Upskill Accelerator was announced on May 11, 2026 as a roughly $25 million Notice of Funding Opportunity, expecting on the order of 5–8 awards between $1 million and $8 million, over 24–36 month performance periods.

Here’s the catch that the headlines miss: a single private business cannot apply for itself. The program funds employer-led sectoral partnerships, and the lead entity of the partnership must be an EDA-eligible applicant — typically a state or local government, an institution of higher education, a nonprofit intermediary, a workforce board, or an economic-development organization. A lone employer or a standalone training vendor is not the intended applicant. (The application deadline and the exact eligible-applicant taxonomy should be confirmed directly in the official NOFO, as those details are most reliably read there.)

So if you’re an individual company hoping to get AI-training cash from this program, the realistic route is to join a partnership convened by an eligible lead — not to apply alone.

The Good Jobs Challenge is not an option

You’ll still find blog posts pointing businesses to EDA’s Good Jobs Challenge. Skip them. It was a one-time American Rescue Plan (FY21) program, its awards went to regional workforce “system lead” intermediaries rather than individual businesses, and a final installment was announced in early 2025. EDA has indicated it will not run further competitions. Presenting it as available funding would be misleading — so we won’t.

What to pursue instead

If you’re a business that wants to fund an AI project, the realistic federal levers live elsewhere:

  • R&D tax incentives — the federal R&D tax credit (IRC §41) and Section 174A expensing, if you build or customize AI. These are the strongest broad levers because they don’t require winning a competition. See AI Tax Incentives for US Businesses (2026).
  • SBIR/STTR and NSF — non-dilutive grants for genuine R&D, if you’re developing new technology. See How to Apply for SBIR/STTR Grants.
  • SBA loans — to finance the purchase and implementation of technology. Not free money, but real capital.

The common thread: there is no broad federal grant that simply pays an ordinary business to adopt AI, and EDA is not the exception to that rule.

How dgm helps

dgm implements osFoundry and other AI software for US businesses — scoping, deploying, automating, and training your team. EDA grants are outside that scope: they flow to governments and intermediaries, not to companies adopting AI. What dgm will do is be honest about which funding levers actually apply to you, point you to the real ones, and then build the AI — rather than send you chasing a program you can’t qualify for.