Are we training people for the jobs that are coming?

Those two numbers, 18% and 78%, tell a story that most workforce planning doesn't account for.

According to the Federal Reserve's January 2026 survey, only 18% of U.S. firms have adopted AI in at least one business function. But those firms tend to be large employers. The result: 78% of the American labor force already works at a company where AI is being deployed. The workers most affected aren't at cutting-edge startups. They're at hospitals, logistics companies, banks, retail chains and government agencies.

At the same time, $700 billion is flowing into AI infrastructure in 2026 alone (Deloitte, March 2026). This includes data centers, GPU clusters, model training facilities, power grid upgrades and the human capital to run all of it.

A small share of firms account for the vast majority of the workforce, meaning AI's impact is broader than adoption rates suggest.

Where the $700 billion is going

The Deloitte analysis breaks AI infrastructure investment into five categories:

1. Data centers and compute, the physical infrastructure (40% of spend)
2. Cloud and AI services like AWS, Azure and Google Cloud (25%)
3. Enterprise AI applications, industry-specific tools (15%)
4. Talent and training, hiring AI engineers and retraining existing workers (12%)
5. Power and grid infrastructure, meeting the energy demands of AI compute (8%)

Category 4 is the one that matters most for workforce planners. $84 billion in talent-and-training investment means companies are spending at scale to either hire or reskill AI-capable workers. The question is whether the education and workforce pipeline can meet that demand.

Connecting graduates to employers: the Georgetown data

The Georgetown University Center on Education and the Workforce (CEW) released updated projections in February 2026 that quantify the gap:

• By 2032, the U.S. economy will need 5.25 million more skilled workers than the current pipeline will produce.
• The shortfall is concentrated in five sectors: healthcare, IT/cybersecurity, advanced manufacturing, clean energy and data analytics.
• Texas alone will account for ~480,000 of those unfilled positions, the second-largest state shortfall after California.

This isn't a future prediction. It's a present reality that compounds every year the pipeline doesn't adjust.

Healthcare alone accounts for more than a third of the projected shortfall, driven by aging demographics and expanding access mandates.

Five things happening across the ecosystem

1. States are racing to attract AI data center investment

Texas, Virginia, Georgia and Ohio are the leading states for new data center construction. Each new facility creates 50-200 permanent high-skill jobs (network engineers, data center techs, security analysts) plus 500-2,000 construction jobs. But the competition for these projects depends on three factors: power availability, fiber connectivity and local workforce readiness.

Texas leads all states in AI-related project announcements, driven by available land, favorable energy prices and corporate-friendly regulation.

2. Healthcare is the largest and most urgent sector

The Georgetown data shows healthcare accounting for 1.8 million of the 5.25 million shortfall. This includes registered nurses, medical technicians, home health aides, behavioral health counselors and clinical data analysts. In Texas, the healthcare shortfall is acute in rural regions and exurban growth corridors (Williamson, Denton and Collin counties).

3. The "middle skills" gap is widening

The most severe shortages aren't at the PhD level. They're at the associate degree, certificate and apprenticeship level. Electricians, HVAC technicians, medical coders, cybersecurity analysts, CNC machinists and data center technicians. These are the jobs where demand is growing fastest and the pipeline is thinnest.

4. Employer-led training is outpacing academic-led training

Companies like Amazon (Career Choice), Google (Grow with Google), JPMorgan Chase ($350M workforce investment) and Salesforce (Trailhead) are building their own credential and training programs because the academic pipeline can't move fast enough. The risk for universities: if employers bypass the degree entirely, the value proposition of traditional enrollment erodes further.

5. Site selection decisions increasingly hinge on talent data

When a company evaluates relocating to Texas, the first question used to be about real estate and tax incentives. Now the first question is: "Show me the talent pipeline." EDCs that can produce real-time data on local program completions, skills alignment and employer demand have a decisive advantage in competitive site selection processes.

Who's building the bridge

The institutions and organizations getting this right share three characteristics:

1. They use data to align programs to demand, not anecdotes or five-year-old surveys, but real-time labor market signals matched to their specific region.
2. They build employer partnerships into the curriculum. Co-ops, apprenticeships, clinical rotations and embedded industry credentials are not afterthoughts but required components of every program.
3. They measure what matters, not just enrollment and graduation rates but placement rates, earnings premiums and employer satisfaction. They make those metrics visible to prospective students, governing boards and state legislatures.