Amazon, Anthropic, Microsoft and OpenAI Back $500M RAISE US Coalition to Retrain Workers Displaced by AI

The numbers land with a kind of blunt inevitability. In 2026, artificial intelligence has been cited as the cause of 87,714 U.S. job cuts—more than in every prior year of AI-driven displacement combined. The IMF has issued fresh warnings about an entry-level job shock that will hit recent graduates and career-changers disproportionately hard. And in the background, a survey commissioned by Anthropic found that 35% of American workers now expect AI to handle the majority of their current tasks within twelve months.

Into this landscape, on June 25, a group of the very companies building those AI systems announced they would fund a coordinated national response. RAISE US—Rebuilding America’s Infrastructure, Skills, and Economy—launched with more than $500 million in committed funding and a stated goal of reaching $1 billion. The founding funders include Amazon, Anthropic, Microsoft, the OpenAI Foundation, Bank of America, and IBM.

Leadership Designed for Credibility

The choice of Gina Raimondo as CEO is deliberate and calculated. As the 40th U.S. Secretary of Commerce under the Biden administration, Raimondo was the architect of the CHIPS and Science Act, the landmark legislation that reshaped semiconductor manufacturing policy in the United States. She also played a central role in AI policy development, including the Commerce Department’s AI Safety framework, and presided over the early export control negotiations that now define how cutting-edge AI chips flow across borders.

Co-founding the organization alongside her is Eric Holcomb, the former Republican Governor of Indiana, whose inclusion signals the bipartisan political positioning RAISE US is explicitly seeking. “The workforce transformation we’re facing is not a partisan problem,” Raimondo said at the launch. “It’s an American problem, and it requires American institutions—business, government, labor, and education—working together.”

The advisory board reinforces that positioning. Former Republican House Speaker Paul Ryan sits alongside AFL-CIO President Liz Shuler, the nation’s top union leader. Billionaire investment manager Stephen Schwarzman of Blackstone is paired with MIT labor economist David Autor, whose widely-cited research on job polarization from technological change has defined the academic understanding of automation’s workforce effects. Economists Erik Brynjolfsson of Stanford and Raj Chetty of Harvard complete the advisory lineup.

What RAISE US Will Actually Do

The organization’s operating model deliberately avoids the federal government channel. Rather than directing funding through Washington agencies—where appropriations battles, political reversals, and bureaucratic lag can paralyze programs—RAISE US will work directly with state governors and major employers.

Four states have signed on as initial pilot partners: Arkansas, Connecticut, Maryland, and Utah. Each represents a different economic profile: a rural manufacturing state, a coastal financial services hub, a mid-Atlantic government contractor corridor, and a Mountain West tech corridor. The idea is to test retraining models that can later be adapted to the vastly different labor market conditions across the country.

Critically, RAISE US has embedded a non-standard success metric into its mandate. Programs will be measured not by enrollment numbers—the traditional output metric that allows job retraining initiatives to claim success while participants fail to find employment—but by whether participants land and keep jobs in less automation-exposed roles. That outcome-based accountability is explicitly designed to address the documented failure of prior federal retraining efforts.

A 2026 study that tracked 23 million participants in federal retraining programs found limited success in actually placing workers into roles with lower automation risk. The researchers concluded that most programs prioritized speed and scale over matching participants with genuinely durable job categories.

The Tension Underneath the Initiative

There is an awkwardness embedded in this announcement that observers have not let pass unnoticed. The companies funding RAISE US—Amazon, Anthropic, Microsoft, OpenAI—are precisely the organizations whose products are generating the displacement the initiative aims to address. Amazon’s AI warehouse and logistics automation has eliminated roles across its fulfillment network. Microsoft’s Copilot suite is actively replacing tasks previously handled by knowledge workers across finance, legal, and administrative functions. OpenAI’s and Anthropic’s models are being deployed by enterprises specifically to reduce headcount in customer service, content generation, and data analysis.

“It’s a bit like tobacco companies funding cancer research,” noted one labor economist who asked not to be identified. “The intent may be genuine, but the structural conflict doesn’t disappear just because you acknowledge it.”

Raimondo has addressed this directly in interviews. “The technology is advancing regardless of whether any one company builds it,” she said. “The question isn’t whether the transition happens—it’s whether American workers are supported through it or left to figure it out alone.”

Against the Clock

The urgency is real. Anthropic’s own survey data puts 35% of workers expecting near-complete AI task takeover within a year. The IMF’s spring 2026 outlook specifically flagged entry-level positions in law, accounting, and knowledge work as facing immediate, structural risk rather than the gradual displacement economists historically modeled for physical automation.

The $500 million RAISE US has secured—and the $1 billion it is pursuing—sounds substantial until set against the scale of the challenge. Meaningful retraining for a displaced worker typically costs $15,000 to $40,000 in direct program expenses when accounting for instruction, credentialing, income support during transition, and job placement services. At $500 million, the initiative can fund perhaps 15,000 to 30,000 full retraining pathways. The U.S. Bureau of Labor Statistics estimated in May that 2.4 million Americans are currently in roles at high displacement risk over the next 24 months.

That math gap will not escape legislators, union leaders, or the workers the initiative is meant to serve. Whether RAISE US represents the beginning of an adequate national response or a well-funded acknowledgment that the problem is larger than any private coalition can solve may take years to answer. But June 25, 2026 marks the moment the AI industry formally accepted, in writing and in dollars, that its products have a workforce cost that cannot be left entirely to others to manage.