Anthropic launches Claude Corps: $150M fellowship to place 1,000 AI workers inside US nonprofits
Anthropic's Claude Corps will train and pay 1,000 early-career fellows $85K each to spend a year building AI tools inside US nonprofits, starting October 2026.
Anthropic has announced Claude Corps, a $150 million national fellowship that will train 1,000 early-career workers to build AI-powered tools and workflows using Claude, then place them full-time inside US nonprofits for a year. The first cohort of 100 fellows starts in October 2026, with applications open until July 17.
The structure is straightforward: Anthropic funds the program and provides the AI expertise. CodePath, a nonprofit and the largest provider of collegiate computer science education in the US, acts as the official employer of record, handling payroll, W-2s, and benefits. Social Finance leads measurement and evaluation, and is building the longer-term financial infrastructure to help the program scale. Host nonprofits direct the day-to-day work.
What fellows actually do
This is not an internship. Fellows are full-time CodePath employees earning an $85,000 salary, plus a $10,000 grant, with full benefits for the duration of the 12-month placement.
Training covers prompt design, building with Claude and the Anthropic API, and running AI evaluations. After placement, fellows receive around five hours of structured training each week. The rest of their time goes toward live projects at their host organization. Each fellow also gets support from a CodePath mentor, office hours with Anthropic for technical questions, and an expanded Claude token budget.
The work spans a wide range of mission areas: workforce development, public health, housing, food security, veterans services, and education. Fellows own at least one project from start to finish, which means they leave with something concrete to show.
Who can apply
The eligibility bar is deliberately low. Applicants need to be 18 or older, have authorization to work in the US, and have no more than two years of full-time work experience. A college degree is not required.
That’s a meaningful design choice. Anthropic is explicitly trying to reach people who are early in their careers and haven’t had access to the kind of AI training that’s increasingly becoming a professional requirement. Applications for the October 2026 cohort close July 17, 2026. Subsequent cohorts are scheduled for January and August 2027.
What this means if you run a nonprofit
Host organizations get a fully funded, 12-month placement at no cost to them. Fellows arrive trained and supported, and the host directs the actual work. Organizations also receive Claude access, implementation funding, and direct support from the program.
Among the confirmed hosts for the first cohort are RAINN, Goodwill Industries International, Code for America, Year Up United, the YMCA of Greater Charlotte, and the Montgomery County Food Bank in Texas. More than 400 nonprofits are expected to participate overall.
To be eligible, organizations must be US-based 501(c)(3)s and Claude for Nonprofits Team or Enterprise customers. For the October 2026 cohort specifically, host organizations need an in-person or hybrid workplace. Fully remote nonprofits can apply but won’t be eligible until the 2027 cohorts. A host information webinar ran on June 17, and host-selection notifications are expected in late August.
If you work at an eligible nonprofit and want to explore hosting a fellow, the host FAQ is the clearest starting point.
The wider context worth knowing
Anthropic CEO Dario Amodei published an essay the same day Claude Corps launched, acknowledging that AI-driven job displacement may be unavoidable and calling for universal basic income funded by AI company taxes. On the same day, Anthropic also announced a separate $200 million Economic Futures Research Fund focused on empirical research and policy trials around how AI reshapes labour markets. These are distinct initiatives, not one combined program.
The timing matters. AmeriCorps, the federal agency for volunteer service, was significantly cut by the Trump administration. CodePath CEO Michael Ellison had been thinking about redesigning that model to account for AI adoption. Claude Corps lands into that gap, funded privately rather than federally.
It also lands shortly after Anthropic closed a $65 billion funding round in May 2026 and filed confidentially for an IPO at a reported $965 billion valuation. The company has the capital to make a $150 million long-term commitment, and this is clearly meant to demonstrate something beyond goodwill. Anthropic has said it plans to rigorously measure the program’s impact, potentially open-source core infrastructure, and eventually replicate the model in other countries.
What to watch
Claude Corps is starting small by design. One hundred fellows, a handful of confirmed host organizations, a structured training program that hasn’t been road-tested at scale. The real question is what the measurement data shows after the first cohort completes. Anthropic has committed to publishing that, and Social Finance’s involvement as the evaluation lead gives it some credibility.
If the model works, the open-source infrastructure commitment could matter more than the fellowship itself. A replicable, documented playbook for embedding AI capacity inside mission-driven organizations would be genuinely useful, well beyond what any single fellowship cohort can achieve.
For now, the immediate opportunity is concrete: a paid fellowship for early-career workers who want to build real AI skills, and a fully funded AI resource for nonprofits that need hands-on capacity but don’t have the budget to hire it.
Full program details, fellow applications, and host information are at anthropic.com/claude-corps.