Anthropic has committed $100 million to the Claude Frontier Academy, a training program that aims to produce 10,000 “frontier deployed engineers” by the end of 2027. CNBC reported the details on 2 October, citing the company’s press release. Nobody has completed the program yet, so every number here is a target, not a result.
The trainees are not Anthropic staff. They come from companies in the Claude Partner Network, the group of businesses that help other organizations put Claude to work. Engineers from Accenture, Morgan Stanley, and Novo Nordisk are in the first groups, alongside several consulting firms. CNBC did not name the other firms.
The curriculum borrows from medicine. Every trainee opens with a simulated enterprise deployment, which is scored. Passing it earns a place in a residency modeled on how hospitals teach doctors: instructors guide, and trainees work real-style cases. A second evaluation and a credential come at the end. Anthropic expects the first certified engineers in early 2027.
Shambhavi Shambhavi, who leads strategy and operations on Anthropic’s partnerships side, told CNBC that customers and partners keep asking for people who pair business context with top-level AI fluency. The demand is visible elsewhere. Draup, a labor-market data provider, gave CNBC figures showing a surge this year in finance listings for forward deployed engineers and similar AI roles.
The $100 million figure comes with gaps. CNBC’s account does not say over what period the money will be spent, or whether it covers instructors, infrastructure, trainee time, or something else. If every dollar went to training and the target were met, the cost would be about $10,000 per engineer. That is an illustration, not a disclosed budget.
The program also has a commercial logic that the announcement does not spell out. Anthropic sells access to Claude, and the people who decide how much of it a bank or drugmaker uses are often the engineers wiring it into daily operations. Training thousands of them, on Claude specifically, produces a workforce that already knows the product and carries a credential tied to it. That does not make the program insincere. A shortage of skilled deployers is a genuine bottleneck, and Anthropic benefits from easing it.
Timing adds context. CNBC noted that Anthropic is widely expected to list on the public market before the year ends, and that a valuation near $2 trillion has reportedly been floated. It also pointed to Reuters, which, working from a leaked copy of the prospectus, said Anthropic took in just under $4.6 billion last year while losing more than $8 billion at the operating line. Against those figures, $100 million is a modest line item. Whether it is a deliberate signal to investors is not something the reporting addresses, and it would be speculation to claim so.
The open questions are practical ones. The credential’s value depends on whether employers outside the partner network recognize it. It also depends on how hard the scored assessments turn out to be, and Anthropic has not published pass rates because no cohort has finished. Training is also only part of the picture: a certified engineer still needs a company willing to let them deploy.
For consulting firms and enterprises already inside the partner network, the near-term decision is whether to enroll staff in the first cohorts. Those still outside should expect the badge to start appearing in client requirements once the first certifications land in early 2027.
Originally reported by CNBC (Caleigh Keating), published 2 October 2026.
