Anthropic projects that its Claude models will account for 80-90% of frontier model training in 2027, according to figures the company has shared about its AI systems' growing role in developing next-generation models. The claim hinges on “frontier models,” the industry term for the most capable AI systems at the leading edge of capability — the same class of systems the company says its own models are increasingly building.

From zero to a quarter in eight months

According to the company, Claude's contribution to next-generation model development stood at 0% at the beginning of 2026. By April, that figure had climbed to 3%. It then accelerated sharply, with Anthropic indicating that Claude's share reached 26% by August 2026.

The company has framed the 80-90% target for 2027 as a projection rather than a settled outcome.

The coding precedent

The trajectory echoes Claude's earlier performance in software development. Shortly after a major model release in August 2025, Claude went from producing a negligible share of code to 50% of Anthropic's internal code. The company then set a goal of 80-90% by the end of that year.

A $100 million bet on human operators

On October 2, 2026, Anthropic announced the Claude Frontier Academy, a $100 million initiative aimed at training 10,000 “Frontier Deployed Engineers” by the end of 2027.

The program follows a residency-style curriculum, comparable to how doctors learn on the job under supervision, with the first certifications expected by early 2027. Anthropic is working with major firms including Accenture and Morgan Stanley on the effort, with those partnerships designed to give trainees real-world placements rather than classroom hours alone.

The compute side of the equation

On the infrastructure front, Anthropic is targeting approximately 9.5 gigawatts of compute capacity by the end of 2027. For scale, a single large nuclear power reactor typically generates about one gigawatt. To reach that goal, it is relying on partnerships with cloud providers including AWS and Google, among others.

Who Anthropic is, and why it frames things this way

Anthropic was founded by former OpenAI researchers and positions itself as an AI safety-focused laboratory. Its projections will also draw scrutiny over how the company measures and reports these figures: percentages such as 26% depend heavily on how “contribution” is defined, and outsiders cannot easily verify internal metrics.

What this means

For enterprises, the Frontier Academy may be the most practical element of the announcement. A pipeline of 10,000 certified engineers could make it easier for large firms to deploy Claude at scale, assuming the program meets its targets and the first certifications arrive in early 2027 as expected.

Key things to watch include whether Anthropic publishes updated contribution figures in the coming months, how it defines what counts as Claude “leading” training, and whether the compute buildout stays on schedule.