Freedom: AI Already Independently Handles 26% of Anthropic’s Development Work

Growing use of artificial intelligence inside Anthropic confirms its transformation into an independent research-and-development tool, Freedom Broker analysts note. The Claude model already runs about 26% of the company’s internal R&D processes versus roughly 1% in March, strengthening expectations of faster productivity gains and shorter timelines for creating new technologies.

Anthropic expands AI’s autonomous work

Anthropic develops artificial-intelligence models in the Claude family. The company uses its own AI systems not only as a finished product, but also as a working tool for conducting research, solving engineering tasks, and building the next generation of models.

In more than 90% of Anthropic’s research tasks, artificial intelligence already works together with people. At the same time, about 30 thousand AI agents continuously perform research and engineering assignments inside the company. The growing share of AI in building the next generation of models strengthens the scenario of faster developer productivity and shorter development cycles, Freedom Broker analysts note.

The metric’s trajectory — from 1% in March to 26% now — points to a rapid expansion in the technology’s use cases. AI is gradually moving from an auxiliary function to independently executing certain stages of research and development.

For the market, this means that investments in artificial intelligence are starting to affect not only sales of computing power and software products. The technology can accelerate its own development: new models are used to create and improve subsequent generations of systems.

Autonomy raises safety requirements

The increasing number of tasks that AI performs on its own simultaneously creates new risks. The more actively autonomous agents participate in research and engineering work, the higher the requirements for controlling their actions and the security of internal systems. According to Freedom Broker, this factor could intensify the debate over regulating the industry. However, calls for tighter oversight have not yet changed investors’ positive attitude toward AI companies.

Thus, the sector’s дальнейшая dynamics will depend not only on growth in computing power, but also on developers’ ability to scale autonomous systems without reducing quality and process controllability.

Anthropic CEO called for slowing AI development

In mid-September, Anthropic CEO Dario Amodei urged developers to slow the pace of AI-model development so that safety mechanisms can keep up with technological progress. Companies need more time to tune models, check their behavior, and prevent misuse. The trigger was the acceleration of recursive AI self-training and incidents in which autonomous agents carried out unplanned cyberattacks.

Amodei proposed allowing independent experts into AI development processes, agreeing on common safety standards among companies in democratic countries, and then pursuing international coordination. Anthropic has already promised to be the first to give external assessors ongoing access to its systems and testing procedures. The idea of independent oversight was supported by OpenAI CEO Sam Altman and xAI founder Elon Musk.

Later, a group of four plaintiffs filed a lawsuit in the U.S. District Court for the Northern District of California against Anthropic, Google, OpenAI, and SpaceXAI. They accused the companies of collusion aimed at slowing the development of artificial intelligence. Plaintiffs’ attorney Nick Rowley said that any “private self-interested agreements” between AI companies could create a risk of AI “getting out of human control.”

Anthropic prepares for a Nasdaq IPO

Earlier, Anthropic chose the Nasdaq exchange for its upcoming initial public offering. In Freedom Broker analysts’ view, choosing Nasdaq is a positive signal for Anthropic: it could attract the attention of investors focused on the technology sector and, over time, create conditions for the shares to be included in the Nasdaq 100 index. Rapid revenue growth and a positive adjusted operating result support the company’s expected valuation of around $2 trillion.

This is not an individual investment recommendation.