Can artificial intelligence accelerate its own development?

The rapid progress of artificial intelligence has largely depended on human researchers designing models, writing software, running experiments, and interpreting results.

A new generation of startups is exploring whether AI systems can take on more of that research process.

Mirendil has raised a reported $1 billion Series A financing round led by Kleiner Perkins and Clearlake Capital, according to reporting from The Wall Street Journal on November 13, 2026.

The financing reportedly values the company at approximately $5 billion.

The investment follows an earlier $200 million seed round, reflecting significant investor interest in technologies intended to accelerate AI development.

Introducing Lor

Mirendil is developing a platform called Lor, designed to coordinate large numbers of AI agents that write software and conduct technical experiments.

The concept involves using autonomous systems to assist with parts of the research cycle that traditionally require substantial human effort.

For example, AI agents might generate code, test possible approaches, analyze experimental outputs, and help researchers identify promising directions.

This approach could potentially allow smaller research teams to investigate more ideas simultaneously.

However, the practical effectiveness of these systems depends on the quality of the experiments, the reliability of their conclusions, and the extent of human supervision.

The promise of faster research

Mirendil reports that one of its researchers trained an initial pretrained model in approximately three and a half months using $10 million in computing resources.

The company presents this as evidence of a more efficient research approach.

That account is company-reported. It does not independently establish that Mirendil can consistently match the performance of leading frontier AI laboratories or that its methods will generalize across research problems.

Training costs and timelines also depend heavily on model size, objectives, data, and evaluation standards.

Beyond artificial intelligence

Investors see potential applications in fields such as robotics and drug discovery.

In principle, automated research systems could assist scientists in evaluating computational hypotheses, exploring design alternatives, or organizing complex technical experiments.

But scientific research requires more than generating plausible results.

Experimental reproducibility, data quality, rigorous validation, and domain expertise remain essential.

In healthcare and biotechnology, computational findings must be assessed through appropriate laboratory and clinical research before they can support medical conclusions.

What this means for women in STEM

Research automation could influence career pathways across computer science, engineering, and applied science.

As AI systems become capable of assisting with more technical tasks, researchers may spend greater portions of their time defining problems, evaluating evidence, designing experiments, and supervising computational workflows.

For women pursuing research and technology careers, this creates both opportunities and uncertainties.

AI tools may reduce barriers to experimentation for smaller teams and emerging founders. At the same time, access to computing resources, specialized training, mentorship, and funding will continue to influence who can participate.

The challenge is ensuring that advances in research automation expand participation rather than concentrate technical capabilities among a small number of organizations.

The importance of oversight

Systems capable of autonomously conducting extended research workflows introduce important questions about reliability and control.

An AI agent may generate code that appears functional but contains errors. It may select inappropriate evaluation methods or produce conclusions unsupported by available evidence.

When many agents operate simultaneously, those problems can become harder to detect.

Researchers and organizations therefore need clear evaluation standards, monitoring, and human accountability.

Mirendil’s founders have expressed confidence that their technology can be supervised safely, but the broader scientific and technical questions surrounding autonomous AI research remain unresolved.

Looking ahead

Mirendil’s financing illustrates how investors are backing the possibility that AI can accelerate not only commercial workflows but the development of future technologies.

The next tests will involve independently demonstrated research performance, repeatability, responsible deployment, and practical applications.

For women in STEM, the opportunity is to participate in defining how these systems are built, evaluated, and used.

The future of scientific discovery will depend not only on what AI systems can accomplish, but on the expertise and judgment of the people directing them.

Source: The Wall Street Journal — Mirendil financing and research automation, November 13, 2026.

Looking for Your Next Opportunity in STEM?

Explore emerging technologies, startup developments, and career opportunities through Fems in STEM and Career Connections.

This article is independent editorial coverage. Financing and technical performance figures are based on published reporting and company statements. Future applications remain prospective.


Discover more from Fems in STEM

Subscribe to get the latest posts sent to your email.

Share.

Promoting Women in Science, Technology, Engineering, and Mathematics

Leave A Reply

Discover more from Fems in STEM

Subscribe now to keep reading and get access to the full archive.

Continue reading