Anthropic, the AI company behind the Claude models, has run an autonomous AI protein design study — and the lab equipment company that measured the results is now talking about it.
According to a Business Wire announcement carried by Google News and republished via TMCnet on August 24, 2026, Carterra highlighted the role of its high-throughput surface plasmon resonance platforms, or HT-SPR, in what it describes as one of Anthropic's autonomous AI protein design studies. Carterra describes itself as a leading provider of HT-SPR platforms for antibody and small molecule drug discovery.
A short translation for non-biologists: proteins do their work by sticking to other molecules. A drug antibody, for example, is useful only if it binds tightly and selectively to its target. Surface plasmon resonance is a lab technique for measuring exactly how strongly and how quickly two molecules bind. "High-throughput" means running very many of those measurements in parallel rather than one at a time.
That throughput is the reason the pairing matters. An AI system that designs proteins can propose candidate sequences far faster than a traditional lab can test them, so measurement becomes the bottleneck. A platform that screens large numbers of binders quickly is what lets a design-test loop run at machine speed — and "autonomous" implies the loop runs with limited human intervention at each step.
The available source material is a vendor announcement, and it does not disclose the study's targets, its scale, its results, or the terms of any arrangement between the two companies. Anthropic's own account of the work is not included in these items, so treat the framing as Carterra's.
Why it matters: drug discovery has long been limited by how fast molecules can be tested, not imagined, and an AI lab publicly running an autonomous design-and-measure loop signals that the bottleneck is starting to move.