Carterra HT-SPR Helps Turn Anthropic’s Autonomous Protein Design Into Measurable Results

Artificial intelligence is rapidly changing protein design and drug discovery, allowing researchers to generate and rank large numbers of potential protein binders in a fraction of the time required by conventional approaches. But as AI protein design accelerates, experimental validation is emerging as a critical bottleneck.

A recent study from Anthropic illustrates the challenge. Published August 18, 2026, the study reported that Claude models autonomously conducted de novo protein binder design campaigns against 15 challenging targets. The models researched each target, selected epitopes, ran open source protein design tools and produced ranked designs without human input into individual design decisions.

The resulting protein designs were experimentally analyzed by Twist Bioscience and Adaptyv Bio. Across the two laboratories, researchers evaluated 1,320 protein binders using Carterra high throughput surface plasmon resonance, or HT-SPR, platforms.

The work is among the largest published wet lab validations of AI designed proteins to date. More importantly, it highlights the growing importance of high throughput protein binding analysis as AI systems become capable of generating increasingly large numbers of candidates for drug discovery.

AI Protein Design Is Moving Faster Than Experimental Testing

Protein binder design is an important stage of early drug discovery. Researchers need to identify molecules that bind to biological targets with sufficient affinity and selectivity before candidates can progress through subsequent stages of characterization and development.

AI is beginning to compress this design process significantly.

In the Anthropic study, Claude generated functional protein binders against multiple challenging targets in days. The company reported hit rates exceeding those achieved by prior methods, demonstrating the potential for AI models to explore protein design problems at a scale and speed that would be difficult for human researchers to match.

The resulting increase in candidate volume creates a new challenge. Generating thousands of potential binders is useful only if laboratories can test those molecules efficiently.

Traditional surface plasmon resonance workflows typically measure interactions sequentially. At the scale of an AI driven protein design campaign, that approach can create a significant experimental bottleneck.

High throughput SPR provides an alternative by enabling researchers to measure many molecular interactions in parallel while generating quantitative data on both binding affinity and binding kinetics.

"This study shows the enormous potential of AI and Lab in the Loop automation to accelerate drug discovery, when paired with high throughput analysis platforms," said Josh Eckman, CEO and co founder of Carterra. "An AI system generated thousands of novel binders in a matter of days. Two independent labs experimentally validated the protein designs in a few weeks. Carterra was built for this moment, when measurement has to keep up with design."

High Throughput SPR Enables Protein Binding Analysis at Scale

The scale of the Anthropic study demonstrates why throughput is becoming increasingly important for AI enabled drug discovery.

Evaluating more than 1,300 protein binders against 15 targets using conventional SPR approaches could require many months of instrument time and significantly larger quantities of purified protein.

Carterra's high throughput SPR technology is designed to address these constraints through an array based approach. Its proprietary flow printing microfluidics technology enables researchers to create arrays of molecular interactions that can be analyzed using real time SPR.

According to Carterra, its platforms can deliver up to 100 times the throughput of traditional label free platforms while using a fraction of the sample. The company says its array based approach can reduce sample consumption to as little as 1% of that required by traditional systems.

For protein design campaigns, the benefit is not simply the number of interactions that can be measured. High throughput analysis allows researchers to examine multiple candidates and targets within the same experimental framework, making direct comparisons possible.

In the Anthropic study, researchers compared Claude's strongest RBX1 binder with the winner of an earlier open protein design competition. Under identical experimental conditions, the Claude designed binder recorded an affinity of 3.9 nM compared with 45 nM for the previous winner.

The researchers also tested human, mouse and cynomolgus versions of a target in parallel. This generated species cross reactivity data as part of the primary experiment, providing information relevant to subsequent preclinical research without requiring a separate campaign.

Binding Affinity and Kinetics Are Critical AI Drug Discovery Data

The value of high throughput SPR extends beyond confirming whether an AI generated protein binder interacts with its target.

SPR provides quantitative information about binding affinity and kinetics, including how quickly a molecule associates with a target and how quickly the interaction dissociates.

These measurements can provide a more detailed picture of a candidate's performance than a simple binding or no binding result.

That distinction is particularly relevant to AI drug discovery. Models require high quality experimental data to assess which design characteristics are associated with successful binding and to improve subsequent rounds of molecular design.

"The bottleneck in AI drug discovery is the experimental validation of all those molecules that the AI models come up with," said Julian Englert, CEO and co founder of Adaptyv Bio. "For large campaigns like this one, high throughput SPR is the best method to get real binding kinetics data, which is why we're using Carterra SPR in our automated lab. That's what generates the data to train the AI models and improve the next round of designs."

The result is an increasingly integrated workflow in which computational protein design, automated experimentation and data analysis operate as a continuous discovery cycle.

Lab in the Loop Connects AI Protein Design With Experimental Data

The emerging Lab in the Loop approach aims to connect AI driven molecular design with automated laboratory experimentation.

Instead of treating computational design and laboratory testing as separate stages, Lab in the Loop workflows allow AI systems to generate candidates, automated laboratories to test them and experimental results to inform the next round of computational design.

This creates a design build test learn cycle that can potentially reduce the time required to move from an initial computational hypothesis to experimentally validated candidates.

High throughput binding analysis is an important component of that process because it determines how quickly large numbers of AI generated molecules can be characterized.

A July 2026 report from Leerink Partners identified Carterra as an enabling technology in the emerging Lab in the Loop market for AI driven antibody discovery and highlighted binding affinity measurement as a central component of the workflow.

Carterra's Role in AI Driven Drug Discovery

Carterra develops high throughput surface plasmon resonance platforms for antibody and small molecule screening and characterization.

Its LSA, LSA-XT, Carterra Ultra and Carterra Vega platforms combine proprietary microfluidics with real time high throughput SPR and analysis software.

The company has more than two decades of experience in label free biosensor technology, with its platforms used by pharmaceutical and biotechnology organizations for biologics and small molecule research.

The Anthropic study illustrates how that technology can fit into an increasingly automated drug discovery workflow.

AI models can now generate large numbers of novel protein designs in days. The ability to experimentally evaluate those designs at comparable scale is becoming an equally important part of the discovery process.

As AI protein design continues to advance, the competitive advantage may increasingly depend not only on generating better molecules, but on how quickly researchers can obtain high quality experimental data, identify promising candidates and feed those results back into the next design cycle.

For AI driven drug discovery, computational power and experimental throughput are therefore becoming increasingly interconnected.

The protein binders described in the Anthropic study are research stage molecules characterized for binding only. They have not been evaluated for therapeutic activity, and the findings should not be interpreted as claims regarding a drug candidate. The experimental results described are those reported by Anthropic and the participating research organizations. For research use only. Not for use in diagnostic procedures.

A July 2026 report from Leerink Partners identified Carterra as an enabling technology in the emerging Lab in the Loop market for AI driven antibody discovery and highlighted binding affinity measurement as a key component of the workflow.

Carterra's platforms combine proprietary flow printing microfluidics with real time array based HT-SPR. The company's LSA, LSA-XT, Carterra Ultra and Carterra Vega platforms are designed for high throughput screening and characterization of antibodies and small molecules.

For drug discovery organizations adopting increasingly automated design workflows, the ability to generate binding data at scale can become an important consideration in determining how quickly a campaign can progress.

The Anthropic study demonstrates the point. Claude generated and ranked protein designs across a broad set of targets. Twist Bioscience and Adaptyv Bio then applied experimental workflows capable of evaluating a large number of those designs within a matter of weeks.

As AI systems take on more of the work involved in molecular design, the bottleneck in discovery may increasingly move downstream. The ability to measure, interpret and feed experimental results back into the design process will become an increasingly important part of the overall platform.

For companies developing AI enabled approaches to drug discovery, the lesson is clear: computational design and experimental capacity need to scale together.

The protein binders described in the Anthropic study are research stage molecules characterized for binding only. They have not been evaluated for therapeutic activity, and the findings should not be interpreted as claims regarding a drug candidate. The results described are those reported by Anthropic and the participating research organizations. For research use only. Not for use in diagnostic procedures.