TechBio Survey Shows a Mixed Industry Outlook, Claude Gaining Ground, More Machine Learning

January 9, 2026

By Allison Proffitt 

January 9, 2026 | The Bits in Bio Slack-first community has released its fourth annual State of TechBio survey results. Year over year findings include a jump in Claude usage, an increase in daily machine learning work, and a serious cooling of the “Very bullish" sentiment across the broader biotech market.  

The Bits in Bio community is a word-of-mouth, Slack-first community of TechBio professionals, mostly in industry in the United States. The entire community was invited to participate in the fourth annual State of TechBio survey through the Slack channel. The survey was sponsored by Gunderson, Ropes & Gray, and AWS, with technical support provided by Bunsen, Nitro Bio, and Wunderdogs. 

More than 200 Bits in Bio members participated, with 65% representing biotech or pharma and the rest from IT, academia, or basic research. Diagnostics made up just over 10% of the respondent industry affiliations. About a third of the survey respondents are computational biologists or bioinformaticians with data scientists, software engineers, ML or data engineers being the other most common role descriptions.  

The industry outlook has cooled considerably since last year and the community remains cautious on the broader biotech market. Most of the community was split between “somewhat bullish” and “somewhat bearish”, but the "very bullish" camp dropped significantly to 9% (-11 YoY), while "somewhat bearish" continued to gain ground (27%, +12 YoY).  

Tools & Tricks 

Claude and OpenAI tools were used with about equal regularity by more than half of the respondents, though Claude enjoyed some significant gains year over year. Proprietary or in-house models were used by only 44% of the respondents, a sharp decline from last year.  

Only about a third of the community has worked with lab automation, but another third are exploring it with data acquisition, high throughput screening, liquid handling and sample prep being the most popular use cases.  

Illumina is the most used equipment vendor, but Thermo Fisher, Agilent, Hamilton, and Oxford Nanopore were not far behind. (Interestingly for sequencing, PacBio was only used by 6% of this community.)  

More than 40% of respondents answered that they primarily write code within a biotech or pharma and 67% reported that their work already involves machine learning. Python and Shell were top programming languages 

The number of respondents who reported that they “never” do ML work dropped to 14%, nearly half of last year’s percentage, though about half of the respondents were split between “daily” and “rarely” doing ML work. Scikit-Learn was the machine learning tool (frameworks, libraries, models, etc) most frequently used by the community at 67%, followed by Pytorch (64%), OpenAI (41%), HuggingFace (38%), and Anthropic (32%).  

Half of the respondents said they personally do data analysis at work and they visualize data with Matplotlib, Plotly, Excel, and Ggplot. While most of this community does not use an electronic lab notebook or LIMS, those that do use Benchling (33%).  

Explore the full survey data online.