Stock Markets March 17, 2026

AI Model Linking H&E Slides to Spatial Proteomics Spurs Sell-Off in Bruker and 10x Genomics

Wolfe Research calls the market reaction disproportionate given limited clinical use and the early-stage nature of the evidence

By Jordan Park BRKR TXG
AI Model Linking H&E Slides to Spatial Proteomics Spurs Sell-Off in Bruker and 10x Genomics
BRKR TXG

Shares of Bruker Corporation and 10x Genomics fell after a tweet by Microsoft’s CEO highlighted GigaTIME, an AI model developed with Providence Health and the University of Washington that predicts multiplex immunofluorescence results from H&E pathology slides. Wolfe Research says the decline reflects investor concern that AI could reduce demand for spatial proteomics in clinical settings, but notes the underlying study is not new and that the tools from the affected companies are not widely adopted clinically.

Key Points

  • GigaTIME, an AI model by Microsoft, Providence Health, and the University of Washington, maps H&E slides to multiplex immunofluorescence results and was highlighted publicly by Microsoft's CEO.
  • Wolfe Research links the recent sell-off in Bruker (BRKR) and 10x Genomics (TXG) to investor fears of reduced demand for spatial proteomics, but calls the reaction disproportionate since these tools are not widely used clinically.
  • The study supporting GigaTIME trained on 40 million cells across 21 proteins and tested on 14,256 patients, but analysts note the protein panel may be too limited to substitute for traditional multiplex assays.

Shares of Bruker Corporation (NASDAQ:BRKR) and 10x Genomics (NASDAQ:TXG) dropped on Monday following public attention to GigaTIME, an artificial intelligence model created by Microsoft in collaboration with Providence Health and the University of Washington that predicts multiplex immunofluorescence test outputs from hematoxylin and eosin (H&E) pathology slides.


Market reaction and analyst perspective

Wolfe Research attributed the sell-off to investor concern that clinical demand for spatial proteomics - the category that includes multiplex immunofluorescence assays - could shrink if AI were to supplant some clinical applications. The research firm, however, cautioned that the price moves appeared disproportionate to the underlying evidence.

Specifically, Wolfe Research noted two points that temper the perceived threat to companies that supply spatialomics tools. First, Bruker and 10x Genomics' spatialomics products are not yet widely integrated into clinical workflows. Second, the paper underlying GigaTIME was published in Cell several months prior, meaning the key findings were already in the public domain.


What GigaTIME does - and the scope of the study

The GigaTIME approach applies multimodal AI to translate H&E pathology results into spatial proteomics readouts. According to the published study, the model was trained on 40 million cells using paired H&E slides and multiplex immunofluorescence results across 21 proteins, and then evaluated on 14,256 patients. The set of markers included PDL-1, a protein cited for its relevance to immunotherapy.

Wolfe Research emphasized limitations within the study, noting that assessing only 21 proteins may not be sufficient to demonstrate that the AI approach could replace traditional multiplex methods, which often survey many more markers. The firm described the development as early stage.


Clinical context reported in the study

The Cell paper framed multiplex immunofluorescence as a tool used to characterize the tumor immune microenvironment, a factor that influences cancer progression and response to immunotherapy. The study also described multiplex immunofluorescence as expensive and slow, which is part of the motivation for exploring AI-based translation from routine H&E slides.

Taken together, the market reaction reflects concern about potential downstream effects of AI on niche clinical testing markets, while analysts underline that the technology and evidence remain in early stages and that current clinical adoption of spatialomics platforms is limited.

Risks

  • Market risk - investor concern that AI translation of H&E slides could diminish demand for spatial proteomics equipment and reagents, affecting companies that supply those tools.
  • Scientific and clinical uncertainty - the study assessed 21 proteins, which Wolfe Research says may be insufficient compared with conventional approaches that evaluate many more markers; the approach is characterized as early stage.
  • Adoption risk - spatialomics platforms from Bruker and 10x Genomics are not widely used in clinical applications, creating uncertainty about how advances in AI will impact near-term commercial demand.

More from Stock Markets

Golden Dome Program Budget Rises to $185 Billion as Three Primes Join Effort Mar 17, 2026 U.S. Boosts Golden Dome Budget to $185 Billion, Adds Top Defense Primes Mar 17, 2026 EGA to Reroute Aluminium Flows Through Oman’s Sohar Port Amid Strait of Hormuz Disruption Mar 17, 2026 Brussels Stocks Close Higher as BEL 20 Advances 0.74% Mar 17, 2026 Paris Stocks Close Higher; CAC 40 Advances on Strength in Energy, Utilities and Financials Mar 17, 2026