Tesorai
4 étapes1 pièces5 Sources
Tesorai, a San Diego-based AI startup founded by former Verily (Alphabet) leaders, announced a $5.3 million seed round on November 18, 2026, bringing its total funding to $8 million. The company is building an infrastructure layer that processes raw experimental data from mass spectrometry and other lab instruments, rather than relying solely on summarized findings in published papers. Backed by investors including Perplexity Fund, IQT, and South Central Ventures, Tesorai aims to make historical and new biological data searchable and comparable to accelerate drug discovery.
The core of Tesorai’s approach is reprocessing existing data. In August 2026, the team demonstrated this capability by recovering p-tau217, a key Alzheimer’s biomarker, from the raw files of five public brain studies where it had not been previously reported. This finding was posted as a preprint and has not yet undergone peer review. Additionally, researchers reported in an August preprint that adding protein data allowed a small single-cell AI model to match or outperform larger models trained only on RNA data.
While competitors like Isomorphic Labs raise billions to design drugs using AI, Tesorai positions itself as complementary by solving the bottleneck of data accessibility. The disagreement in the field centers on whether generating new data (as seen in the recent $1.8 billion effort by Biohub, Google DeepMind, Meta, and others) or harmonizing existing raw data is more critical for advancing AI in biology. Tesorai argues that much of the necessary evidence already exists but is unusable due to fragmentation and poor annotation.
Readers should watch for the peer-review status of the Alzheimer’s marker study, further expansion of the platform beyond proteins to other molecular data types, and new commercial partnerships with biopharmaceutical companies.
Comment on en est arrivé là
August 10, 2026: JAMA Neurology publishes an Alzheimer's panel; Tesorai later recovers p-tau217 from raw data of five related studies.
techfundingnews.com ↗