Prompt: Review the startups listed on Altitude Lab and identify your top three picks.
My assumptions: Not prioritizing any particular sector, funding stage, or investment thesis. No access to pitch decks, just public information.
Top 3 picks: and stood out. would've been third, but it was acquired by Pathos, so I suggest instead.
Leash Bio
Founded: 2021
Capital: $9.3M Seed round in April 2024, led by SpringTide Ventures.
Technology: Leash Bio is an AI-driven drug discovery platform using a proprietary high-throughput screening system to generate biochemical data by screening small molecule binding against proteins. They’ve generated over 17 billion protein-compound interaction data points—orders of magnitude more than what’s available in PubChem. This dataset powers a cyclical machine learning loop that rapidly designs and refines small molecules against protein targets.
Market and Differentiation: Leash operates in the small molecule drug discovery market, a core part of the $50B+ annual biopharma R&D spend. It differentiates by building a specialized experimental system to directly measure molecular interactions—creating a dataset tailored to train ML models for binding prediction. This contrasts with Recursion (phenotypic imaging) and Isomorphic Labs (structure prediction), which train on different biological layers. Leash’s experimental-first approach may offer a more grounded training signal, but needs to be validated through downstream results like lead identification.
Risk and Considerations: Leash is operating in a fast-moving field with well-capitalized players building autonomous labs for large-scale data generation. The key question is how they plan to scale not just higher-quality data, but the most relevant data for the ML task at hand.
Team: Strong | Six of the eleven team members are ex-Recursion, bringing relevant experience in AI-driven drug discovery and a proven track record of working together. CEO and CTO . Overall Take: Strong team and thoughtful TechBio approach. The key question is how well it translates into high-performance, translatable hits—especially in a field with increasingly capable players. Worth engaging to dig into the model (performance, compute infrastructure) and the experimental system—how binding is measured (e.g., enrichment vs. affinity), technical limitations, and whether they’re capturing the right experimental signals to drive therapeutic relevance and lead success.
Sethera
Founded: 2024
Capital: $3M Seed round in January 2025.
Technology: Sethera is building a peptide discovery platform that combines enzymatic cross-linking with AI-guided design to generate polymacrocyclic peptides—complex, multi-ring structures with improved stability and binding potential. The platform enables rapid generation of diverse and structurally novel peptide libraries for hit discovery and is backed by a filed patent and four academic publications.
Market and Differentiation: Sethera operates in the multi-billion-dollar peptide therapeutics market, where demand is growing for more stable and tunable drug formats. Most other platforms use traditional or automated chemical synthesis. Sethera uses enzymatic cross-linking to generate structurally novel polymacrocyclic peptides with more control and structural diversity. Combined with AI-guided design, the platform offers a modular, scalable approach to hit discovery.
Risk and Considerations: Operates in a competitive space with other macrocycle-focused platforms. No disclosed pipeline of leads, and the key question is whether the structural novelty translates into meaningful functional or therapeutic advantages.
Team: Good starting team composed of a faculty founder and his former PhD student, who have worked closely together. Team is now five. The recent addition of Robert Langer to the board brings high-profile visibility and strategic credibility. CEO and CSO . Overall Take: Strong early team and differentiated platform aimed at unlocking new peptide structural design space, but success hinges on whether that novelty translates into meaningful functional breakthroughs. Worth engaging to review data, any early leads, and their discovery strategy (indication focus, internal vs. partner-led).
Vistim Labs
Founded: 2021
Capital: $932K in seed funding across two rounds (2023 & 2024)
Technology: Vistim Labs is building a software-only AI tool that analyzes EEG data to estimate key dementia-related biomarkers—like amyloid, tau, and cognitive function. It runs on existing EEG hardware, offering a non-invasive, lower-cost alternative to PET scans for earlier, more accessible diagnosis of Alzheimer’s and related conditions.
Market and Differentiation: The EEG device market is moderate (~$1.7B) but part of a much larger and growing neurodiagnostics space. Traditional diagnostic tools for dementia—like PET scans and CSF analysis—are invasive, expensive, and hard to scale. Vistim takes a lighter approach: software-only AI that runs on standard EEG machines. Most competitors bundle AI with proprietary hardware; Vistim’s SaaS model is cheaper, more scalable, and easier to integrate into existing clinical workflows.
Risk and Considerations: Validated in a HITLAB pilot, but still early. No regulatory clearance or clinical adoption yet. Key questions include model accuracy and regulatory pathway—but most importantly, market adaptability, especially given the competitive space.
Team: Team is credible but has room to bring in strong operators. Advisory support includes neurodiagnostics veteran Frank Zanow (CEO of ANT Neuro, largest EEG manufacturer). CEO Overall Take: Software-only AI running on existing EEG infrastructure—simple, scalable, and differentiated, but the key question is market potential and value capture. Worth engaging to assess clinical performance, regulatory path, and scalability strategy—including pricing, adoption drivers, and commercial model.
Analysis Table (See final column for decision rationale)
Market Size & Opportunity