Biotech is about to do to healthcare what SaaS did to software in the 2010s: compress cycles, break silos, and move from “one big bet every 10 years” to a portfolio of fast, data-driven experiments.

2026 won’t be about one miracle drug or one unicorn IPO. The real story will be a handful of “tech-bio” startups that redesign how we discover, develop, and deliver therapies. The five companies below are worth following not only for what they build, but for what they signal about the next innovation wave.

As you read, pose yourself three questions:

• Where could this logic apply in my own industry?
• What will this change in the way value is created and captured?
• Am I building (or investing) for the last cycle or the next one?

Recursion: turning wet labs into data factories

Recursion (USA) is one of the clearest examples of the “AI-native” biotech model. They’re not just using machine learning as a tool — they’ve rebuilt drug discovery around it.

In simple terms, Recursion does three things:

• Automated labs run millions of experiments in parallel on cells.
• High-throughput imaging generates huge amounts of phenotypic data (how cells change under different conditions).
• Custom AI models spot patterns humans would never see and suggest new drug candidates.

Their real asset isn’t a single molecule; it’s a vertically integrated data engine: robotic lab infrastructure, proprietary datasets, and an AI stack running on top. That stack powers both internal programs and partnerships with big pharma.

What this reveals about the next wave:

• Biotech is becoming a platform business. The most valuable companies won’t live or die with one drug. They’ll run multiple discovery programs on the same data and automation backbone.
• Physical capacity is a constraint and a moat. Anyone can spin up cloud GPUs. Very few can design, finance, and operate automated labs that generate high-quality, standardized biological data at scale.
• Data feedback loops are the new IP. Every experiment feeds the model; every improved model designs better experiments. That compounding loop is hard to copy and gets more valuable over time.

If you run any science-heavy or R&D-heavy business, the Recursion playbook is a warning: if you’re still treating data as a by-product and not the core asset, someone else will build the platform around you.

Insitro: closing the loop between patient data and drug design

Insitro, founded by Daphne Koller (co-founder of Coursera), pushes the “tech-bio” idea one step further: instead of starting from biology alone, they start from patient data at scale.

Their approach, simplified:

• Use large patient datasets (genomics, clinical records, phenotypes) to identify patterns and disease mechanisms.
• Build highly controlled in vitro models (like induced pluripotent stem cells) that mimic those patient subgroups.
• Run automated experiments and apply machine learning to discover and optimize drug candidates for those specific subgroups.

Insitro isn’t just trying to find “a drug for disease X”. They’re trying to find “the right intervention for the right biological subtype of patients, backed by hard data before the first clinical trial.”

What this reveals about the next wave:

• Personalization will be baked in from day one. Instead of developing a broad drug and then discovering who it works for, discovery itself will be informed by real-world patient data.
• Data partnerships are strategic weapons. Access to biobanks, health systems, and genetic databases will make or break pipelines. Negotiating data access becomes as important as lab science.
• Clinical risk is attacked upfront with modeling. More of the uncertainty is shifted from the clinic (slow, expensive) to the lab and the cloud (fast, cheap).

For healthtech founders, Insitro illustrates a key shift: electronic health records, insurance claims and genomics are not just “nice to analyze” — they’re raw material for entirely new therapeutic businesses.

Mammoth Biosciences: CRISPR as a toolkit, not a buzzword

CRISPR hype peaked years ago, but Mammoth Biosciences (co-founded by CRISPR pioneer Jennifer Doudna) is one of the few players methodically turning that promise into diversified products.

Their core bet: CRISPR isn’t just one gene-editing tool. It’s a toolkit that can power multiple categories:

• Diagnostics – ultra-sensitive, rapid tests for infectious diseases and beyond, leveraging CRISPR’s ability to detect specific genetic sequences.
• Therapeutics – targeting genetic diseases by editing DNA or RNA precisely inside the body.
• Platform licensing – offering CRISPR toolsets to partners rather than owning every downstream application.

Instead of betting the company on a single flagship drug, Mammoth is building a modular CRISPR platform that can be plugged into different markets with different partners.

What this reveals about the next wave:

• “Deep tech” doesn’t have to mean “single-product risk”. If your core asset is a versatile technology, you can structure the business around multiple shots on goal.
• Platform + partnership beats vertical integration for frontier tools. Trying to own the entire stack from molecule to market can kill you on time and capital. Smart partnering accelerates use-cases and derisks the roadmap.
• Biology is getting more “programmable”. CRISPR tools increasingly look like APIs: you can target, edit, and fine-tune biological systems with growing precision.

For non-biotech founders, the lesson is simple: if you have a foundational technology, ask yourself whether you’re building a single product… or a platform others should build on.

Generate Biomedicines: generative AI for proteins, not pictures

If you’ve seen what generative AI can do with images and text, now imagine applying similar principles to proteins. That’s Generate Biomedicines’ game.

Instead of randomly screening nature’s proteins and hoping something works, Generate trains AI models to:

• Understand the relationship between protein sequence, 3D structure, and function.
• Design novel proteins with desired properties from scratch (stability, specificity, manufacturability).
• Iterate designs rapidly based on experimental feedback from their own labs.

The promise: move from “searching a massive biological haystack” to “asking the model to propose needles” that never existed in nature, but are tailored to specific therapeutic uses.

What this reveals about the next wave:

• AI won’t just optimize existing workflows; it will create new design spaces. We’re moving from discovery to invention of biological components.
• Compute becomes as central as chemistry. Capital expenditure isn’t just labs and clinical trials; it’s GPUs, engineering teams, and model infrastructure.
• Biotech talent stacks are hybrid. Tomorrow’s “drug discovery team” is part wet-lab, part ML research, part software engineering. Companies that can’t integrate these cultures will fall behind.

For leaders in any R&D-heavy sector, Generate is a preview: generative models transform how you explore your design space, whether that’s proteins, materials, or even industrial processes.

NewLimit: longevity and the business of aging

NewLimit is tackling a problem every country quietly worries about but few systems are set up to address: aging itself as a modifiable biological process.

Their focus: use epigenetic reprogramming (resetting how cells “read” their DNA without changing the DNA itself) to extend human healthspan – the number of years we live without chronic disease.

In practice, that looks like:

• Mapping how epigenetic markers change with age across different cell types.
• Identifying interventions that can push aged cells back toward a youthful state without triggering cancer or other side effects.
• Targeting specific tissues first (e.g. immune system, liver) where modest rejuvenation could have massive health impact.

NewLimit is at an earlier stage than the other companies on this list, but its ambition is representative of a broader trend: aging and longevity are moving from fringe to frontier.

What this reveals about the next wave:

• “Indications” will expand beyond traditional disease categories. Instead of treating one organ at a time, we’ll increasingly target systemic processes like aging, inflammation, and immune dysregulation.
• Longevity is a macro-economy driver, not a niche. The cost of aging populations hits pensions, productivity, healthcare budgets, and consumer markets. Any credible way to delay functional decline has trillion-dollar implications.
• The regulatory framework will be stress-tested. Aging isn’t officially recognized as a disease in most systems. Startups and regulators will need new endpoints, new trial designs, and maybe new categories.

If you work in insurance, pensions, HR, or policy and you’re not tracking the longevity space, you’re flying blind. Your business model quietly assumes that healthspan won’t change much; NewLimit and others are betting that assumption is wrong.

What these five startups signal about the next biotech wave

These companies operate in very different niches, but together they point to a coherent shift in how biotech will create value over the next decade.

Here are seven cross-cutting themes worth watching — and acting on.

1. From “drug companies” to “data platforms that ship drugs”

• Recursion and Insitro show that the core asset is not the pipeline alone, but the integrated stack: data, models, automated labs, and partnerships.
• Expect more “platform-first” companies that monetize via multiple programs, co-development deals, and eventually internal products.

Action point: If you’re building in biotech, define clearly: what’s your accumulating advantage? Is it a one-off molecule or a system that gets smarter and more valuable with every project?

2. AI is shifting risk earlier in the funnel

• AI won’t magically eliminate risk, but it can move more of it from clinical trials (slow, expensive) to modeling and lab experiments (faster, cheaper).
• Better target selection, better patient stratification, and in silico prediction of toxicity/efficacy will become standard expectations, not “innovation theatre.”

Action point: Investors: start asking founders how they’re reducing development risk structurally via data and models, not just “moving fast” in the lab.

3. Biology is becoming programmable infrastructure

• Mammoth and Generate illustrate a shift: genes, proteins, and cells are turning into design variables.
• Once biology becomes programmable, it stops being just “healthcare” and starts infiltrating manufacturing, materials, agriculture, and even computing.

Action point: Non-biotech CEOs should treat biotech like they treated cloud in 2010: a horizontal capability that will eventually touch every sector.

4. Partnerships are not optional; they are the go-to-market strategy

• All five companies rely heavily on partnerships: with big pharma, health systems, data owners, or manufacturing players.
• Regulation, capital intensity, and trust requirements make “go it alone” unrealistic in most cases.

Action point: Founders: design your partnership strategy as deliberately as your product roadmap. Who has the data, the distribution, the regulatory muscle you don’t?

5. The talent stack is hybrid — and hard to copy

• Teams now combine ML researchers, software engineers, biologists, chemists, clinicians, and regulatory experts.
• The real moat is not just hiring these people, but getting them to actually work as one system, not in silos.

Action point: Leaders: invest early in org design. Cross-functional squads, shared metrics, and compatible incentives beat “brilliant but disconnected” teams.

6. New business models will challenge regulation and reimbursement

• Personalized and longevity-focused interventions don’t fit cleanly into existing reimbursement and approval pathways.
• Some of the early revenue will come from adjacent models (self-pay, employer benefits, premium services) before insurers and regulators catch up.

Action point: Don’t just ask “can we build it?” Ask “who pays, under what code, with what evidence, and in what timeframe?” early in your planning.

7. The line between “healthcare” and “everything else” is blurring

• Longevity affects labor markets, pensions, and housing.
• Programmable biology affects food, fashion, packaging, and energy.
• Better healthspan changes how consumers spend, save, and work.

Action point: Whatever your sector, stress-test your 10–20 year strategy against a world where people live longer, healthier lives and biology is a design tool, not a constraint.

How to position yourself for this wave

You don’t need to be a CRISPR scientist or an AI researcher to play in this new landscape. But you do need to adjust how you think about opportunity, timing, and risk.

Three practical moves:

• Map overlaps with your existing strengths.

  Are you strong in data infrastructure, regulatory navigation, clinical operations, manufacturing, or enterprise sales to hospitals? Those are scarce skills in tech-bio — and they travel well.

• Build a learning pipeline, not just a watchlist.

  Don’t just “follow” Recursion or Generate on social media. Set up internal sessions every quarter where your team dissects one tech-bio company: business model, partnership strategy, talent stack, and what it could mean for your market.

• Experiment at the edges.

  If you’re a corporate, that might mean pilots with diagnostics or longevity startups. If you’re a founder, it might mean starting with a narrow, data-rich problem instead of trying to “solve aging” on day one.

Biotech is leaving the era of blockbuster serendipity and entering the era of engineered iteration. The five startups above are early proof points of that shift. If you understand what they’re really doing — not just the science, but the business design — you’ll be better equipped to decide where to build, where to invest, and, just as importantly, where not to waste the next ten years.