backing founders turning scientific & engineering breakthroughs into companies @KdT_Ventures
DC
- Tweets 1,146
- Following 1,698
- Followers 8,106
- Likes 3,551
Pinned Tweet
humbled to announce KdT IV, an oversubscribed $100M+ fund dedicated to supporting early-stage, science-driven companies.
at KdT, we believe the most iconic companies of our time are those that leverage scientific breakthroughs to build enduring businesses with significant societal impact. our conviction is that science-driven ventures require a focused, technical, and strategic team of investors to support founders from the earliest stages.
our mission remains the same: to partner with founders building at the intersection of technology and the physical layer of the world, across multiple verticals including therapeutics, digital health, chemicals, materials, and beyond.
capital alone is not enough; we are committed to providing technical expertise and strategic guidance to help navigate the long, windy road from scientific discovery to commercial success.
we are deeply grateful to our limited partners for their trust, co-investors for their collaboration, and most importantly, the founders who give us the privilege of participating in their journey. we have the best job in the world.
coverage in the comments below.
29
12
270
18,074
startup i want to help build/fund: infrastructure for the coming wave of incidental findings.
between consumer screening companies like Prenuvo/Neko and AI detecting more abnormalities in routine imaging, we’re going to find a lot more things we weren’t looking for.
“incidental finding” will become one of the most anxiety-provoking phrases in healthcare. “indeterminate 6mm lesion, recommend follow-up" will leave millions of patients asking whether this matters and what to do next.
they need a system: triage, risk stratification, surveillance, and routing to the right care. there’s a huge opportunity to combine better diagnostics with longitudinal follow-up, using what happens to each patient to improve risk prediction for the next.
who’s building this?
14
4
60
4,393
Patrick Malone, MD PhD retweeted
Despite pipelines filling with SOST x DKK1 bispecifics and RANKL x PTH fusion proteins there is actually a lot of bonafide innovation needed in osteoporosis and fracture prevention. Boston is a bone town!
4
17
705
when purchasing data, pharma wants to know how the data will help advance a particular program or make a specific development decision. for the AI labs, the question is whether the dataset improves a model’s ability to understand and predict biology across many problems. that creates demand for data/experiments who’s value is distributed across many future programs.
initially, the labs may buy whatever useful biological datasets exist (indiscriminate), but eventually demand will shift toward commissioning specific experiments that resolve gaps in a model’s understanding (discriminate).
6-12 months ago the large labs started buying bio data
this spend is starting to ramp very aggressively. the labs are somewhat indiscriminate buyers with enormous budgets. the therapeutics market has long had buyers of data (large and small pharma) but they've historically been the exact opposite of huge-budget and indiscriminate
this shift will distort the bio startup market in a bunch of ways. there will be a lot of short-termist behavior to try to get in front of this capital firehouse. but unlike the AI data labelers (mercor, et al) and the robotics labelers (mecka, et al), the bio startups serving up this data have a chance to emerge from this period with businesses that don't rely on selling data
the best bio startups will use these sudden resources the labs are dumping on them to supercharge their novel assay and work towards independently powerful models and eventually directly produce therapies
bio is the final frontier and its a very exciting time ahead for bio startups IMO
6
11
131
12,996
something interesting happening in AI drug discovery: Twist and GenScript may be getting paid to reduce their own future demand. both are increasingly selling datasets linking sequence to expression/binding/developability, not just DNA and protein. the molecules are becoming intermediate products.
these businesses have mostly made money when customers order molecules to test, but the data they're now selling trains models that get better at de novo binder design, so customers will need fewer designs and fewer synthesis rounds to find a good binder. building the model takes a ton of data/experiments (the spend we're now seeing), but using it in a mature workflow will take far fewer.
more programs could offset that. but as binders get cheaper to make, spending will shift from synthesis/design toward functional testing: does the molecule actually do what you want in cells and in vivo.
for Twist/GenScript and similar businesses, the question is whether customer spending follows them as the bottleneck moves from binder design to nominating drug candidates, and whether they can maintain attractive margins on that more complex work.
10
74
6,184
would love to see this extended to bio/chem/physics. lots of effort into using AI to generate new science, but figuring out which existing findings are solid enough to build on will arguably be more impactful.
Very interesting project: Living Science uses an AI agent to reproduce key findings from seminal papers in economics, document what holds up and extend the analysis with newer data. It's like a follow-up paper with robustness checks and extensions. New paper every week.
1
1
12
994
an underappreciated reason NIH funding is so durable: it's less about Congress believing in science, and more about the jobs it supports in lawmakers' home states and districts.
universities and academic medical centers are often among the largest local employers. a vote to cut NIH is a vote to cut jobs back home.
there's a lesson here for biotech. we spend a lot of time explaining why people should be excited about what excites us (cool science, new modalities). we should spend more time making sure our success shows up in things they already care about: jobs, their communities, affordable medicines, national security.
biotech's political future will depend as much on how widely it shares the benefits of innovation as on how much innovation it produces.
5
10
54
4,703
repeated, temporary U.S.–China trade deals create a paradox: ongoing uncertainty can convince manufacturers that they need alternatives to Chinese suppliers, while reducing their willingness to commit to one.
uneven progress in trade negotiations (tariff relief on $60 billion of select goods announced this morning, two-month extension of the broader truce, etc) creates a difficult timing problem for startups.
manufacturers urgently seek alternative suppliers when China restricts exports of critical materials. but temporary truces can give them a reason to wait and see whether access to Chinese supply improves before making a multiyear commitment to a newer, potentially more expensive domestic supplier.
startups need those commitments to finance production and achieve the scale needed to lower costs. customers can end up waiting for an alternative to become competitive while withholding the orders that would help it get there.
curious how other founders/investors have handled this, but seems logical to make it easy for customers to start small while urgency is high, eg paid qualification, followed by a recurring share of purchases that grows with demonstrated performance. agree upfront that those purchases continue when supply returns to normal, and secure minimum-volume commitments before expanding capacity.
1
8
621
every hard tech investor is staking out SpaceX and Anduril alumni (us included).
but the underrated founder is 48, works at GE, knows exactly what to build, and is done asking permission.
Made with AI
5
51
3,347
Patrick Malone, MD PhD retweeted
a preview of the future of AI for science: the list of discoveries worth investigating will grow much faster than our ability to understand them.
a growing backlog of unexplained AI discoveries will increasingly shape the questions human scientists pursue.
Claude has discovered a previously unknown enzyme system hidden in the DNA of bacteriophages. Beside the enzyme’s gene sits a long array of repeating DNA—a structure that looks somewhat similar to CRISPR.
We don’t yet understand what this system does, but only a handful of known systems share its features, and all of them are able to cut, copy, and paste DNA. Historically, the discovery of such programmable systems has helped revolutionize medicine. CRISPR, for instance, is now the foundation of genetic medicines. But it will take much more work to learn what this system does, and whether it can be put to similar use.
Read more: anthropic.com/news/claude-di…
5
3
24
2,842
surprised this hasn't generated more discussion, but openevidence is getting into the drug dev biz.
OE has a specific distribution advantage unlike any other drug developer: physicians actively seek its input at the moment they're deciding how to treat a patient.
this changes two constraints that often influence which drugs get developed:
1. how much complexity a physician has to remember. a drug can be valuable in a narrow patient subgroup / in context of specific biomarker / at a specific point in treatment. these details make it hard to commercialize. OE helps physicians navigate this clinical complexity and prescribe the drug.
2. how large a market needs to be to support commercial infrastructure. small / fragmented patient populations can be expensive to reach. OE could reach physicians through a platform they already use and spread engagement costs across many drugs, lowering the revenue each drug needs to justify commercialization.
with large pharma needing ever-larger commercial opportunities to justify investment, good to see a model that could make smaller indications and more complex therapies commercially viable.
8
4
83
10,249
ironically, the most likely cause of the ultimate failure of the China-to-West in-licensing model will be... the success of the China-to-West in-licensing model.
Western NewCos in-license assets from China, succeed, and exit. capital flows back to biotechs in China. those companies reinvest in global development and commercialization and become global pharmas. then they become the acquirers and licensees for the next generation of chinese biotechs.
the best assets stop leaving China, and Western NewCos get what's left.
china just released their latest 5 year plan for biotech. it is clear that china has ambitions of building globally dominant pharma companies, not just licensing assets to the West:
1. it explicitly supports building overseas sales networks, R&D centers, and manufacturing bases. these are the capabilities needed to commercialize internationally and capture product revenues directly.
2. china is addressing their domestic reimbursement bottleneck: faster hospital adoption of innovative drugs, a new pricing mechanism, and a push for commercial health insurance. chinese biotechs that can make money domestically can afford to keep global rights instead of selling them for upfront cash.
3. china is pushing regulatory alignment with global standards so chinese data can gain credibility with FDA/EMA. historically, a big part of what Western licensees paid for was the ability to "translate" a chinese asset into a package global regulators would trust. if that translation layer disappears, so does much of the reason to license.
4. it supports international multicenter trials and overseas registration. these capabilities will make western partners less essential to getting a drug through global development and approval.
5. the partnership language includes joint development AND joint commercialization. that creates a path to retaining more economics while building experience operating internationally.
6. china is targeting 50 pharma companies with annual revenue above RMB10 billion and at least 5 products with global annual sales above $1 billion. that scale will create more domestic acquirers, giving chinese biotechs non-western alternatives to exit/license. over time, more promising assets will stay within chinese pharma companies.
china-to-west licensing volumes will keep rising, but i suspect western partners’ share of the upside has already peaked. the west may keep licensing more chinese drugs while capturing less of the upside. the economics change when your partner no longer needs you to reach the market.
4
2
35
4,797
china just released their latest 5 year plan for biotech. it is clear that china has ambitions of building globally dominant pharma companies, not just licensing assets to the West:
1. it explicitly supports building overseas sales networks, R&D centers, and manufacturing bases. these are the capabilities needed to commercialize internationally and capture product revenues directly.
2. china is addressing their domestic reimbursement bottleneck: faster hospital adoption of innovative drugs, a new pricing mechanism, and a push for commercial health insurance. chinese biotechs that can make money domestically can afford to keep global rights instead of selling them for upfront cash.
3. china is pushing regulatory alignment with global standards so chinese data can gain credibility with FDA/EMA. historically, a big part of what Western licensees paid for was the ability to "translate" a chinese asset into a package global regulators would trust. if that translation layer disappears, so does much of the reason to license.
4. it supports international multicenter trials and overseas registration. these capabilities will make western partners less essential to getting a drug through global development and approval.
5. the partnership language includes joint development AND joint commercialization. that creates a path to retaining more economics while building experience operating internationally.
6. china is targeting 50 pharma companies with annual revenue above RMB10 billion and at least 5 products with global annual sales above $1 billion. that scale will create more domestic acquirers, giving chinese biotechs non-western alternatives to exit/license. over time, more promising assets will stay within chinese pharma companies.
china-to-west licensing volumes will keep rising, but i suspect western partners’ share of the upside has already peaked. the west may keep licensing more chinese drugs while capturing less of the upside. the economics change when your partner no longer needs you to reach the market.
11
32
159
20,102
US reindustrialization so far has mostly been a compute build out. the interesting question now is whether AI demand simply absorbs more power/equipment/labor/capital, or whether the bottlenecks it creates pulls forward enough new capacity that other industries benefit too.
the thing to watch is whether new capacity remains dedicated to AI, or whether it starts reducing costs and lead times for physical infra in entirely different industries. that would be the transition from AI capex boom to broader reindustrialization.
5
382
one of my favorite events of the year. this is a virtual workshop for scientists that want to build a neuroscience startup. we'll cover team building, development and fundraising strategies, navigating partnerships, and more. we've had hundreds of scientists from labs and companies across the world.
whether you're actively building or still ideating, check it out.
Join us for our 3rd annual Future Neuro Founders Program co-hosted by PsyMed Ventures and KdT Ventures!
Last year 400+ founders from 190 universities joined our annual virtual, complimentary program again for scientists who want to build.
1
12
1,291
as the US rebuilds domestic supply chains, it will be interesting to see to what extent today's materials markets reflect true demand vs how much they reflect constraints engineers have learned to work around.
this point from @packyM about battery cell companies moving downstream is a good example.
we usually size a market by asking existing customers what they'll buy, but the supplier may understand what becomes possible when its input gets dramatically cheaper or more available before the customer does.
a battery maker might realize cheaper cells make a storage project viable, then build and own the project itself. that demand would never have appeared in its customers' forecasts.
curious how many of the domestic materials companies being built today eventually move downstream and become their own most important customers.
2
5
994
important topics!
I'll be debating the benefits and costs of US biotech's "secrecy solution" with @davidycli + @patricksmalone, talking bio-risk with Caitlin Frazer, Ed You + Jen Crooks, and we've got @maxonwifi on stage with @RepMoolenaar + the FDA's Grace Graham + Karim Mikhail
7
1,448
if you're working at the intersection of bio and defense, including biosecurity, biomanfacturing, and human performance, this is worth checking out.
luma.com/kfgh2fvu
7
1
27
4,923
some thoughts on Claude’s protein design work, and implications for the future of AI drug discovery:
1. the optimal biotech company will get shockingly small. a couple great scientists will supervise hundreds of agents doing literature work, modeling, design, analysis, and experimental planning.
2. the future looks like a human+AI scientist orchestrating many specialist models, not one giant biological foundation model. the moat shifts toward the data and infra needed to know which models to use, how to route between them, and how to combine their outputs.
3. value may sit in the harness, not the model. orchestration, tool use, QC, and failure handling may matter as much as raw model capability.
4. the huge 16k+ word prompts are basically an attempt to make tacit scientific expertise executable. a major source of leverage will be converting the knowledge currently trapped in scientists’ heads into rules, examples, tools, and evals agents can use.
5. candidate ranking may matter more than candidate generation. generating 100k plausible molecules is only useful if you can identify the 20 worth testing.
6. surprisingly little of the prompt is actually about biology. most of it is model selection, compute allocation, controls, provenance, scoring, and managing an unsupervised scientific workflow.
7. hard targets are still hard. large, flexible, conformationally heterogeneous proteins remain difficult, so GPCRs, IDRs, dynamic complexes, etc. should remain relative white space even as easier protein design gets commoditized.
17
18
107
9,023
Patrick Malone, MD PhD retweeted
AI transformed coding, science is next
@ReactorfieldAI makes scientists and deep tech startups AI-native
Built by me and @berkbuilds to accelerate science with frontier AI
186
199
630
238,789