Ron Alfa
@Ronalfa
Followers
5K
Following
13K
Media
147
Statuses
3K
CEO/Co-founder, @NOETIK_AI Decoding Cancer | Ex-@RecursionPharma $RXRX | @StanfordMed MD-PhD | @PDSoros | Build the Future
San Francisco, CA
Joined February 2011
Announcing OCTO-VirtualCell (vc) a multi-scale, multimodal transformer trained to predict gene expression for a virtual cell in cellular contexts within patient tissue samples. Complete wth the Celleporter demo app to explore the data! 1/
12
84
420
A note from a friend nicely captures this moment: "A non-exclusive license for AI model access is a dream that I honestly didn't think was possible after some of the partner conversations I saw."
2
0
26
Grateful for shout out for @NOETIK_ai leadership in innovating business models in AI x bio. We also agree that the next frontier of AI is solving clinical / translational.
Today, we’re sharing Breyer Capital’s 2026 Healthcare and Life Sciences Predictions. As always, what an honor to work with and collaborate with @morgancheatham.
0
0
7
today, we’re sharing Breyer Capital’s 2026 Healthcare and Life Sciences Predictions. the 2026 forecast calls for the evolution of business models for clinical and bio AI, the maturation of infrastructure for prevention and the biology of aging, and a bioengineering renaissance
8
29
258
We kicked off the year with the FIRST foundation model licensing deal with GSK. @NOETIK_ai been quietly building the ML stack for human biology and clinical translation, accumulating more data than anyone. This is the most important problem in therapeutics, and the space is
EXCLUSIVE: GSK will pay millions to license AI models, rather than drug candidates, from Noetik $50M in upfront and near-term milestones, centered on accessing virtual cell models for lung and colon cancer. Talked to @Ronalfa in my latest: https://t.co/bu7a0qNxXG
0
1
24
Bio foundation models pretrained on massive human data.
Palantir CEO Alex Karp on the commercial lessons of AI on the battlefield: “What’s valuable is doing something no other enterprise in the world can do, pairing your tribal knowledge with efficiencies no one else has."
1
0
10
Going Founder Mode On Cancer https://t.co/pGxM34IrC3 Sid Sijbrandij is a generational founder. He founded and led GitLab, one of the largest remote companies in the world, from idea-stage startup to NASDAQ-listed software giant. But in 2022, a six centimeter mass growing from
39
139
850
This is an amazing piece by @ElliotHershberg on @sytses journey going Founder Mode on Cancer. An optimistic view for science, if less so for broad access to innovation. Have been there, we can do much more for patients, but we need to escape the confines of standards of care.
1
2
40
This was a fun panel this morning! Thanks to @turbine_ai for hosting and awesome presentation by @fabian_theis!
Today by @Turbine_AI: Virtual Biology Meetup Top AI x bio panel: @Ronalfa @fabian_theis Amaro Taylor-Weiner Discussing the future of virtual cells and more. Link below ⬇️
2
3
23
Meanwhile the US continues to massively under invest bio, and bio x ai. We are effectively bootstrapping a frontier bio ai lab and a biotech at the same time.
Apparently multiple Bio-x-ML people I know are getting credible ~$1M offers from Chinese state-affiliated recruiters. They mostly don’t want to move to China — but these offers are big multiples of USA prevailing salaries. Time for USA to wake up in this new deep tech rivalry…
5
11
71
We are not a virtual cell company. @NOETIK_ai builds bio foundation models. One of them is a world model that simulates biology through virtual cells. Clinical response prediction models have been in use for 6 months and coming next.
3
6
91
Glad we’ve been generating all this data for 3 years.
The AI healthcare gold rush is here
1
0
19
We are 12+ months ahead of what we’ve released publicly
demis recently said in an interview that AI labs are roughly 4-5 months ahead of what they release publicly, and in AI they might not have AGI yet, but even if they do, they'll probably focus first on scaling data centers and energy which is why we're seeing so much investment
1
0
23
Always just trying to bring together the best people to build as fast as humanly possible.
0
0
9
People keep asking. A big value driver in our Foundation Model deal is we 100% own the PRETRAINING data. It’s not public, it’s NOT owned by a university, etc
4
8
71
Also good to consider in bio data. There are plenty of large volume bio datasets that are not useful for training ai.
A data moat isn’t just about volume. It’s the compounding understanding of edge cases, user preferences, correction patterns.
1
0
23
Closing off JPM26 it can only be concluded that 2026 kicked off as the year for bio x ai. @NOETIK_ai <> @GSK $50M FM licensing DEAL @boltz_bio <> @a16z <> @pfizer Seed Launch + DEAL @chaidiscovery <> @EliLillyandCo huge FM licensing DEAL @proximabio <> @DCVC Seed Launch
4
14
126
💯
i try not to rag on tech guys for believing that applying ml to biology is really easy, because, tbh, i think the same about their field. i do kinda believe that if i were handed the reigns of power at a robotics or text or vision frontier lab, id figure things out by years end
0
0
7