Pinned Tweet
Excited to introduce Onyx and announce our $8M seed round led by Dimension to build a virtual immune system!
Onyx has always been a story of connection. The first connection was to my cofounder @amlakhan, a deeply curious technologist who I’ve been fortunate to call a friend for 19 years now. I've always said that working together was a cornerstone of starting Onyx, and it truly has been one of the best decisions I've ever made.
Connection is also at the heart of what Onyx does. We believe AI is key to making biological complexity tractable, ultimately transforming life sciences into an engineering discipline. But that requires the right data. AI performance compounds with context, and richer biological context comes from more connected data.
In biology, such context is often missing: molecular data, clinical history, and change over time are typically captured in isolated fragments. Onyx provides the substrate that connects them, pairing genomics, transcriptomics, proteomics, and metabolomics with rich clinical histories. By studying these layers together in people over time, our data platform traces how real human biology changes in response to infection, inflammation, and treatment. Seeing how these trajectories differ across people brings us closer to understanding what drives disease—and to developing precision medicines that work better for real people.
Yet connecting data is only possible because of the people who have connected with our mission. Nearly a thousand members have already joined our research studies. They are more than research participants: they are difference-makers transforming how we understand some of humanity’s most complex and chronic diseases. I always say that our work is for people, but the truth is that people make our work possible. That connection comes full circle, and at Onyx, people will always come first.
Finally, as we raised our oversubscribed round, we sought more than just business partners. We wanted investors we felt a true connection to—not merely sources of capital, but co-conspirators in creating meaningful change. To that end, we feel incredibly fortunate to have joined forces with @_DimensionCap, @haystackvc, @refactor, @standard_vc, @robotventures, and many more.
Now back to building!
Today we're introducing Onyx, a research collective building a virtual immune system: a model of how real people respond to infection, inflammation, and treatment.
We're launching with an $8M seed round led by Dimension Capital.
onyx.bio/news/introducing-on…
Roger Chen :: retweeted
Onyx
AI will learn to reason about bio and medicine, yes.
That will come from startups creatively assembling new multidisciplinary datasets paid for by AI labs and pharma, yes.
But a great dataset alone does not manifest a great company.
That requires a unique point of view on terminal value.
Onyx has that and it stems from a question not frequently asked in AI x Bio: What role does the patient play here exactly?
The Onyx approach takes inspiration from its roots in experimental science, real world ML, cryptography, and internet community stewardship.
On a personal note - Roger has been a close collaborator and dear friend dating back to our days as PhD students (even though he went to Berkeley, a shit school).
Humans like Roger and Amit are who we want shepherding us into this brave new world.
We're thrilled to be a part of the adventure.
Excited to introduce Onyx and announce our $8M seed round led by Dimension to build a virtual immune system!
Onyx has always been a story of connection. The first connection was to my cofounder @amlakhan, a deeply curious technologist who I’ve been fortunate to call a friend for 19 years now. I've always said that working together was a cornerstone of starting Onyx, and it truly has been one of the best decisions I've ever made.
Connection is also at the heart of what Onyx does. We believe AI is key to making biological complexity tractable, ultimately transforming life sciences into an engineering discipline. But that requires the right data. AI performance compounds with context, and richer biological context comes from more connected data.
In biology, such context is often missing: molecular data, clinical history, and change over time are typically captured in isolated fragments. Onyx provides the substrate that connects them, pairing genomics, transcriptomics, proteomics, and metabolomics with rich clinical histories. By studying these layers together in people over time, our data platform traces how real human biology changes in response to infection, inflammation, and treatment. Seeing how these trajectories differ across people brings us closer to understanding what drives disease—and to developing precision medicines that work better for real people.
Yet connecting data is only possible because of the people who have connected with our mission. Nearly a thousand members have already joined our research studies. They are more than research participants: they are difference-makers transforming how we understand some of humanity’s most complex and chronic diseases. I always say that our work is for people, but the truth is that people make our work possible. That connection comes full circle, and at Onyx, people will always come first.
Finally, as we raised our oversubscribed round, we sought more than just business partners. We wanted investors we felt a true connection to—not merely sources of capital, but co-conspirators in creating meaningful change. To that end, we feel incredibly fortunate to have joined forces with @_DimensionCap, @haystackvc, @refactor, @standard_vc, @robotventures, and many more.
Now back to building!
Roger Chen :: retweeted
@rgrchen and I started @onyxbio to understand how real immune systems evolve in real people. By generating multiomic datasets from our members over time, we aim to help cross the trial failure chasm.
Shoutout to @_DimensionCap @haystackvc @refactor and more for believing in us!
Today we're introducing Onyx, a research collective building a virtual immune system: a model of how real people respond to infection, inflammation, and treatment.
We're launching with an $8M seed round led by Dimension Capital.
onyx.bio/news/introducing-on…
Roger Chen :: retweeted
Excited to announce our partnership with @rgrchen, @amlakhan and the team at @onyxbio!
We believe that precision immunology is one of the most interesting directions of new medicines and that multi-modal, longitudinal datasets from real patients are needed to power this field.
Onyx is building a unified data platform for immunology - combining clinical records, multi-omics, sample banking, and patient reported outcomes. This platform helps researchers to understand how immune diseases progress, why patients respond differently to treatment, and uncover new ideas for medicines. At the limit, these data will power the buildout of a entirely virtual immune model to simulate human immunology with high precision.
We have known Roger for over a decade and are excited to reunite! Nearly a thousand patients are already enrolled onto the platform and looking forward to much more to come!
onyx.bio/
Roger Chen :: retweeted
Your health plan should get better year-round. Not just once a year.
Today we're launching @Decent on @ycombinator. We find the best health plan for your company, then keep making it better all year.
Roger Chen :: retweeted
Every successful marketplace has winners and losers
But some participants make it more of a dark forest than a chessboard
We find some direct evidence for this — and some simple math to make these marketplaces more robust to rogue inference providers!
Roger Chen :: retweeted
Announcing Dimension III
An $800M fund to team with the world’s most ambitious entrepreneurs pioneering the frontiers of science and compute
Pumped to continue to partner with our existing and future founders!
Roger Chen :: retweeted
Today we're unveiling Dimension III: a $800M fund to partner with the leading entrepreneurs across the frontiers of compute (silicon through infrastructure) & science (biotech through physics).
To our founders, past, present, and future: now is your time. LFG!
Roger Chen :: retweeted
Thrilled to announce Dimension III, a fresh $800M fund which takes the firm up to $1.6B AUM.
Even since we launched the firm a few years ago, the overlapping exponential curves of science, AI, infrastructure, and hardware have continued to scale at staggering rates. We are more certain than ever that a research-led, long-term oriented investment practice is the right capital partner to ambitious founders in this era. If you are building something technically ambitious and possibly a little crazy, you are in good company here.
While the fund, our team, and our ambitions continue to grow, we remain steadfast on our mission to partner with founders building at the frontiers of science and compute.
Excited for the future ahead, let’s go build it!
top notch investors and even better human beings
[brief update]
we have a new set of funds for @haystackvc
announcing haystack 8 plus more, but it's really about my teammates @aashaysanghvi_ & @DivyaDhulipala
thank you to all of our frequent ecosystem co-conspirators, our LPs, our co-investors, our founders. i'm very grateful 🙏 and this is the most fun role in the world!
semilshah.com/2026/04/26/new…
Roger Chen :: retweeted
Today we're announcing Sardine's $70M Series C 🚀
Money moves at the speed of light. Crime moves at the speed of AI. Between these forces, trust is eroding faster than ever.
We built Sardine to move faster than the threat - and it's working.
LGND 👀
While we haven't officially launched, can't help but be excited to see this great artcle in the @WSJ featuring @lgnd_ai (lgnd.io) and the opportunity for AI applied to earth observation data. Big stuff ahead! wsj.com/tech/ai/artificial-i…
Only because you gave me the opportunity 🙏
Depends on who you’re sending out.
@rgrchen found Fetch ($300M exit) and Expanse ($1B exit) during his short time with us.
Roger Chen :: retweeted
An important systems bottleneck when working with LLMs is model loading times, but if you get the details right, you can speed up standard implementations by around 20x (over 10 minutes down to around 35 seconds for Llama-2-70B).
There are a few bottlenecks to numbers to think through (numbers for a p4de instance)
⏩ S3 -> disk bandwidth (10GB/s) [but it takes a lot of extra work to saturate this]
⏩ disk -> CPU memory (theoretically 16GB/s but practically 5GB/s)
⏩ CPU memory -> GPU memory (16TB/s for PCIe4)
Adding these together for a 130GB model gives around 52 seconds S3 -> GPU memory. But you can do much better.
In this blog post, we got it down to around 35 seconds (compared to over 10 minutes for standard implementations).
Fun fact: you can reverse all of this to build fast model checkpointing (GPU memory -> S3).
How long does it take to download Llama2 70B? On the 4x 25 Gbps NICs that aws.p4de's have, it should take ~10s. Yet in production we've observed much higher times, which makes autoscaling less responsive + more expensive. This blog post shows how we've reduced download & init time from 4m25s to 20s, using techniques such as streaming S3->CPU->GPU and multiple TCP streams. anyscale.com/blog/loading-ll…
Excellent systems work by @bonecheng!
Roger Chen :: retweeted
The Rhind Mathematical Papyrus is the oldest manuscript written in algebra and trigonometry, dating back to 3,550 years ago.
It shows that the Egyptians used first-order equations, geometric series and a second-order algebraic equation, related to the Pythagorean theorem a² + b² = c²
It also describes how to obtain an approximation of π accurate to within less than 1% and one of the earliest attempts at squaring the circle.
Roger Chen :: retweeted
One funny thing about DAO governance is that it makes it hard to manage assets held by a collective (why vote to sell the asset you're voting with?) — but for many collectives, asset management is more than pure yield optimization
Can we do better with a return to futarchy?
Today we're proud to announce Aera, the world’s first autonomous, data-driven treasury management protocol. Aera will provide DAOs with a one-stop solution for managing their token reserves efficiently and transparently.
aera.mirror.xyz/QMSOGh75-_n2…
