Big news! We just raised $1.2 billion for DFJ Growth Fund V — our largest fund yet — to back the next wave of category-defining companies 🚀
📝 @dfjgrowth blog: https://t.co/uEitelucSR
📈 @FortuneMagazine@agarfinks: https://t.co/dm4kZQPwYw
Silico, the platform for ambitious AI research, is publicly available today.
AI is advancing fast. The tools to understand it need to advance even faster. Silico lets you interpret and train your models at frontier scale.
Learn more + get access 🧵
Thrilled that @Qualcomm has completed its acquisition of @Modular!
This move accelerates Modular's mission - providing unifying AI SW for HW from all vendors, going beyond today's fragmented ecosystem. Datacenters are disaggregated and heterogenous, and deserve high quality SW!
New @dfjgrowth founder story with @clattner_llvm Co-Founder & CEO of @Modular. Great read on Chris’s journey building foundational developer tools including LLVM and Swift to now Modular. Congrats to Chris, @iamtimdavis and the team on the recent acquisition by Qualcomm!
https://t.co/dgSgRSEhWp
A huge congratulations to @Modular on the agreement to be acquired by @Qualcomm! From our very first meetings with @clattner_llvm and @iamtimdavis we had conviction this was the right, one-of-a-kind team with the bold and strategic vision to build a unified compute platform for AI. This combination only accelerates their journey to unlock a more open, hardware-independent future.
From our entire team at @dfjgrowth: congratulations on the milestone and we can't wait to see what is ahead! 🔥
I'm excited to share that Qualcomm is acquiring Modular: this will accelerate our path to unifying accelerated compute with an open platform. This will also mark a new era in open software development for Qualcomm. 👇
Open weight models are more than just an open source story.
They are the AI era equivalent to cloud repatriation: a shift from abstraction-first to workload-specific infrastructure.
In the early days of cloud, startups no longer had to buy servers, forecast capacity, rack machines, wire networking, or turn uncertainty into capex. Infrastructure became an API. That changed the cost and pace of company formation.
Frontier model APIs are playing a similar role today. They let teams build without managing GPUs, inference infrastructure, model serving, or reliability. For most startups, that is the right starting point.
But as AI products scale, the abstraction tax becomes visible: token costs, latency, rate limits, model behavior changes, data control, and vendor dependency. That does not mean every company should run open weight models. It means some workloads eventually become predictable, expensive, sensitive, or strategic enough to justify moving down the stack.
Where is this all headed? The future AI stack is likely to be hybrid and workload placement will be paramount.
https://t.co/1Igbl1odJc
Congratulations to @elonmusk, @Gwynne_Shotwell and the incredible @SpaceX team on their historic IPO. An amazing story of vision, grit, and execution built over the past 24 years.
It has been an honor for @DFJgrowth to be part of the journey since 2009 and watch one of the great company-building stories of our time unfold.
To the stars!💫
We are building solar + battery power plants, off-grid to start. 100% renewable. 99.9% availability.
Grateful to our early backer @dfjgrowth. Read more about Ambrosia from DFJ: https://t.co/dbHbMqIEIU
We are building solar + battery power plants, off-grid to start. 100% renewable. 99.9% availability.
Grateful to our early backer @dfjgrowth. Read more about Ambrosia from DFJ: https://t.co/dbHbMqIEIU
We are building solar + battery power plants, off-grid to start. 100% renewable. 99.9% availability.
Grateful to our early backer @dfjgrowth. Read more about Ambrosia from DFJ: https://t.co/dbHbMqIEIU
Venture capital is on the cusp of what could be its greatest returns ever.
This year’s Forbes Midas List—the 25th anniversary edition—reflects the staggering rise of SpaceX, OpenAI and Anthropic, three private companies whose valuations are reshaping the venture landscape and driving a record 25 newcomers onto the ranking.
Read more: https://t.co/6s7cg0qAD2 (Photos: Guerin Blask for Forbes) #ForbesMidas
I keep coming back to the thesis that AI may move faster and broader than most people expect. Not just because the models keep getting smarter, but because AI uniquely compounds on itself and those with the right levers have an accelerating flywheel.
We see it playing out in real time where better models → more usage → more data and feedback → faster product iteration → better models. This loop matters and is propelling the industry with not just new products, but real economic value.
I wrote up a few thoughts on why AI’s compounding dynamics may make this platform shift steeper, broader, and more power-law driven than prior technology cycles.
Let's debate.
https://t.co/Hoq4Q2aHGL
We achieved state-of-the-art performance in predicting which of 4.2 million genetic variants cause diseases by interpreting a genomics model, in a new preprint with @MayoClinic.
We're now releasing an open source database for all variants in the NIH's clinvar database. 🧵(1/8)