Etched shipped its first rack and raised $700M, led by our customer @JaneStreetGroup with participation from @kleinerperkins, @sequoia, @a16z, @BainCapVC, and @blackstone.
We've raised $700M at a $21B valuation from Jane Street, Kleiner Perkins, Sequoia, A16Z, Peter Thiel, BCV, and Blackstone.
We're also excited to share that we've shipped our first rack to Jane Street.
17
15
308
38,458
The round also included @stripesco_, @psumvc, @PrimaryVC, Tiger Global, and @neo. It's always a pleasure working with @sonyatweetybird, @abhishekm1636, and all of our new investors
2
22
1,805
@BSchech and @patrick_oshag were some of our earliest believers, and I'm humbled they have chosen to double down in every single subsequent round.
4
2
36
1,635
I'm so excited to go to work every day and think about how we can make our customers happy. We spent lots of time on our low-voltage inference and cluster-scale memory technologies (and V2/V3/... for this tech), but just as much time on the basics
1
1
32
1,163
A special shoutout to the people at SK Hynix, who have been unreasonably helpful for the entire lifetime of Etched.
@seokhee4 personally gave advice when Etched was still located in a dorm room, gave us HBM parts for use in pre-silicon CoWoS tuning, and brought a team of engineers to our office during bringup to help us tune eye training.
They did all of these things before we had a working chip, and I'm humbled and grateful they believed in us from day 0. I couldn't be more excited for them to be joining our Series C financing.
We’ve raised $300M in Series C funding at a $10.3B valuation from Sequoia, Andreessen Horowitz, Jane Street, Argo, and SK Hynix.
Our mission is to run the world's inference. This round accelerates production of our inference clusters.
We've opened an 80,000-sqft, 10-MW facility 15 minutes from our office to expedite production and prototyping.
10
16
445
74,257
I'm excited to announce we have raised $300M led by @sequoia , with participation from @a16z , Jane Street, Diffusion, and SK Hynix.
We’ve raised $300M in Series C funding at a $10.3B valuation from Sequoia, Andreessen Horowitz, Jane Street, Argo, and SK Hynix.
Our mission is to run the world's inference. This round accelerates production of our inference clusters.
We've opened an 80,000-sqft, 10-MW facility 15 minutes from our office to expedite production and prototyping.
37
16
624
66,347
@robertwachen and I have gotten to know Sonya and Abhishek while building Etched. Sonya saw the future and invested in OpenAI in 2021 - before agents, reasoning, coding, or ChatGPT had been dreamed up.
But it wasn't just them - almost the entire partnership came by our office in person. Inference will be the biggest market of all time, and we're excited to work with Sequoia to scale production as rapidly as we can.
1
39
2,576
They are joined by @RaghuRaghuram, @DavidGeorge83, @sarahdingwang and @shangdaxu at a16z. I first met Raghu at our office two years ago (before he was at a16z) when our company was <20 people and our technology was just a whitepaper.
Raghu is the greatest infrastructure strategist I have ever met, and that's who I want on my team as we build the world's biggest inference clusters.
2
30
1,669
Seeing our A0 silicon and first-gen racks come to life has been the experience of a lifetime. It takes a village - I am so proud to be working alongside @robertwachen @czhu1729 @saptadeep_pal and our world-class team to do what many thought was impossible.
We think our Low-Voltage Inference (LVI) and Cluster-Scale Memory (CSM) tech will help bring down the cost of inference for the world, all while using less power. But there’s still much to do on next-generation products - if you like solving hard problems, join us!
We're coming out of stealth.
We've built our first racks after a successful A0 tapeout, $1B+ in customer contracts, and $800m raised.
Early customer tests show us achieving SOTA throughput, latency, and power efficiency on inference workloads.
Our first racks ship this summer.
20
23
230
36,059
Ben and Asher are some of the smartest people I know, and the team they've assembled is world class as well. So excited to see what they build!
Announcing Flapping Airplanes!
We’ve raised $180M from GV, Sequoia, and Index to assemble a new guard in AI: one that imagines a world where models can think at human level without ingesting half the internet.
1
2
45
14,804
Gavin Uberti retweeted
Mercor (@mercor) scaled from $1-500M in revenue run rate in the last 17 months, making us the fastest growing company of all time. Our growth is accelerating. We averaged 11% week over week growth in July, 18% WoW growth in August, and 19% WoW growth in September.
One trend driving this meteoric growth: the Economy is Becoming an RL Environment Machine.
Reinforcement learning is becoming so effective that agents can hillclimb any benchmark, but humans need to define the rewards to automate everything.
While everyone fears job loss, we’re creating a new category of knowledge work faster than any other time in history. The future of work will converge on training agents.
We're paying out over $1M / day to people in our marketplace and hiring experts rapidly across nearly every domain: software engineers, doctors, lawyers, consultants, bankers, and many more.
145
163
1,431
626,405
happy llama day to all those who celebrate
Today is the start of a new era of natively multimodal AI innovation.
Today, we’re introducing the first Llama 4 models: Llama 4 Scout and Llama 4 Maverick — our most advanced models yet and the best in their class for multimodality.
Llama 4 Scout
• 17B-active-parameter model with 16 experts.
• Industry-leading context window of 10M tokens.
• Outperforms Gemma 3, Gemini 2.0 Flash-Lite and Mistral 3.1 across a broad range of widely accepted benchmarks.
Llama 4 Maverick
• 17B-active-parameter model with 128 experts.
• Best-in-class image grounding with the ability to align user prompts with relevant visual concepts and anchor model responses to regions in the image.
• Outperforms GPT-4o and Gemini 2.0 Flash across a broad range of widely accepted benchmarks.
• Achieves comparable results to DeepSeek v3 on reasoning and coding — at half the active parameters.
• Unparalleled performance-to-cost ratio with a chat version scoring ELO of 1417 on LMArena.
These models are our best yet thanks to distillation from Llama 4 Behemoth, our most powerful model yet. Llama 4 Behemoth is still in training and is currently seeing results that outperform GPT-4.5, Claude Sonnet 3.7, and Gemini 2.0 Pro on STEM-focused benchmarks. We’re excited to share more details about it even while it’s still in flight.
Read more about the first Llama 4 models, including training and benchmarks ➡️ go.fb.me/gmjohs
Download Llama 4 ➡️ go.fb.me/bwwhe9
1
1
19
9,039
Excited to be part of the founding class of Magnificent Grants, and congratulations to all the winners!
Check out the new 2024 Cohort of Magnificent Grants. And happy to report that going forward, the fellowship application process will be on a rolling basis, launching a nomination system... substack.com/home/post/p-157…
2
7
5,634
Hackathon idea - nearly speculative decoding. As of v0.7.3, @vllm_project supports Deepseek R1's Multi-Token Prediction module, letting you "skip" a token generation if the multi-token prediction guessed it correctly in advance.
But what if you accepted almost correct guesses?
We're excited to partner with @Cognition_Labs @mercor @CoreWeave and @AnthropicAI to host an inference-time compute hackathon, featuring >$60K in cash prizes and >1 exaflop of free compute.
2
1
4
5,093
Rather than insist the speculative decoding distribution match exactly, if you are OK with, say, 99% of the distribution being recovered (e.g. according to some metric like KL Divergence), you could accept guessed tokens more frequently
1
1,925