Fearless Engineering: The CS-4 is our first engineering documentary. Through lab footage and conversations with the engineers behind each breakthrough, the film follows the creation of the CS-4, our latest wafer-scale AI system. 0:00 introducing the engineering team 1:09 Wafer-scale history 2:28 Powering the wafer 4:00 The cooling problem 4:18 Designing the cold plate 5:11 Holes in the wafer 6:33 the alignment problem 7:47 The CS-4 PRD 8:35 Designing for deployability and serviceability 11:01 Removing the motherboard 11:25 Removing internal cooling loops 11:41 inventing the power rack 12:35 A one-person lab deployment 14:03 Backpack: 50% fewer components 14:55 Connecting the compute backpack to cooling manifold 17:58 shifting Cerebras’s focus from training to inference 19:09 Building the Cerebras Cloud 20:01 CS-3 vs. CS-4 performance 20:49 3D Stacked DRAM

CS4 uses the same wafer as CS3. So where does the extra speed come from? You asked on Twitter and Discord. Join Angela Yeung, SVP of product and @MilksandMatcha, on the behind-the-scenes on the engineering behind CS4: connecting wafers, recirculating cooling water, and getting compute deployed at scale. 1:08 Where does the extra speed come from? 2:33 Connecting wafers for 10-trillion-parameter models 4:15 What breaks first when you scale? 6:07 Breaking down cooling 7:30 Easy CS4 deployment

We built Money Agent, a personal finance assistant powered by @Alibaba_Qwen 3.8 27B on Cerebras. Now you can turn a home-buying question into a real-time conversation, then into a financial goal, faster.

Connecting the dies and working around defects were fundamental challenges for wafer-scale computing. Making it work also came down to power, cooling, and reliability. @seanlie explains how those lessons are shaping @cerebras’ approach to stacking DRAM. From @seanlie’s conversation with @swyx on @latentspacepod.

Qwen3.8-27B is now live at Cerebras speed. The dense, open-weight model from @Alibaba_Qwen scores 34 on the Artificial Analysis Intelligence Index—making it comparable to models such as GPT-5.6 Luna, DeepseekV4 Pro, and Claude Sonnet 4.6.

Running GPT-5.6 Sol Ultrafast on Cerebras, @MatthewBerman saw his workflow flip: fewer parallel agents, same productivity, lower mental load. When tokens stop being the bottleneck, the rest of the stack becomes the next layer to optimize. Faster inference opens that door.

It is interesting to see the industry move from 12Hi back toward 8Hi HBM just as DRAM stacking is accelerating. Going vertical does not make the fundamental constraints of memory disappear. As stacks get taller, thermals, power, yield, packaging, and reliability become increasingly difficult architectural problems. HBM stacks DRAM vertically beside compute. The next frontier (as seen at hot chips) is stacking DRAM directly on compute, turning the interface from the edge of a chip into its entire surface area.

Today, we announced a new 165 MW data center in Mikkeli, Finland, in partnership with Compute Nordic Finland. Construction of the first 50 MW is already under way. Two things make this project special to me: First, we designed the Mikkeli data center to give back. The facility is designed to capture the heat our computers generate so the community can use it, and the water that cools our systems gets recirculated instead of drawn from the town's supply. Second, it creates long-term jobs. An estimated €1.0–1.7 billion will be invested into the region and it will create up to 250 permanent technical jobs in South Savo. Thank you to Compute Nordic Finland, the City of Mikkeli, and our network of Finnish partners for helping bring this to life.

a great bring your friends to work day @swyx @vibhuuuus @seanlie we have labs/ hardware testing/ equipment/ manufacturing/ racks in our office in south bay friends and connections always welcome to come by for a hardware tour

Ten years ago, we started @cerebras around an approach many believed was impossible. As a computer architect, it is hard for me to imagine a more exciting time. Model releases are accelerating, and hardware tapeout is compressing from multi-year roadmaps to annual launches. Hot Chips is my favorite conference, and it’s where I launched Cerebras 7 years ago. This year’s conference was especially exciting, and so much innovation was shared. I am watching the industry recreate itself: SRAM is mainstream, DRAM is moving into the third dimension, networks are being fundamentally redesigned, and AI is helping design and program the chips themselves. The industry has never moved faster and some of the hardest architectural questions are still wide open.

where in the world is cafe compute this week? 🇯🇵 Tokyo: 🇲🇾 Kuala Lumpur: 🇮🇳 Hyderabad: big thanks to our @cerebras ambassadors for hosting @asahiXXXXXXXXX @hazlijohar @jayant99acharya

At Cafe Compute Fal.Con, we’re giving one lucky winner the opportunity to experience Cerebras-like speed from behind the wheel of a supercar at SPEEDVEGAS. The winner take a supercar onto a purpose-built Las Vegas racetrack with a professional instructor riding alongside. Ferrari, Lamborghini, Porsche, McLaren, Aston Martin, Mercedes-AMG, Nissan GT-R, Lotus. Which car will you drive? RSVP: Come see us at Booth # 1039 at Fal.Con with @CrowdStrike *** No purchase necessary. One winner will be selected at random at Cafe Compute on Sept 1. Must be in attendance at time of drawing to win. Must be 18+ and have a valid driver's license. Prize is subject to vehicle and reservation availability. Participation may require acceptance of SPEEDVEGAS’s terms and liability waiver. Cerebras is not responsible for personal injury, death, property damage, or other loss arising from participation in or use of the prize.

Last week, Cerebras CTO @seanlie announced CS-4, 30x faster than the GPU. This week at #hotchips2026#, Cerebras announced CS-5, another step function faster than anything we've seen before with up to 10,000 tokens/sec/user on models such as Gemma 4 31B and gpt-oss-120B. Seana and I will be doing an AMA on Cerebras, CS4/5/6, Hot Chips announcements. Let us know what questions you have.

Congratulations to our partners at @OpenAI today. Taking a new chip to real lab results at great performance is a huge achievement, and improving speed across the board pushes our whole industry forward. We look forward to having Jalapeno and the next-gen Cerebras solution together in production in 2027, driving OpenAI’s comprehensive fast and ultrafast inference product portfolio to continue evolving what's possible with fast AI.