BTC $64,629.98 -0.00%
ETH $1,830.16 +0.88%
BNB $593.08 +0.11%
XRP $0.9610 +0.47%
SOL $72.84 +2.25%
TRX $0.3062 +1.24%
DOGE $0.0590 +0.18%
ADA $0.1535 +0.66%
BCH $208.77 +0.25%
LINK $9.18 +0.35%
HYPE $69.78 +5.06%
AAVE $112.83 +2.65%
SUI $0.7242 +0.33%
XLM $0.1378 -0.68%
ZEC $922.25 +3.79%
AAPL $352.90 -1.22%
AMZN $257.64 -2.23%
GOOGL $283.98 -3.14%
MSFT $517.58 -1.02%
META $622.99 -1.72%
NVDA $157.68 -2.27%
TSLA $379.02 -1.61%
SNDK $814.69 -3.24%
INTC $54.66 -4.77%
SPCX $137.93 -1.73%
MU $723.80 -3.25%
AMD $510.64 -1.02%
BTC $64,629.98 -0.00%
ETH $1,830.16 +0.88%
BNB $593.08 +0.11%
XRP $0.9610 +0.47%
SOL $72.84 +2.25%
TRX $0.3062 +1.24%
DOGE $0.0590 +0.18%
ADA $0.1535 +0.66%
BCH $208.77 +0.25%
LINK $9.18 +0.35%
HYPE $69.78 +5.06%
AAVE $112.83 +2.65%
SUI $0.7242 +0.33%
XLM $0.1378 -0.68%
ZEC $922.25 +3.79%
AAPL $352.90 -1.22%
AMZN $257.64 -2.23%
GOOGL $283.98 -3.14%
MSFT $517.58 -1.02%
META $622.99 -1.72%
NVDA $157.68 -2.27%
TSLA $379.02 -1.61%
SNDK $814.69 -3.24%
INTC $54.66 -4.77%
SPCX $137.93 -1.73%
MU $723.80 -3.25%
AMD $510.64 -1.02%

consumption

All
Article
Flash

hot_img Cerebras releases the fourth generation AI inference system CS-4: performance doubled, power consumption doubled, more flexible deployment

Cerebras released its fourth-generation AI inference system CS-4 this week, based on the same 5nm WSE-3 wafer, achieving double the performance by doubling the clock frequency and power consumption. A single CS-4 cabinet accommodates 3 wafers (CS-3 has 2), featuring a modular "backpack" design that simplifies manufacturing and deployment, with a TDP of approximately 125 to 135kW. The CS-4 can provide an inference speed of nearly 4000 tokens/second/user, about twice that of the CS-3, and supports decomposed inference with heterogeneous systems such as AMD and AWS Trainium.Cerebras claims that the CS-4 offers about 2000 times the on-chip memory bandwidth of NVIDIA's Rubin (43PB/s), but the 44GB SRAM capacity remains unchanged, and long-context inference still requires multi-wafer stacking. For example, with the DeepSeek V4 Pro (1.6T parameters), approximately 20 systems are needed for a 1M context window, and about 40 systems are required for 256 concurrent users, corresponding to a CAPEX exceeding 20 million USD. Cerebras is collaborating with clients such as OpenAI and plans to achieve approximately double performance improvements each year, aiming for a 20-fold throughput increase by 2027. The "backpack" cabinet design of the CS-4 will continue into the next-generation "Nexus" platform.
app_icon
ChainCatcher Building the Web3 world with innovations.