🚨 “AI chips that are 30x faster than Nvidia” are here—
but the stock price has crashed 13%! Are investors not buying it?
Group: 点击进入玖玖的粉丝群
According to the company, CS-4 can generate tokens at a rate of up to 30 times the speed of traditional GPU infrastructure on a per-user, per-second basis. But the market’s reaction is completely the opposite.
📉 During Tuesday’s trading session, Cerebras’ stock fell by more than 12% at one point, dropping to around $218.
More importantly, compared with the $350 opening price when the company listed in May, the stock is already down by over 35%. Why, after the release of a chip touted as so powerful, are investors voting with their feet instead? The key may be not just the chip.
🔥 This time, Cerebras’ CS-4 is built from three WSE-3 Turbo wafer-scale processors. Each processor has around 40 trillion transistors. The company hopes that this ultra-large wafer-scale architecture will reduce data transfer between multiple chips in traditional GPU systems—thereby lowering latency and improving AI inference speed. Cerebras also uses SRAM as its core storage solution.
🚨 But what truly worries investors is the financial data.
In the second quarter, the company reported an approximate loss of $2.98 per share, compared with a profit of $1.91 per share in the same period last year. In other words, the product performance story is eye-catching, but profitability hasn’t kept pace.
This explains the very clear mismatch the market is showing:
The chip release news is big, yet the stock plunges. Because in the end, the capital markets don’t only look at “how fast the chip is”—they also want to see whether that performance can actually translate into orders, revenue, and profit.
And Cerebras is clearly betting on the future.
The company expects that by the end of 2027, it will deploy about 600 megawatts of computing infrastructure. During this period, it plans to increase overall throughput by 20 times and speed by 4 times. Meanwhile, CS-4 is expected to launch officially in the third quarter of 2026.
Click the avatar to watch the live stream + join the Jiujiu chat group to get daily strategies 🚀
#Cerebras #AI芯片 #NVIDIA
but the stock price has crashed 13%! Are investors not buying it?
Group: 点击进入玖玖的粉丝群
According to the company, CS-4 can generate tokens at a rate of up to 30 times the speed of traditional GPU infrastructure on a per-user, per-second basis. But the market’s reaction is completely the opposite.
📉 During Tuesday’s trading session, Cerebras’ stock fell by more than 12% at one point, dropping to around $218.
More importantly, compared with the $350 opening price when the company listed in May, the stock is already down by over 35%. Why, after the release of a chip touted as so powerful, are investors voting with their feet instead? The key may be not just the chip.
🔥 This time, Cerebras’ CS-4 is built from three WSE-3 Turbo wafer-scale processors. Each processor has around 40 trillion transistors. The company hopes that this ultra-large wafer-scale architecture will reduce data transfer between multiple chips in traditional GPU systems—thereby lowering latency and improving AI inference speed. Cerebras also uses SRAM as its core storage solution.
🚨 But what truly worries investors is the financial data.
In the second quarter, the company reported an approximate loss of $2.98 per share, compared with a profit of $1.91 per share in the same period last year. In other words, the product performance story is eye-catching, but profitability hasn’t kept pace.
This explains the very clear mismatch the market is showing:
The chip release news is big, yet the stock plunges. Because in the end, the capital markets don’t only look at “how fast the chip is”—they also want to see whether that performance can actually translate into orders, revenue, and profit.
And Cerebras is clearly betting on the future.
The company expects that by the end of 2027, it will deploy about 600 megawatts of computing infrastructure. During this period, it plans to increase overall throughput by 20 times and speed by 4 times. Meanwhile, CS-4 is expected to launch officially in the third quarter of 2026.
Click the avatar to watch the live stream + join the Jiujiu chat group to get daily strategies 🚀
#Cerebras #AI芯片 #NVIDIA