|
 |
|
| Nvidia CEO wants TSMC to boost output for AI chip demand |
2/3/2026 |
|
Nvidia CEO Jensen Huang is urging Taiwan Semiconductor Manufacturing Co. to "work very hard" to meet demand for AI chips, noting that Nvidia alone could require TSMC to double its capacity over the next 10 years, with an immediate jump in output needed this year. TSMC has already hinted at a major increase in capital spending, projecting up to a 37% boost in 2026 and further investments in 2028 and 2029.
|
|
|
| Alibaba ships more than 100K AI chips |
2/2/2026 |
|
Alibaba Group has shipped over 100,000 units of its advanced AI chip, the Zhenwu 810E, beating local rival Cambricon, according to sources. The chip's performance is comparable to Nvidia's H20, highlighting Alibaba's growing role in China's push for alternatives to US processors.
|
|
|
| AMD, Qualcomm to use SOCAMM for AI products |
1/30/2026 |
|
Advanced Micro Devices and Qualcomm plan to use small outline compression attached memory modules in their AI products, following Nvidia's lead. SOCAMM, based on low-power double data rate dynamic random-access memory, is upgradable and works alongside high-bandwidth memory to handle memory-bound tasks. AMD and Qualcomm are also exploring a different approach to integrating SOCAMM to improve power control and reduce motherboard complexity.
|
|
|
| ASML sales soar |
1/29/2026 |
|
Q4 total net sales were €9.7 billion at a gross margin of 52.2% with net income of €2.8 billion
|
|
|
| Nvidia, Apple to use Intel for some chips, reports say |
1/29/2026 |
|
Nvidia will partner with Intel for the 2028 Feynman GPU chip, while Apple is considering Intel for entry-level M-series processors. Both companies will still rely on Taiwan Semiconductor Manufacturing Co. for core production. The move is part of a dual-foundry approach to mitigate risks.
|
|
|
| China greenlights Nvidia H200 chip imports |
1/29/2026 |
|
China has approved the import of Nvidia's H200 AI chips after an initial rejection, with several hundred thousand chips allocated, according to reports. The H200 is Nvidia's second-most-powerful AI GPU, behind the restricted Blackwell B200, but is more capable than the H20, which was previously the chip allowed.
|
|
|
| Researchers create energy-efficient memory device |
1/27/2026 |
|
The device processes data only from active pixels, ignoring background noise, and the proof-of-concept system recognizes patterns with up to 92% accuracy, outperforming traditional methods.
|
|
|
|
