According to foreign media Barron’s, Intel CEO Pat Gelsinger said in an interview recently that the performance of Intel’s 18A process will be ahead of TSMC’s N2 (2nm) process.
Gelsinger pointed out that although the transistor density of Intel’s 18A process and TSMC’s N2 process seem to be similar, everyone says that Intel’s backside power delivery technology is better. He said that this allows silicon chips to have better area efficiency, which means lower costs, and better power supply means higher performance.
Gelsinger said that good transistor density and excellent power supply make the Intel 18A process slightly ahead of TSMC N2. In addition, TSMC’s packaging costs are very high, and Intel’s gross profit is expected to increase slowly.
Intel 18A is slightly better than TSMC N2, and US AI will continue to lead
When asked how the Sino-US AI competition will evolve in the next 10 years? Kissinger replied that when people mention AI, they think of computing power, data supply and innovative research algorithms. The United States has clearly taken the lead in terms of computing power. Everyone thinks that China has an advantage in data, but in fact, the openness of the United States has made many data sets readily available. The third is that Americans have integrated innovation into their lives and have an open and challenging character. It is difficult for China to keep up.
Kissinger believes that China’s policies are becoming more conservative and are increasingly unable to produce highly innovative entrepreneurs, and data is becoming increasingly difficult to obtain. As for how to measure the degree of AI innovation? From the literature referenced in top AI papers, we can find that the United States has taken a dominant position.
More about technology news please visit our site.