China is valuing the possibility of increasing your investment in local chip manufacturing. And according to Bloomberg, this new investment could reach up to $70 billion. This figure would be added to current Chinese government initiatives, such as Big Fund III, started a few months ago. If it occurs, this investment could accelerate the design and manufacturing of chips to facilitate the competition of Chinese companies with NVIDIA and other competing companies.
Over the past year, China has announced several large projects as well as various initiatives to try to boost its domestic chip supply as well as reduce its dependence on US chip supplies. Also to reduce its quota of purchasing chips from other countries, and therefore reduce the shortage to which other countries are exposed due to restrictions on their manufacturing.
These projects are quite diverse, and while some have been responsible for promoting chip production locally, others are designed to discourage the use of hardware from other countries. For now, though, Chinese chip designs appear to be several generations behind the most advanced AMD or NVIDIA products, so they still have a long way to go to catch up. Of course, an investment like the one they are considering would undoubtedly contribute to reducing this gap.
There are sources that place the investment that China will make in this regard at a lower figure, which is around 28 billion dollars. But if one like the one valued by Bloomberg, of 70,000 million, occurs, it would be the largest government investment in the manufacturing of semiconductors made by a government of any country in the world.
In any case, it would be in line with President Xi Jinping’s approach to domestic chip manufacturing. The president sees his push as a strategic imperative for his defense, and also to try to break the stronghold that the United States has in the manufacture of cutting-edge chips.
China has already offered energy subsidies to companies using domestically produced chips, and has mandated that Chinese companies running AI inference workloads must use locally produced at least half of the chips they use in their data centers. But for now they have not had much success in forcing local companies to use domestic chips for module training, due to their lower performance compared to NVIDIA’s.
This, however, could change in the future, since China’s chip technology has made progress quite rapidly. Some companies say they have already used advanced assembly and packaging technologies to improve the performance of older process nodes. Although we have yet to see evidence of this, and there are concerns about its heating and the performance of the chips manufactured, China is, without a doubt, Taking steps toward future independence in semiconductor manufacturing.
