Vinay Kumar
Perfection is a luxury when you’re competing with limited resources in a lopsided geopolitical race
In a few months, India will roll out its first homemade chip. Nothing fancy, it will be a 90nm chip for low-end applications. China, meanwhile, has progressed to making machines that make advanced chips. It’s a very big deal. Even US buys most of its chip-making machines from the Dutch company ASML. That’s why China’s feat has lessons for us. US had banned sale of these machines to China, to ensure it didn’t get ahead in tech. So, China decided to make its own by hook and by crook. It has an army of very competent PhDs. The ‘hook’ part meant stepping up R&D and reverse engineering, and ‘crook’ meant poaching staff from ASML and other labs.
Yes, Chinese chip-making machines aren’t the world’s best. In fact, they are unproven at this point. But when the whole tech and AI race depends on chips, they create a “good enough” buffer against US sanctions and outright bans. China intends to make five of these machines this year, and 20 in 2027. And guess who’s helming the project? A state-owned firm called Shanghai Aishengna. Kind of like our HMT, which is defunct now.
Being “good enough” has its advantages, as Chinese AI models have shown. Recently, a US startup that was attacked by a rogue OpenAI agent used a Chinese model, GLM-5.2, to defend itself. Because Chinese models are far cheaper, and good enough, they are rapidly capturing the market. Point is, you don’t have to be absolutely top-notch to succeed. Ambition and matching efforts can give you a fighting chance in any race. That’s why our progress from 90nm to 28nm in chips, and then to 4nm and up, shouldn’t take decades. In AI also we need to move faster. Let’s get more ambitious.
Disclaimer
Views expressed above are the author’s own.
