One of the products in our company is a viscosity controller used in the printing process. The equipment continuously senses the viscosity of the ink and, depending on whether it is above or below the desired level, automatically adds or reduces the solvent to maintain consistency throughout the printing process. It senses a changing environment, responds to it and continuously adjusts its action. 

Yet it is automation, not AI. It is a closed-loop feedback-control system operating according to predefined parameters. This distinction is important because sensing, responding and operating autonomously do not, by themselves, make a system AI.

The recent robotic “Olympiad” publicised in China provides an interesting parallel. Robots participated in running, jumping, football and other athletic events. But if their movements are being executed according to automated programmes and conventional feedback-control systems, the achievement is fundamentally one of robotics, sensors, actuators, control engineering and automation. AI may certainly be incorporated into some aspects of such robots - such as perception, prediction, decision-making or learning - but the mere fact that a robot performs an athletic task autonomously does not make the task AI.

My point is therefore not that AI has no role in modern robotics, but that AI is increasingly being used as a promotional label rather than a precise description of the underlying technological achievement. A sophisticated automated system does not become AI merely because it senses, responds or acts on its own.

China Working on AI for Ten Years

This demonstration by China has caught Americans by surprise a couple of times over the last few years. They remember, of course, the DeepSeek moment, when a Chinese model demonstrated capabilities that challenged some prevailing assumptions about the cost and resources required to develop advanced AI. And then, more recently, came developments from a firm called Moonshot.

This is especially striking because China has had less capital at its command and has also faced export controls on some of the most advanced chips in the world. Chinese AI companies have, of course, been accused by American AI makers of what is called distillation - essentially, using the outputs of more advanced models to develop or improve their own models. At the same time, however, there is now a growing recognition across Silicon Valley and Washington that China has been able to make much faster progress than the United States had expected.

Contrary to what many perceive, China has been at it for more than ten years, with a great deal of focus at the leadership level. This sustained effort, it turns out, has been decisive because China has been able to make much faster progress than the United States expected.

There are essentially three things you need for AI: sophisticated chips, energy and human talent.

Indigenous Computer Chips

China has plenty of human talent, including at the very highest levels of computer science. It also has enormous energy capacity, probably around ten times as much installed electricity-generation capacity as the United States, although it also serves a much larger population.

What China did not have was access to the most advanced computer chips. So how did it get around that?

Partly by importing advanced chips through Southeast Asian countries, and partly by developing its own indigenous chips. This is a story that began during the first Trump term. The first trade war was a real wake-up call for Beijing. China realised then that it was entering a period of greater strategic competition with the United States and began focusing much more strongly on self-reliance in critical technologies.

Even if the hype is set aside - and China is certainly very good at hype - there is now growing recognition that China has made substantial progress in AI, and at a pace that has surprised many in the world.

What is an AI Race?

But what exactly is an AI race? Is it only between China and the USA? There is a possibility that this whole idea of a race is misconceived, because the whole world is going to use AI for all kinds of things.

Take the analogy of electricity. There was no electricity race. There were pioneers, competing technologies, companies and countries that made important advances, but there was no finish line at which one country could declare itself the winner. Thomas Edison played a major role in commercialising electricity and built General Electric, which went on to become a major company. But electricity did not belong to Edison, America or any one country. It was progressively adopted, developed and transformed by the whole world.

AI may well follow a similar path. The question may not ultimately be who wins the AI race, but who is able to use AI most effectively and diffuse it most widely through the economy.

The conversation in Silicon Valley about AI is very different from the conversation in Beijing, Shanghai, Delhi or Bangalore. There is a lot of talk about achieving artificial general intelligence, or gaining a decisive strategic advantage, the idea that there might one day be a point where we cross the tape and, therefore, the game is ours.

Beijing and Delhi see it quite differently. They say: “What we are trying to do is get AI into the economy, and diffuse it as widely as possible.”

We can already see this happening in China's factories, where robots are working on production lines. There are, of course, pros and cons to this. This is not an unambiguous good. But it is giving China an economic advantage. If AI and robotics reduce the cost per unit of production, Chinese manufacturers become more competitive around the world.

And in some ways, that is the ballgame.

The real significance of AI, therefore, may not be about who is the first to cross some hypothetical finish line called AGI. It may be about who is able to diffuse AI most rapidly through the real economy, into factories, services, logistics, research and everyday business, —and convert technological capability into productivity and competitiveness.

That is a very different conception of a race. And, like electricity, it may ultimately be a race without a finish line and without a single winner.

(The author is an Indian Army veteran, a tech entrepreneur and a contemporary affairs commentator. The views expressed are personal. He can be reached at  kl.viswanathan@gmail.com )