news

한어Русский языкEnglishFrançaisIndonesianSanskrit日本語DeutschPortuguêsΕλληνικάespañolItalianoSuomalainenLatina

Text | Quadrant, Author | Li Wei, Editor | Sun Erxi

Burning the cold stove means paying attention to unpopular fields in advance and investing in them for a long time. To some extent, "burning the cold stove" is consistent with the successful experience shared by Nvidia CEO Huang Renxun of Nvidia's search for the "zero dollar" market to make a breakthrough. In China, there are also successful cases of new energy vehicles and short video platforms that seize the "zero dollar" market by burning the cold stove of innovation in the track. In the current development environment, burning the cold stove has become the key for enterprises to achieve breakthroughs. To achieve this, vision, patience and flexibility are indispensable.

TSMC's factory in Japan is located in Kikuyo Town, Kyushu Island. This is a small town with only 43,000 people, but now it shoulders the heavy burden of making Kyushu Island a "Silicon Island". According to the estimates of Japan's Kyushu Financial Group, TSMC will bring 7,000 jobs and US$28 billion in wealth to this town in the next ten years.

This is a microcosm of Japan's grand plan to revive the semiconductor industry. In the past few years, Japan, from the government to enterprises, has continuously emphasized the importance of revitalizing the semiconductor industry and quickly put this cognition into practice. Introducing Taiwanese semiconductor companies has become the key to this action. Like Kikuyo Town, Taiwanese semiconductor companies are injecting strong development momentum into the revival of Japan's semiconductor industry.

However, both the Taiwanese semiconductor companies that are flocking in and the Japanese semiconductor industry practitioners need to admit that Japan is not currently the main target customer location for Taiwanese semiconductor companies. Among the top seven customers announced by TSMC in 2023, only Sony, a Japanese company, ranks seventh. In the second quarter of 2024, TSMC's revenue will come from North American customers, 16% from mainland China, and only 6% from Japan.

It is possible that in the future, Japanese companies with deep industry accumulation in industries such as automobiles, industry, consumer products and high-performance computing (HPC) will be able to generate greater semiconductor demand growth, providing a new growth curve for Taiwanese semiconductor companies. It is also possible that Taiwanese semiconductor companies will be able to establish deeper cooperation with Japanese semiconductor companies at the upstream of the industry chain, tap into their technological accumulation, and export factory management experience to them.

This kind of "burning a cold stove" behavior with high uncertainty greatly tests the vision, patience and flexibility of enterprises. First, you need to have a deep insight into the long-term development trend of the industry and accurately judge the growth of an opportunity in five or ten years; second, you need to have enough patience and wait unremittingly for the opportunity to mature; of course, new factors will also appear, turning opportunities into mistakes, at this time, enterprises need to have the flexibility to change lanes in time.

This "burning a cold stove" concept is very similar to a successful experience shared by Nvidia CEO Huang Renxun: In many cases, companies need to find a "zero dollar" market and invest in it with great courage to lock in greater industry development opportunities in advance. This so-called "zero dollar" market refers to a blank market that has neither large-scale customers nor strong competitors at present, but has huge development potential.

Especially when the industries that originally led the development are entering the mature stage and new leading industries are about to take over, "burning the cold stove" and looking for the "zero dollar" market like TSMC and NVIDIA have become the only way for Chinese companies to become stronger.

Finding the next “zero dollar” market

Lin Yifu once divided China's current industries into five categories: catching-up industries, leading industries, transition industries, overtaking industries and strategic industries. He believes that both leading industries and overtaking industries need to rely on their own research and development to obtain new technologies, and then maintain or achieve leadership in the market. The research and development of new technologies requires companies to be brave and good at starting over.

In the context of the transition from the old to the new in the industry, the current leaders are more likely to face the possibility of being overthrown by technological innovators. Therefore, companies such as Microsoft, Google, Tencent, and Alibaba, which have experienced the transition from the Internet to the mobile Internet, will be more urgent to lay out new tracks and new trends. Traditional car companies such as Volkswagen and Toyota have also begun to bear the increasing competitive pressure from Chinese new energy car companies.

At the same time, for new energy vehicle companies, changing lanes means betting on the "zero dollar" market and burning the cold stove of next-generation technology. Just like Nvidia, in 1993, when CPUs were still thriving and popular, it chose to make accelerated computing chips, turning this then "cold stove" into a "hot stove" that is now booming.

The same is true for new energy vehicle companies. Over the past 100 years, developed countries have established patent barriers in the field of internal combustion engines that are difficult to bypass. Traditional Chinese car companies that lack core technological advantages have found it difficult to achieve breakthroughs in this field. New energy vehicles provide a new track for everyone to start anew, and on this track, Chinese new energy vehicle companies have been able to establish a competitive advantage.

In 2021, 2022 and 2023, China's automobile exports exceeded 2 million, 3 million and 4 million respectively, and its market share in Europe, South America, Australia, Southeast Asia and other regions has achieved great development. Among them, the export volume of new energy vehicles increased from 310,000 in 2021 to 1.203 million in 2023, from one new energy vehicle in every seven exported vehicles to one new energy vehicle in every three exported vehicles.

Before this achievement, China's new energy vehicle industry had been a niche track with immature technology for a long time. Although BYD developed the pure electric sedan F3e in 2006, and the state proposed the development goal of forming a production capacity of 500,000 new energy vehicles in the "Automotive Industry Adjustment and Revitalization Plan" issued in 2009, by 2014, the annual production and sales of new energy vehicles were only 78,500 units.

In the past ten years, the scale of China's new energy vehicle industry has achieved a rapid growth of more than 100 times. This achievement is not only due to the strong support of the country at the policy level, but also to the persistent R&D investment and bold innovation of Chinese auto companies in this once "zero-dollar" track. BYD alone has invested hundreds of billions of yuan in R&D in the past decade, has tens of thousands of technical R&D personnel, and has applied for more than 40,000 patents.

"Cooking a cold stove" requires more patience

This shows that one must have enough patience and tenacity on the road of "burning a cold stove". The biggest difference between "burning a cold stove" and "chasing the wind" is whether to bet your long-term vision on a budding track and become a trend creator, or to make short-term operations on a track that has already taken off and obtain short-term benefits by becoming a vassal of the growth trend.

The result of a company's choice will largely determine whether it is qualified to become a great company. Take Nvidia as an example. In 2016, Huang Renxun personally donated the world's first DGX-1 to OpenAI. At that time, compared with its then CEO Musk, OpenAI was still a little transparent in the eyes of the public, and it was still six years before the company detonated the wave of generative AI. And this device, which was worth $129,000 at the time, was just a drop in the bucket for Nvidia's investment in the future.

Nvidia could not have predicted that OpenAI would inject new vitality into the AI ​​industry by the end of 2022. Just like they could not have predicted when they were promoting CUDA that three people named Jeff Hinton, Alex Krzyzewski, and Ilya Sutskever would use Nvidia's CUDA to train AlexNet, which detonated deep learning. However, they firmly believe that the era of accelerated computing will come sooner or later.

"When the company believes in something, we should take action. So we delved into deep learning and systematically reshaped everything over the next decade," Huang said in his speech. "We went on to invent computing, interconnects, systems, networks, and almost every other aspect of computing. And of course, software. We invested billions of dollars in unknown areas. For a decade, thousands of engineers worked on deep learning and advancing and extending deep learning, but they didn't really know to what extent we could really apply this technology."

During this process, Nvidia was also questioned by shareholders due to its long-term poor performance caused by R&D investment, and was advised to focus on improving profitability. But Nvidia's current market value proves how much the patience and resilience based on precise insights can pay off. The company's market value has long been around $1 billion, but now it has become $3 trillion.

NVIDIA is both the implementer and the sufferer of "burning the cold stove". In 2017, when Quanta gave up on manufacturing DGX-2 for NVIDIA due to too little demand, NVIDIA found VEICHI to manufacture it for it. During the production of DGX-2, VEICHI realized the importance of GPU to the development of AI technology, and began to allocate a lot of resources to cooperate with NVIDIA's research and development before NVIDIA opened up the market. Finally, a few years later, VEICHI tasted the dividends of the AI ​​era brought by NVIDIA.

For many Chinese companies that are already in a leading position in China, they need to be more patient with the "zero-dollar" market like Nvidia and Flextronics, and face the darkness before dawn with greater resilience. In particular, today's leading Chinese companies are increasingly facing global competition, which will be more directed towards the research and development and promotion of next-generation basic technologies and infrastructure-based industrial applications. This requires more patience and resilience.

Persistence is not stubbornness, you need to change at the right time

Of course, the emphasis on patience and resilience does not mean that you have to stick to one path until the end. Persistence is not stubbornness. When the competitive environment and technological development change, companies need to change in time and re-select breakthroughs. When sharing his secret to being a CEO, Huang Renxun said: "On the one hand, you have to persist and stick to your beliefs; on the other hand, you can't be stubborn, so that you have flexibility. This is contradictory and complex. Many CEOs of startups are very talented, but they are determined to prove that they are right and forget about agility."

When introducing NVIDIA's entrepreneurial journey, Huang Renxun always expressed his gratitude to Shoichiro Iriko, who was once the president of Sega. This gratitude came from a wrong choice that almost killed NVIDIA in the cradle. After NVIDIA was established, it chose to use quadrilateral image technology to build its own graphics accelerator product NV1. In the same year, Microsoft released Windows 95, which swept the market, but the Direct 3D standard it adopted was built based on triangle image technology.

As a result, Nvidia was caught in a compatibility crisis and failed to provide Sega, its first major customer, with graphics accelerators that continued to use quad graphics technology. Huang Renxun told Sega frankly that due to Nvidia's R&D strategic mistakes, NV2 was no longer necessary to continue to develop, but he still hoped that the other party would pay according to the contract. In the end, it was Shoichiro Irikyo who pushed Sega to pay the funds that Nvidia urgently needed to restart R&D.

At the beginning of Nvidia's journey to a market value of 3 trillion, Huang Renxun promptly chose to embrace Microsoft's Direct 3D standard, and did not continue to insist on developing graphics accelerators using quadrilateral image technology to prove himself. This led Nvidia out of a dead end, and Shoichiro Iriko's kindness to the startup company gave it the opportunity to correct its mistakes. "It was because we faced up to our mistakes, stopped losses in time, and humbly sought help that we survived the crisis of bankruptcy," Huang Renxun later concluded.

Japan, which is trying to revive its semiconductor industry, has seen many negative examples of "changing at the right time" during the decline of the industry. In his book "The Lost Thirty Years: A History of Japan's Economy in the Heisei Era", Yukio Noguchi regards sticking to the old business model as one of the reasons for the decline of Japan's manufacturing industry. The specific manifestation of this is the obsession with building large vertically integrated factories.

"At that time (2004), the development trend of the world's manufacturing industry was to gradually shift to a horizontal division of labor production mode, and its representative company was Apple... Because they were all standard production processes, they could be completely replaced by other companies... But even after the Lehman incident, Japanese companies still insisted that they must establish large-scale vertically integrated factories... It was not until the mid-term settlement in the fall of 2011 that Japanese companies realized the era of the vertical integration model. At that time, Japan's electronics industry fell into a state of deficit across the board." Yukio Noguchi analyzed.

Finally, "Guishan Manufacturing", a synonym for exquisite LCD TVs, switched from producing large LCD panels for TVs to producing small and medium-sized panels, and was eventually acquired by Hon Hai Group, which benefited from the horizontal division of labor model, in 2016. But many years later, BYD also pursued a vertical manufacturing model that controls everything from the production of parts to the final product, and it has embarked on the fast track of development.

Timely changes will ultimately determine whether a cold stove can be turned into a hot fire. Chinese companies are not lacking in flexibility, but they need to have a more precise vision and more courageous persistence to grasp trends and opportunities.