Summary of Key Points
After Yushi Technology's listing on the Hong Kong stock market, founder Wu Gansha reflected on his ten-year entrepreneurial journey: From his time at Intel, he witnessed numerous failed innovations and learned to set a minimum standard for success. When starting a business, he avoided the chip and robotics sectors and chose the autonomous driving industry. He began by tackling challenging scenarios in closed environments such as airports (referred to as “Yizhou”) before aiming for more general applications (referred to as “Jingzhou”). After enduring the hardships of being five years ahead of others in technology, his company has now matured in terms of technology, cost, and efficiency. Yushi is ready to expand globally, with the goal of becoming the “AI driver for the world” by providing subscription-based services. His experience offers practical advice to tech entrepreneurs: survive first, then seek growth.
1. Intel’s “Death Valley” taught him to set a minimum standard and avoid reckless ventures
Wu Gansha spent 16 years at Intel, where he saw many innovative projects fail. Some technologies were too radical (for example, the Larrabee chip, which aimed to compete with NVIDIA but failed due to reliance on the outdated X86 architecture), while others were disconnected from market needs (for instance, a chip designed for convolutional processing in 2009 was abandoned due to internal decisions). These experiences made him extremely cautious about potential pitfalls:
- Avoid chips: The entry barrier for small companies into the chip industry is too high; even Intel had difficulties with partnerships like Unisoc.
- Avoid robotics: Robotics were costly at that time (the chassis, arms, and hands combined cost hundreds of thousands of dollars), making them far from practical for household use.
- Choose autonomous driving: The application scenarios are standardized (similar traffic rules worldwide) and the control is simple (only acceleration, deceleration, and steering are required). He followed Tesla’s three-step approach to development.
In short, entrepreneurship is not about betting on a “sexy” idea; it’s about ensuring survival first.
2. Don’t chase trends! Focus on practical challenges first
In 2016, L4 Robotaxi (driverless taxis on public roads) were the hottest topic in autonomous driving, but Wu Gansha took a counterintuitive approach:
- The “Longzhong Plan” strategy: He used closed environments like airports and ports as secure bases (similar to Yizhou) and general applications like passenger vehicles at L2–L3 levels as stepping stones to broader markets (Jingzhou).
- Why airports?: He calculated the benefits using Product-Market Fit (PMF): An airport requires 3.5 drivers per vehicle, with each driver costing hundreds of thousands of Hong Kong dollars annually. The environment is also highly controllable since all vehicles follow rules, ensuring the safety of autonomous vehicles.
- Abandon car manufacturing: In 2018, he invested tens of millions in car manufacturing but realized that automakers were focusing on advanced control systems, while his company was stronger in autonomous driving technology.
As a result, Yushi has a 90% market share in airports and is the only company operating large-scale autonomous services—first proving the concept before expanding to buses and ports.
3. The hardship of being five years ahead: Technology lagged behind, so they had to endure
Wu Gansha admitted that most autonomous driving companies started five years earlier than them. By 2017, end-to-end and global model technologies were emerging, but data and computing power were insufficient, forcing them to rely on rule-based methods (e.g., stopping at red lights). It took two years to meet safety standards, another three years to improve efficiency (e.g., ensuring vehicles didn’t delay passengers or brake excessively), and another two years to reduce costs (e.g., reducing the number of remote operators from 1:3 to 1:10). Many companies failed; five of the top 10 autonomous driving firms listed by CCID in 2021 are no longer in business. Wu Gansha compares Chinese entrepreneurs to “squirrel fish”—they remain resilient despite difficulties and only admit defeat when they can no longer hold on.
4. Ten years of preparation: Ready to expand and become a global leader
Now that technology, costs, and application scenarios are mature, Yushi is ready to expand:
- Goals: Increase market penetration from 1% to 10%, expand from 6 countries to 60 markets, and develop 5–10 new products with tenfold growth.
- Business model: Provide subscription-based AI driver services—without manufacturing or operating vehicles directly, they will supply AI drivers to various industries (e.g., airports and ports). According to his calculations, 100,000 AI drivers generating $10,000 each year would total $1 billion; 1 million drivers would generate $10 billion.
In short, the initial focus on developing general-purpose technology pays off in large-scale revenue.
5. Advice for tech entrepreneurs: Survive first, then pursue dreams
Wu Gansha’s experience can be summarized as follows:
1. Set a minimum standard: Avoid overly ambitious projects and find a stable foundation (like an airport). “First, ensure you can’t be defeated; then wait for the right opportunity.”
2. Wait for the right time: If your direction is correct but you started too early, make compromises to survive (e.g., focus on specific scenarios instead of broader applications).
3. Universal technology is key: Make your technology versatile (e.g., AI drivers that can operate buses, trucks, and passenger vehicles) to achieve lower marginal costs and scalability.
He emphasizes that entrepreneurship is not about a dramatic breakthrough; it’s about turning seemingly wrong choices into viable solutions. Only by surviving can you pursue your dreams.
This report offers a practical guide to entrepreneurial survival, reminding tech entrepreneurs to be humble and persistent in their efforts.