Source Code Capital | I'm 28 This Year, and I've Been Working on Power Supplies for Ten Years and One Month

源码资本源码资本··0·6

#Source Code Capital

Shuang Liu, founder and CEO of Senmu Leishi, is currently a PhD candidate in Electrical Engineering at Huazhong University of Science and Technology. In his fourth year of undergrad, he started his own power supply design studio, earning his first bucket of gold through technical work. In 2017, while pursuing his master's degree, he founded Senmu Leishi.

Through his entrepreneurial journey, Liu met various power supply companies and discovered that the industry's customization attributes were particularly pronounced, with heavy reliance on R&D across the board. Inspired by hardware-in-the-loop simulation equipment, he decided to tackle the industry's non-standard customization problem through "controller standardization." As the first company in the visible global landscape to build a "standardized digital power controller," Liu's leap of faith is filled with ambition and boldness. Yet simultaneously rooted in the traditional power supply industry, he possesses a sophistication and clarity beyond his years.

Livestream Excerpts

Source Code Capital Runtao Zhang: Why are you so persistent and obsessed with power supply design?

Shuang Liu: I studied electrical engineering in college. Throughout that process, I didn't really make many choices — from my first year entering the lab, I worked on power supply-related research. Actually, power supplies were a sunset industry back then (2012). Nobody chose it; there were only two or three people in the entire lab. But I think I'm a somewhat "dumb" person. Once I choose a path, I just keep my head down and walk to the end.

But it's precisely this "dumb" quality that let me persist in this industry and keep accumulating. After accumulating to a certain point, I made innovations. Although nobody chose the power supply direction back then, now with the new energy industry's development, including electric vehicles, the entire electronics industry is desperate for talent. This seems like the reward for persistence. We never chase trends — we just stick to this one path, and as we walk, this path becomes a trend.

Source Code Capital Runtao Zhang: Throughout your entrepreneurial journey and on this path of power supply design, what positive feedback have you received?

Shuang Liu: Positive feedback comes from internal and external aspects. Internally, I think it comes from within. Having been in this industry for so many years, my firm internal understanding is that as long as human civilization doesn't fundamentally change, power supplies will always be indispensable — this field will never be eliminated.

Looking at human evolution, all major transformations have been accomplished through energy utilization. From the earliest fire-by-friction and flint inventions, when humans began harnessing energy; to the steam engine sparking the Industrial Revolution; to the current new energy revolution. Ultimately, the prerequisite for all energy to be usable by humans is conversion to electrical energy. Therefore, I am deeply convinced that the industry I work in is something humanity needs for the long term.

From external factors, for me personally it was more direct. In my third year, I opened my own studio, helping others draw schematics. I earned 8,000 RMB in two hours, which made me realize that knowledge could actually be converted into money. The excitement of making money opened a new door for me, and I've walked this path ever since. From both the macro and micro levels, I've continuously received positive feedback, which has made me more determined to persist in this field.

Source Code Capital Runtao Zhang: Senmu Leishi is the first company to create a "standardized digital power controller." From a technical difficulty standpoint, what's the approximate gap between developing a standardized power controller versus an ordinary digital power controller?

Shuang Liu: Let me use a human body analogy. A digital power controller is equivalent to a human brain — one person's brain only controls that one person's body, only adapted to that specific body. But there are all kinds of people: some do investment, some do sanitation work, some do other things — whatever they do, it's all controlled by the brain. A standardized power supply is like a power brain that manages different power supply types together, without interference between them, without unwanted reactions, allowing the people controlled by this brain to do investment, sanitation, or anything else.

With dedicated controllers, all interfaces are fixed. With general-purpose controllers, many interfaces are uncertain. This requires a product that can both accommodate different application scenarios and allow users to do secondary development for adaptation. Solving this problem is difficult — it's equivalent to performing a "head transplant," where you need to resolve various blood vessel matching, immune system adjustments, and many other issues.

Source Code Capital Runtao Zhang: How does a standardized digital power controller prioritize different application scenarios, and how do you price it?

Shuang Liu: Currently we choose high-end products with relatively high power, such as particle accelerator power supplies. These products have high industry specifications and high acceptance of new things. We first choose fields with high barriers — these are suitable at the current stage. For example, we choose semiconductor processing equipment power supplies, which are frequently constrained by overseas suppliers; we mainly do domestic substitution. For now, we won't choose red-ocean markets like charging piles; we'll find suitable opportunities before entering.

On pricing: although our product is entirely new, from a competitive standpoint, we appropriately benchmark against TI's DSP, STMicroelectronics, or domestic ARM chips as coordinates, trying to price slightly lower. For very high-end power supplies with much higher requirements, we price higher.

Source Code Capital Runtao Zhang: Why didn't this kind of standardized digital power controller emerge in the past? Why is this the right moment, and why is your team the one to do it?

Shuang Liu: First, the entire power electronics industry is enormous — wherever electricity is used, there are power supplies. China has roughly 400 billion RMB in annual power supply demand. But compared to the massive power market, it's like facing a huge mountain of gold. While some people are digging for gold, we're making the tools to dig gold.

Second, in the gold mine there are small gold flakes and large gold veins — what if you can't lift them? We've created a tool that can both lift large gold veins and dig small gold flakes. Moreover, this tool was honed through practice. For example, DC power supplies, AC power supplies, or pulse power supplies — different power supplies require completely different considerations. In practice, you need long-term accumulation of different types of algorithms, such as DC power algorithms, AC power algorithms. For gold-digging companies, spending time to accumulate these algorithms isn't cost-effective in terms of cycle and cost.

Finally, why our team? Probably because we come from a student background — we've accumulated ten years and one month in this field, amassing a large quantity of algorithms. If another company were to do the same thing, it would take half a year to a year, and require a sufficient number of engineers — very high costs. This is why the industry focuses more on the "gold mountain" and less on how to make a small shovel or hammer for digging gold.

Source Code Capital Runtao Zhang: Have there been any underlying technological changes that could enable this kind of standardized controller to emerge?

Shuang Liu: We reference hardware-in-the-loop simulation. Hardware-in-the-loop simulation is already very mature abroad. After silicon semiconductors matured, the power electronics industry basically settled into its form, with only some detailed optimizations and algorithm improvements remaining.

Some foreign companies discovered this hardware-in-the-loop simulation point and quickly verified the feasibility of these technologies. When we discovered this point, we believed we should not only do technical verification but also turn it into a product. This is how we iterated to PPEC (Programmable Power Electronics Controller), achieving rapid standardized development of digital power supplies.

Source Code Capital Runtao Zhang: So you saw similar value in something adjacent that inspired you. How did you judge at the time that this was a real problem worth solving?

Shuang Liu: I think it's still "from practice, to practice." Initially I didn't think this was particularly great either. We spent half a year making the hardware, plus previously accumulated algorithms, and made our current product — the digital power core controller (PPEC). Then we gave it to the most junior engineers in our company to use, and they found their efficiency improved greatly.

Then we took it to customers, and they also said that using Senmu Leishi's product, the speed exceeded expectations. Things they thought would take one or two years to gaoding, we could do in one or two months. Whether junior engineers or senior engineers, the feedback was that the experience was particularly good. This made us feel that this should be the right direction.

Source Code Capital Runtao Zhang: After resolving the company's direction, how do you define Senmu Leishi's "1" in the journey from 0 to 1? And why define it this way?

Shuang Liu: From 0 to 1, there are two factors: one is team, one is business.

From the team perspective, we have a "relatively" stable team. Although people join and leave along the way, our organizational framework is stable and can continue moving forward.

From the business perspective, we've planted our feet in the industry, can sustain ourselves, and can replicate a system. What comes next is to sustain this and keep replicating it.

Therefore, I believe when the team is sustainable and the business is sustainable, we've completed the journey from 0 to 1.

Source Code Capital Runtao Zhang: At the team organization level, through your exploration so far, as a young founder, what do you think are the core essentials for attracting high-end talent in this industry?

Shuang Liu: Chinese history has a very classic model for recruiting talent: "three visits to the thatched cottage." First is sincerity — I sincerely hope to work with you. I'm very patient when recruiting. I spent two years recruiting one entrepreneur to our company, constantly "egging him on" until he finally came. So far, people I've seriously invested effort in recruiting have stayed at the company continuously — this value is greater.

Second, satisfy the talent's material or development needs. Sincerity plus satisfying talent's needs regarding space and material aspects. Combine these two, and it basically solves the problem.

Source Code Capital Runtao Zhang: The power supply industry requires not just technical experience but also extensive business development networks. As a young engineer initially plunging into the business world, what detours did you take?

Shuang Liu: These detours were probably all related to student entrepreneurship. When I first started, I had to handle many things in parallel, with my heart racing every day. External problems were easier to overcome; internal adjustment took more effort.

After starting a business, I discovered that all efforts ultimately need to convert into corporate profit — you need to capture the market. But I was still a relatively green student. My clients were much older than me, even people my parents' age. I needed to buddy up with them as brothers, rather than treating them as elders like a child would. This transformation was very challenging for me.

Entering the business market, clients have various needs. Sometimes I couldn't quite grasp what they wanted; clients felt we were outsiders and couldn't connect. Through the process of exploration and change, I experienced much tempering and spent a great deal of time.

Source Code Capital Runtao Zhang: When you first encountered these situations, what were your genuine feelings?

Shuang Liu: Curiosity and fear from not seeing the full picture. When you can't see the full picture, there's some fear of the unknown. But once you see the full picture, you realize it's just that — though the process is painful and can't be skipped. I often tell my team: all pain that can be walked through looks beautiful in retrospect.

Source Code Capital Runtao Zhang: Can you share one or two stories, the most impressive ones? The most genuine psychological journey?

Shuang Liu: In undergrad, we participated in the National College Student Electronic Design Competition. We thought victory was within sight when a teammate suddenly said he had other things to do and quit. Looking back now, it wasn't that big a deal, but at the time, so many things suddenly piled on, and I felt my heart racing every day — very uncomfortable.

It was Selected Works of Mao Zedong that saved me. Volume One — I read it several times. It explained the problems faced in early entrepreneurship very thoroughly; you could directly apply it. That period lasted about two months, handling many things in parallel, which tempered my ability to handle problems concurrently.

Source Code Capital Runtao Zhang: Looking back, what do you think was the most important thing you learned from Selected Works of Mao Zedong?

Shuang Liu: The two most important words — what I saw most in Volume One of the Selected Works was "essence." At any time, you must grasp the principal contradiction, find the essence, and once found, everything else falls into place.

Once the principal contradiction is resolved, other contradictions can be easily resolved. Without grasping the principal contradiction, that problem remains there. So I often ask myself: is this problem really the real problem? I persist in asking myself this.

Source Code Capital Runtao Zhang: Were there moments when you felt like just giving up?

Shuang Liu: No. I'm a relatively "dumb" person. I always believe "dumb people have dumb luck." I never thought about giving up — I only thought about how this path should be walked. Like from undergrad to PhD, I've been working on power electronics, and will likely do it for life.

I'm from Hunan, living in Wuhan. There's a dialect expression in Hunan and Hubei called "ba de man" — very stubborn. People around me growing up were all stubborn; it's a thinking habit. The more others say it's wrong, the more I want to do it. Young people need to make mistakes — without mistakes, there's no growth. I don't feel I've taken the wrong path; I feel I just haven't taken the right one yet. Not taking the right path and taking the wrong path are completely different. Not taking the right path means I still have a chance to come back. I'll step in pits and make mistakes, but "when the cart reaches the mountain, there will be a road through; when willows darken, flowers brighten, another village appears" — just keep walking forward.

Source Code Capital Runtao Zhang: Standardized power controllers are still very innovative things. Early promotion was relatively difficult. Now when promoting such concepts or products in the market, what challenges or skepticism do you frequently receive from customers?

Shuang Liu: Our biggest problem is that although this thing is good, it faces changing user habits. Everyone currently makes power supplies using DSP, ARM, and other chips. With what we've made, they need to abandon their previous approach. Although it's simpler, many people say Apple's iOS is easier to use than Android, but when I'm used to Android, switching to iOS also feels inconvenient. In this process of getting customers to make the transition, customers will raise doubts.

Source Code Capital Runtao Zhang: Among all customer voices, some are valuable feedback, some are emotional noise. How do you distinguish signals from noise?

Shuang Liu: Mainly through data. Whether data is good or bad, first take it all in — absorb everything first.

During promotion, the team has a framework of thinking, equivalent to building a big data model, letting signals come in to judge which data is useful and which isn't, filtering accordingly. Data isn't good or bad; noise can actually help judge what's wrong, enabling iteration.

Because people have different perspectives, non-recognition actually brings us better things, helping us not be limited by fixed cognition, not become stubborn, triggering thought, and continuously adapting to new scenarios.

For example, customers often come to us to solve adaptation problems, which also makes us constantly consider issues from an adaptability perspective. Different application scenarios — how to adapt and quickly connect, equivalent to considering at the hardware level what to conveniently open up to customers, to open more interfaces.

Source Code Capital Runtao Zhang: Senmu Leishi is currently the first visible company globally doing power supply controllers. Throughout your business exploration, who is the most suitable learning object?

Shuang Liu: Learning objects aren't fixed. Entrepreneurship is a systematic project. Qi Baishi once said, "Those who learn from me live; those who imitate me die." In entrepreneurship, you need to eat "hundred-family rice" — draw broadly from many strengths.

In early entrepreneurship, Chairman Mao was my learning object: grasp the principal contradiction, and the principal aspect of the contradiction; in team management, "oppose formalism" and "oppose book worship." Huawei's team execution — "let those who hear the gunfire make decisions," standing on the front lines — is also worth learning from.

Senmu Leishi will form its own set of management or operational thinking — this is Senmu Leishi's characteristic. In the process of building our systematic project, we're eating "hundred-family rice" to construct our system.

Source Code Capital Runtao Zhang: Where do you think your competitors will come from? From large power supply companies, or from teams similar to yours?

Shuang Liu: For large power supply companies, they certainly have the capability to produce products like ours, but the time cost and human cost are very high. Now power electronics industry engineer compensation is almost catching up to the chip industry. Besides accumulating algorithms, they also need to build different platforms and verify effects — too costly for small companies, and for large companies it's not a main business, not really necessary. University-background entrepreneurial teams like ours are more likely, because they can complete accumulation without very high costs.

Source Code Capital Runtao Zhang: For a young company, how do you build a moat in this field?

Shuang Liu: First, create a business that will never let the company die — this is why we've always used PPEC to make power supplies. Now we'll select some high-end power supplies and use PPEC to build the business. On one hand, this creates industry benchmarks; simultaneously it tells customers that PPEC can solve high-end applications. On another hand, it can scale company revenue to a certain level, which is quite promoting for us.

At the PPEC level, during the modularization and chip-ization process, building moats and ecosystems is very important. When we accumulate to a certain user base, we plan to open-source — open-source all power supply schematics except the PPEC portion, making power supplies have no barrier.

Actually, the biggest moat is user habits. Like TI's moat in DSP power supplies. TI has a university program, constantly sending chips to universities. We young engineers all used TI chips; after graduation, our first reaction is, I've solved this work with TI chips before, so I choose TI chips. The barrier to cultivating user habits is very high. We now also need to reference this thinking, building moats in our early model.

Source Code Capital Runtao Zhang: Many young engineers in our livestream audience are also potential entrepreneurs. If you could give advice to young founders with innovative potential and engineering backgrounds, what do you think is most important?

Shuang Liu: Engineers first need to look at development direction. First, see what your engineering or research pursuit is. If you're doing research, you don't necessarily need to pursue market-oriented things. For research, just keep exploring — don't care whether this path explores right or wrong things out, and often being wrong more leads to being right, and you can find a right path.

Second is replicating your professional capabilities. In the entrepreneurial process, you might bury yourself alone in technology for months, living and eating in the lab. This causes social and communication skills to decline. Once the first-generation product is made, replicate as soon as possible, find someone to replace your technical role, so you have time to explore more forward-looking problems.

Third is finding a methodology to capture the market, and replicate yourself as soon as possible. As a technically-background entrepreneur, when you see your product being widely used, that sense of pride is very strong — even higher than the achievement of initially building the product.

Interact with Source Code Capital

How did you learn about Source Code Capital? And how did you find out about this episode of the Leap of Faith livestream? Did Runtao and Liu's conversation about entrepreneurial experience inspire you? Welcome to leave us a comment. We will select 10 readers with the most likes and gift them CEO Shuang Liu's favorite set of books.

For more MaHui members, click "Read Original" to view