DeepSeek in the context of "The China Innovation Method"

Over the past week, much of the AI world has been discussing the emergence of DeepSeek, a ChatGPT-like service and model allegedly developed at a fraction of the cost of similar Western solutions. However, allegations are now surfacing that much of its development was achieved by copying elements from ChatGPT through a process called distillation —a practice that, depending on your perspective, is either strategic and pragmatic or outright intellectual property theft.

Rather than diving into DeepSeek's technical specifics, I want to focus on a broader takeaway—what I learned about the "China innovation method" during my five years of working closely with some of China’s best engineers. My experience informs how I interpret DeepSeek's rise and what it means for competition in AI and beyond.

Copy First, Then Innovate

While leading global product management during 3Com's turnaround, one of my responsibilities was revitalizing our product lineup. A small part of this effort was launching a new desk phone—one of the few visible products we made, as our core networking gear (switches and routers) was typically hidden in wiring closets.

At the time, most of 3Com’s R&D was handled by H3C, the former Huawei enterprise division we had acquired. Since the technology for building a desk phone was straightforward, our main goal was to introduce a fresh industrial design. We informed the engineering team that they would handle the tech in-house, but we planned to hire a Western design firm to create the phone's physical look.

A few days later, our China engineering team surprised us by presenting a fully designed desk phone—complete with what they claimed was a "new" industrial design. The problem? It looked almost exactly like the existing Cisco desk phone.

When we asked why, they didn’t see an issue. Their logic was straightforward:

"Why waste money on a design firm when we can just copy a design that customers already know and accept? It saves time, reduces risk, and frees up resources for true innovation elsewhere."

Our Western perspective saw this as unacceptable—a blatant copy of a competitor's product—but from their point of view, it was simply smart business. This cultural difference is key to understanding China’s approach to innovation.

Understanding the "China Innovation Method"

This experience at 3Com perfectly illustrates the fundamental principles of China's innovation method:

1. Everything that can be copied will be copied. Any "accessible" ideas or technology—whether from Western firms or local Chinese competitors—are fair game. This isn’t an anti-Western stance; it’s how Chinese companies compete with each other as well.
2. Copying isn’t the end goal—it’s a starting point. Many assume China simply copies without innovating, but that’s a misunderstanding. Copying gives them a "free" head start, allowing them to refocus resources on rapid iteration and competitive innovation.
3. Speed, execution, and efficiency are the real competitive advantages. Since its assumed everyone can copy, success depends on how quickly and efficiently you can improve on the base model. This makes Chinese companies highly competitive against their Western counterparts.
4. Customer-driven development dominates. When I spoke to my Western engineers, they typically described the technology they were working on. In contrast, my Chinese engineers started with customer use cases and financial goals. This approach ensures that resources are focused on real-world applications, though it can sometimes limit free exploration and breakthrough innovations.

Competing Against the "China Innovation Method"

For Western companies, competing against this approach requires a clear-eyed strategy:

• Assume that anything that can be copied, will be. If they can get access, it will be used to the maximum extent, regardless of perceived IP protections.
• Expect rapid innovation on top of what is copied. Chinese companies will aggressively iterate, refine, and adapt based on customer engagement and market experimentation. Early iterations may have security and quality issues, given the focus on speed and efficiency, but these will be addressed over time if they affect customer adoption.
• Core innovation that is not tied to immediate customer requirements remains a Western advantage. The Western model encourages exploration beyond customer-driven needs, leading to breakthrough technologies. In this area, Western firms still have the upper hand.
• Educate customers on the risks. Customers should be made aware of potential IP issues, security risks, and quality concerns in hyper-competitive Chinese products.
• Stay paranoid and move fast. As Andy Grove famously said, “Only the paranoid survive.” The best way to compete isn’t through legal complaints—it’s through relentless innovation and execution. I found there was a lot to learn from parts of the China method, especially the relentless focus on customers by engineering and overall innovation velocity.

Final Thoughts

The emergence of DeepSeek is just the latest example of China’s well-honed innovation playbook in action. Understanding the China method is critical—not just for AI, but for any industry that expects to compete in a global market. The key takeaway? Assume they will copy. Expect them to improve. Be faster and smarter, and leverage your core research.

#DeepSeek #AI #China