Summary of Key Points
Spuruiyuan is a startup company focused on industrial X-ray inspection. Founded by Wu Mingwei, who has observed the industry for 20 years, the company started with a small team during the pandemic. By utilizing a technical approach that combines "advanced functions (complex functions, Fourier transformations) and AI," they have solved a detection problem with molybdenum-containing materials that foreign equipment could not address. The company's revenue grew from 1.55 million in 2021 to over 100 million this year, making them a leader in the field of industrial inspection within the context of the new AI infrastructure.
I. The Core of an X-ray Machine Isn't Hardware, but Mathematics: The Magic Combination of “Advanced Functions and AI”
Many people think that X-ray machines rely on hardware (such as radiation sources and detectors), but Wu Mingwei believes the key lies in mathematics. For example, X-rays from hospitals show all organs layered together, while CT scans can reveal layers by using Fourier transformations to separate them. This principle is also applied in industrial inspection; to detect defects in solder balls on circuit boards, clear internal images must first be obtained through mathematical methods, and then AI helps improve detection accuracy.
- Advanced Functions: Fourier transformations are fundamental; without them, it would be impossible to obtain usable internal images (just as noise needs to be filtered out in a noisy environment to hear clearly).
- AI: Once the images are obtained, AI can identify faint defects (such as welding bubbles) with accuracy increasing from 50% to over 99%.
In contrast, foreign equipment uses only functions without AI, requiring manual image analysis; domestic competitors use only AI without the underlying mathematical methods, failing to produce clear images. Spuruiyuan’s approach combines both, enabling fully automated inspection.
II. Why Can We Detect Molybdenum-Containing Materials When Foreign Equipment Fails?
Molybdenum is commonly used in semiconductors (e.g., in the heat dissipation layers of drone infrared chips), but it absorbs X-rays strongly, making them appear dark on foreign equipment. Spuruiyuan’s success is due to a set of digital filters added to their custom software:
- Molybdenum emits a weak signal after absorbing X-rays, which is filtered out by common software (e.g., VG Studio); Spuruiyuan’s software amplifies this signal, allowing them to see the internal structure.
Hitachi has verified this with ultrasound, and the results matched Spuruiyuan’s X-ray images, becoming their unique technical advantage.
III. 20 Years of Observation and the Pandemic: The Journey from 1.55 Million to Over 100 Million in Orders
Wu Mingwei’s entrepreneurship wasn’t impulsive:
- 20 Years of Observation: He saw 3D X-ray machines in European and American customers’ facilities in 1998 and understood the principles behind Fourier transformations, focusing on this industry ever since.
- The Pandemic Turning Point: At 47, he was comfortable in a foreign company but decided to change his life during the pandemic. He assembled a small team of software engineers and experts from various fields to start the business.
- Critical Moments: The first year’s revenue was 1.55 million, and maintaining the team cost 5 million; they survived thanks to unexpected support from friends and investors.
- Order Surge: Companies like C (chip packaging) and M (NVIDIA power supply supplier) chose Spuruiyuan due to foreign equipment’s limitations, leading to orders in the tens of millions this year.
IV. AI New Infrastructure Is Booming, and They’ve Become Leaders in Inspection
The new AI infrastructure (e.g., GPUs, AI servers) requires extensive testing:
- For example, a damaged power management module in a GPU can cause significant losses, so rapid inspection is essential.
- PCBs (essential for AI servers) were previously neglected but are now in high demand.
Spuruiyuan doesn’t just sell components; they provide quality inspection services. The surge in demand for AI infrastructure has driven their orders upward.
V. Future Goals: Billion-Dollar Markets and Advanced Technologies
Wu Mingwei’s next targets include:
1. TSV/TGV Inspection: When chips are stacked vertically, vertical “elevator shafts” need to be created on silicon/glass for signal transmission; no domestic equipment can perform this, while foreign NXP equipment costs $2 million per unit. Spuruiyuan plans to develop their own prototype.
2. Advanced PCB Inspection: A company that supplies PCBs for NVIDIA ordered 150 X-ray machines last year; Spuruiyuan’s technology gives them a competitive advantage.
These markets are worth billions, and Spuruiyuan has the technical edge because they understand the underlying mathematics, solving problems others can’t.
Final Words: Advice from an Entrepreneur
Wu Mingwei advises entrepreneurs to:
- Maintain a positive mindset: Don’t rely on success formulas; be resilient to setbacks.
- Think broadly: Learn from multiple disciplines and unite experts to solve problems.
- Love life: Engage in hobbies (music, calligraphy) to relax under pressure, which can inspire creativity.
AI won’t replace industrial inspectors; instead, it will create more data labeling jobs. The core of industrial inspection remains basic mathematics.
This story is inspiring: an ordinary person achieved success through 20 years of observation and dedication to mathematics, seizing the opportunity in the new AI infrastructure era. It shows that sometimes the most critical technology isn’t high-tech AI but often overlooked fundamental disciplines.