Zhongwei Precision Instruments Pushes Silicon Carbide Cutting Loss Down to 40 Microns: How Breaking Process Limits Rewrites Chip Manufacturing Cost Structure

Event Background

Shenzhen Zhongwei Precision Instruments Technology recently completed a multi-million yuan angel round of financing, exclusively invested by Skyworth. The company was founded in November 2024 and focuses on advanced industrial equipment including SiC ingot laser exfoliation and diamond laser precision machining. Its core breakthrough is: using programmable pulse and crack-direction-guided growth technology to control the laser kerf loss in SiC substrate laser slicing to under 40 micrometers, a reduction of over 50% compared to the industry standard of 80-120 micrometers.

Founder Chen Jingyu comes from the University of New South Wales in Australia and has long engaged in specialized chip integration; Chief Scientist Sun Hongbo is an academician of the Chinese Academy of Sciences, professor at Tsinghua University, and a pioneer in nonlinear laser ultra-precision manufacturing. The team's disciplinary background spans materials science, laser physics, and chip engineering—a combination that itself signals this is not one-dimensional refinement, but normalization-level reorganization.

Deep Logic of Process Bottlenecks

Silicon carbide (SiC) is a third-generation wide-bandgap semiconductor material, widely used in power devices for new energy vehicles, photovoltaic inverters, and other high-frequency, high-power applications. However, SiC hardness approaches that of diamond. When cut with traditional mechanical grinding wheels:

1. Large loss: Each cut produces an 80-120 micrometer "waste layer" on both sides of the ingot—material that is ground away without producing chips 2. Crack propagation: Mechanical stress induces microcracks requiring additional repair processes 3. Low yield: After multiple cuts, a 300mm ingot achieves only 60-70% effective yield

This bottleneck has persisted for over 10 years. The industry has continuously pursued microscopic optimization through "finer wheels" and "more precise feed rates," but none have broken through the 80 micrometer ceiling. The reason is simple: the loss inherent in mechanical cutting is fundamentally determined by grain size and cutting width, and cannot violate physics.

Normalization Shift: Laser Replaces Mechanical

Zhongwei Precision Instruments' breakthrough lies in changing the mechanism of action. The logic of programmable laser pulse technology is:

1. Thermal stress-induced cracking: High-power laser pulses create precise temperature gradients inside the SiC ingot, inducing ordered cracks at preset locations 2. Crack direction-controlled growth: By adjusting pulse width, power, frequency, and other parameters ("programmable"), control the crack propagation along the crystal's weakest direction 3. Separation rather than grinding: Finally, a gentle break along the crack separates the material, with loss requiring only a 40 micrometer buffer layer

The fundamental difference in this approach is: no longer fighting material hardness, but leveraging its internal structural weaknesses. This is a shift from "mechanical thinking" to "thermal + materials science thinking."

Why It's Difficult to Replicate

Everyone understands the broad direction of "using lasers to cut," but Zhongwei Precision Instruments' moat lies in:

1. Nonlinear laser physics: Pulse shape and frequency design involve deep laser physics knowledge. Sun Hongbo's 30-year accumulation in this field cannot be made up in a few years 2. Materials-process coupling: For different SiC crystal orientations and doping concentrations, the threshold for crack growth varies completely. This requires extensive experimental data accumulation 3. Engineering integration: The gap between laboratory lasers and batch processing of 300mm ingots on production lines encompasses temperature control, vibration isolation, automation, and other engineering challenges

Supply Chain Ripple Effects

Upstream (laser equipment manufacturers): - The traditional laser cutting machine market is relatively saturated with 20-30% margins - Zhongwei Precision Instruments' specialized laser control system requires customized pulse management chips, potentially driving demand for high-end lasers

Downstream (SiC ingot manufacturers, chip producers): - Yield improvement from 60-70% to 80%+, directly reducing chip costs by 15-20% - As SiC chip costs decline, the BOM cost of new energy vehicle power modules will be rewritten, potentially accelerating SiC penetration rates (currently only 5-8% globally)

Competitive landscape: - Companies mastering efficient cutting processes will become "hidden champions" in the SiC supply chain - Skyworth's investment may represent a layout around "chip cost competitiveness" as a long-term issue

Why This Is a Textbook Case of Process Limit Breakthrough

This event exhibits the three typical characteristics of process breakthroughs:

1. Old process approaches physical limits: The loss in mechanical grinding is determined by grain size; 80 micrometers is not engineering failure but physical necessity 2. Breakthrough comes from normalization shift: Not "better wheels" but "lasers replace wheels"—the mechanism of action fundamentally changes 3. New normalization has different cost structure: Laser equipment has high upfront investment, but unit cost (per piece) is lower, reorganizing the competitive foundation of the supply chain

This is precisely where Schumpeter's "creative destruction" and Christensen's "disruptive innovation" overlap: technology is not a long gradual slope, but a sudden leap at the limit.

Open Questions

• Is 40 micrometers the new limit of laser processes, or is there further crack control potential (such as 20 micrometers)?
• When SiC cutting costs drop significantly, who will be the biggest beneficiary—upstream ingot manufacturers or downstream chip design companies?
• Will Zhongwei Precision Instruments become a process platform company (like ASML in lithography), or ultimately be acquired by or face competition from larger firms?