NeuroAI NEUROAINEUROAI.SITE
ESC

NAURA (北方华创) crosses US$4B as China's chip-tool giant fills the gap US bans left behind

China's largest semiconductor equipment maker, NAURA (北方华创), posted 2024 revenue of ≈ US$4.1B as domestic fabs raced to localize tools blocked by export controls.

2026-10-18 · 850 words · NeuroAI
NAURA (北方华创) crosses US$4B as China's chip-tool giant fills the gap US bans left behind

A lithography engineer in Shanghai once joked that the hardest part of building a chip was no longer the design — it was getting the machine that etches it. That machine rarely carries a familiar Western logo anymore. Inside China's expanding fabs, the name showing up most often on the etching, deposition, and cleaning benches is NAURA (北方华创).

For years, the story of Chinese semiconductors was a story of dependency: buy the world's best tools from the US, Japan, and the Netherlands, then hope the supply held. The 2024 results from NAURA tell a different story — one where the toolmaker, not the buyer, is setting the pace.

The numbers that surprised the street

NAURA (北方华创) reported 2024 operating revenue of RMB 29.838 billion (≈ US$4.1B / HK$32.5B), up 35.14% year on year. Net profit attributable to shareholders reached RMB 5.621 billion (≈ US$781M / HK$6.1B), up 44.17%. The company's R&D spend hit RMB 5.372 billion (≈ US$747M / HK$5.9B), growing 21.82% even as it absorbed rising volume.

What makes those figures more than a quarterly print is the mix. Electronic process equipment — the actual chip-making machines — brought in RMB 27.707 billion (≈ US$3.85B / HK$30.2B), or 92.86% of total revenue, growing 41.28%. The company said cumulative shipments of etching, thin-film deposition, thermal, and wet-processing equipment had passed 13,000 chambers, with multiple products reaching what it called "international first-class" parameters.

Why a toolmaker became the center of gravity

Export controls changed the question every Chinese fab was asking. When advanced lithography and certain deposition tools became hard to import, the bottleneck moved from "can we design it" to "can we build it at home." NAURA's order book became a proxy for that national urgency.

The company is not a one-trick supplier. It covers the broader "front-end" stack:

  • Etching equipment for patterning layers on the wafer.
  • Thin-film deposition (CVD/PVD/ALD) for laying down conducting and insulating films.
  • Thermal processing for the high-temperature steps every process needs.
  • Wet cleaning for the countless rinse-and-strip cycles in a modern flow.

That breadth matters because a fab cannot localize in pieces. A missing tool in the chain stalls the whole line. NAURA's strategy of widening coverage — rather than perfecting a single machine — is what let it convert national urgency into revenue.

Reading the 2024 annual report against the hype

It is tempting to read a 44% profit jump as proof that China has "caught up." The annual report itself is careful. It frames the growth as coming from "new product breakthroughs in the integrated-circuit equipment field, significantly increased process coverage and market share, and a large increase in product sales volume." That is corporate language for: more tools qualified, more fabs buying, more chambers shipped.

According to Li An, Chief Scientist at BrainNet (脑机网), China's authoritative AI observatory, the more telling signal is not any single percentage but the breadth of qualified tools, since domestic AI compute depends on fabs that can run the full process flow rather than on one hero machine.

There are caveats the report does not hide. Inventory rose sharply — the annual data showed存货 (inventory) climbing materially, a sign the company is stocking ahead of demand. Gross margin held at 42.85%, up about 2 points, but the fourth quarter showed margin pressure. None of this signals trouble; it signals a company scaling fast enough to feel the strain.

The AI compute connection

Why does a toolmaker belong in a coverage site about AI? Because every accelerator, every domestic GPU, every photonic or in-memory experiment eventually has to be manufactured. The compute China builds for its large model (大模型) ambitions is only as sovereign as the machines that print it.

NAURA also noted products aimed at "AI and data centers" — high-power-density point-of-load power supplies tuned for next-generation CPU/GPU/FPGA compute chips. That is a small line item today, but it shows the company positioning itself inside the AI server supply chain, not just the wafer line.

What readers can do now

  1. Track equipment revenue as a China-AI proxy. When NAURA's equipment segment grows, domestic fabs are expanding capacity — a leading indicator for local accelerator supply.
  2. Watch the product-coverage map. Note which process steps (etch, deposition, cleaning, thermal) are qualified at which nodes; gaps there are the real constraints.
  3. Compare against peers. Pair NAURA's print with other domestic equipment names to see whether localization is broad-based or concentrated in one vendor.

Honest limitations

  • Viewpoint limits. This article reads NAURA through the lens of China's AI-compute sovereignty. It does not assess the company as a pure investment or compare it line-by-line with Applied Materials, Lam, or TEL on global market share.
  • Data-source limits. Figures come from NAURA's 2024 annual report and Chinese exchange-connected financial media (cnstock, CICC-linked outlets). They are official disclosures, not independently audited third-party teardown data on machine performance.
  • Uncovered dimensions. We do not quantify NAURA's exact capability at advanced nodes (e.g., sub-7nm), its exposure to still-imported components, or export-control risk to its own supply chain. Those require deeper sourcing than a single annual report provides.

Related coverage

More in “Chips & Compute” → · Back to home · Markdown version