虎嗅

How long does it take from the production of a lithography machine to its actual use?

原文:光刻机从造出来到用起来有多远?

Summary of Key Points

An anonymous report about a Chinese state-owned company's small-scale production of immersion DUV lithography machines has caused a sharp decline in the semiconductor sector of the U.S. stock market (especially ASML, whose stock price fell by nearly 6%, resulting in a loss of hundreds of billions in market value), while A-share stocks related to lithography machines have risen against the trend. By analyzing the credibility of the news, the logic behind market reactions, and the actual progress and challenges of domestic lithography machine development, the report reveals a core contradiction: what the market is pricing is not the current production capacity, but the expectation that "domestic substitution is no longer a question mark, but a matter of time." The U.S. stock market is concerned about the shortening of ASML's monopoly period, while A-share investors are betting on opportunities for domestic orders.

1. Is This News About Domestic Lithography Machines Reliable?

The report comes from the American technology media outlet The Information and claims that a Shanghai-based state-owned company plans to produce 5 immersion DUV machines this year and 20 more next year, to be delivered to companies like SMIC. However, there are three key doubts:

1. Lack of Official Confirmation: All information comes from the same anonymous source; ASML has not confirmed these plans, and the Chinese side has also remained silent, without even mentioning the company's name.

2. Many Past False Reports: In May this year, a report claiming that Shanghai Microelectronics had delivered 28-nanometer lithography machines in bulk was debunked (the so-called summit never took place, and the equipment was not displayed), and there was a correction of the misconception about "lithography machine yield" – yield refers to the percentage of chips that are qualified, not a measure of the machine itself.

3. Supply Chain Dependence: The report acknowledges that "some key components are from Japan," indicating that these 5 machines do not represent a completely independent production line.

In short, while the news sounds plausible, it lacks official confirmation and appears to be more of an "exploratory leak" within the industry.

2. Why Did ASML's Stock Drop So Sharply? What Is the Market Worried About?

On the surface, the production capacity of 5/20 machines is negligible compared to ASML's annual shipments of 130 (according to Bank of America estimates, 20 machines would only affect 2.4% of ASML's sales). However, the market is concerned about the "future":

1. Monopoly Premium Erosion: ASML holds a 98.7% market share in immersion DUV lithography and is valued with a premium for its perceived permanent monopoly. This report makes investors realize that this monopoly may be temporary.

2. Climaxing of Regulations and Catch-Up Efforts: The Netherlands will ban the sale of advanced DUV machines in 2023, and the U.S. also plans to restrict sales (lithography machines require maintenance every 4-5 years, and the restricted equipment will gradually become obsolete). The timeline for domestic production (targeting mass production by 2027) coincides with these regulatory changes, suggesting that China may no longer need to buy from ASML in the future.

3. Market Bet on Probability: Stock prices reflect expectations. Even if only 50% of the news is true, any increase in the likelihood of domestic substitution reduces ASML's long-term value.

3. How Far Are Domestic Lithography Machines From Being "Useable"?

Even if all the information is accurate, there are still four major hurdles before these machines can be mass-produced:

1. Certification Time: After the machines are delivered, they need to be installed, calibrated, and integrated with production processes (a process called "certification"), which takes months.

2. Capacity Gap: ASML produced 130 machines this year and plans to increase capacity by 30% in 2027; domestic production of 5 machines this year and 20 next year is still a very small proportion.

3. Efficiency and Cost: ASML's flagship models can process over 300 wafers per hour, while domestic machines are significantly less efficient. Producing 7-nanometer chips will require multiple exposures, increasing costs and lowering yield (TSMC took more than a year to improve its yield for 28-nanometer chips, and new domestic equipment will face a similar challenge).

4. Supply Chain Dependence: Key components still rely on Japan, and delays from local suppliers have affected production this year.

In summary, these 5 machines represent a breakthrough from "0 to 1," but there is still a long way to go before reaching "100."

4. Why Did the U.S. and Chinese Stock Markets React Differently? What Are Their Different Calculations?

The same report has led to contrasting reactions in the two markets:

  • U.S. Stock Market Drop: Investors are concerned that ASML's monopoly is weakening, which could affect future revenue growth.
  • A-share Market Rise: Investors are seeing opportunities for domestic orders; as long as there is an expectation of domestic substitution, related companies (such as Zhangjiang High-Tech and its affiliated enterprises) may receive orders, driving up their stock prices.

Essentially, both markets are betting on the timeline for domestic substitution, but from different perspectives: the U.S. market is concerned about ASML's past monopoly, while the A-share market is optimistic about the potential of domestic alternatives.

5. The Significance of Domestic Lithography Machines: A Milestone Is Not the End

The value of these 5 machines lies not in their ability to replace ASML's production capacity, but in:

  • Demonstrating that China has overcome the technical barrier of producing immersion DUV machines.
  • Providing domestic chip companies (such as SMIC and HuaHong) with an alternative to foreign imports.
  • Showing the global market that China's semiconductor autonomy is not just a slogan but a reality being implemented step by step.

It is important to remain realistic: mass production by 2027 and reaching commercial scale by the mid-2030s are still distant goals. Current stock price fluctuations are largely driven by sentiment, and ultimately, the reality will be determined by actual production data (certification, yield improvement, and capacity increase).

In conclusion, while domestic lithography machines hold promise, we should not treat this milestone as the end of the journey.

(Note: This article does not constitute investment advice; all information should be confirmed by official sources.)