My scattered thoughts on today’s CMD (Capital Markets Day). I thought the analysts present did a great job challenging Canatu’s management, so big thumbs up for that. In 2024, Canatu reported the TAM (Total Addressable Market) for CNT pellicles to be approximately 3 billion euros; now, management has reduced the TAM to 1.5 billion euros. This raised a valid question about whether Canatu had been a bit overzealous in their initial estimates. Canatu also seemed somewhat surprised by how slow the transition in ASML’s machines has been (400W → 600W). This was used to explain why financial targets were pushed back.
I still believe there are relevant drivers for using Canatu’s CNT pellicle, even in, for example, 500W machines:
- 8-15% better transmittance than the current ‘metal silicide’ pellicle. This means the ASML scanner can expose wafers faster, which directly impacts fabs’ top line.
- CNT can withstand at least 10,000 wafer exposures (ASML recommendation), and according to Canatu, even up to 15,000 wafers. My own notes say 10k wafers, depending on the power.
- According to Canatu, by using their pellicle, if it’s estimated that a 10% better transmittance can be achieved compared to the current pellicle, 1/10 of the ASML scanner acquisition cost can be saved. This would result in approximately 180 million euros in CAPEX savings for the fab.
Regarding revenue development, it was commented that in the coming years, revenue is expected to be generated solely by the reactor business, and by 2030, the situation will shift so that royalties bring in the majority of the revenue. If this materializes, it sounds like that famous “Teukka Salama’s curveball,” but time will tell.
I have tried to estimate how long one customer can manage with one reactor, i.e., what is the throughput of one reactor. I believe this is one of the most important parameters when trying to project reactor sales. Unfortunately, management did not want to answer such specific questions and invoked trade secrets.
A couple of weeks ago, I found an article mentioning Mitsui’s pellicle transmittance for EUV light (>94%). I asked if Canatu had a similar figure and was told it would be a few percentage points better. I somewhat suspected this, as Mitsui has to use several post-process steps. At the same time, I inquired how many post-process steps Canatu has, and the answer was “fewer than 10,” which they still aim to reduce through R&D. In my notes, this emphasizes that Canatu or the customer can produce pellicles with good yield. The more post-process steps there are, the more residues remain already in the manufacturing stage, which must be removed. My own estimates suggest Mitsui might have as many as 20-30 post-process steps, but there’s no precise public number for this.
We’ll await Ate’s comments after the CMD, but Ate’s forecasts regarding the new financial targets were quite accurate.