News ASML teams up with Imec for sub-2nm process technologies with High-NA EUV chipmaking tools

[Admin]

ASML to install its High-NA EUV lithography tools in Imec's pilot production line to give research and development personnel access to leading-edge equipment.

ASML teams up with Imec for sub-2nm process technologies with High-NA EUV chipmaking tools : Read more

 

[shady28]

"However, each High-NA EUV system costs $350 billion..."

Should read million, not billion.

 

[acadia11]

"However, each High-NA EUV system costs $350 billion..."

Should read million, not billion.

I read that and thought ... so basically the US would spend 10% of it'a annual budget just on a single High NA EUV system .... or half of what it spends on military each year. new players there would be like 3 countries in the world that could buy these machines if that were the case, let alone new players or academic institutions. The ROI would take how long ....

 

[Roland Of Gilead]

"However, each High-NA EUV system costs $350 billion..."

Should read million, not billion.

Your right. But, what stumps me is, why are ASML not charging more for these machines. Yes, they are absurdly expensive to begin with, but what would the likes of TSMC, Intel or any other fab do if ASML said they are $1 billion each? Those companies would have no choice to pay it, given how essential they are in the those companies producing chips in massive quantities, and the multiples of billions they make from it.

 

[thestryker]

Your right. But, what stumps me is, why are ASML not charging more for these machines. Yes, they are absurdly expensive to begin with, but what would the likes of TSMC, Intel or any other fab do if ASML said they are $1 billion each? Those companies would have no choice to pay it, given how essential they are in the those companies producing chips in massive quantities, and the multiples of billions they make from it.

They wouldn't buy them until they absolutely had to. Just look at the adoption of the High-NA machines at their current pricing. Intel is the only company buying a quantity of them that can deliver volume manufacturing. Despite ASML being a monopoly they do seem to be run by people who want to keep making money long term and fleecing your customers is a good way to get them to invest in competition.

 

[thestryker]

While I'm sure this is costing ASML more up front than they're getting out of it the long term benefits should more than pay for themselves. It sure seems like the cost to design these High-NA machines must have been immense. They've been making a lot of moves that aren't necessarily peak capitalist in the name of just getting them out there.

 

[jp7189]

Your right. But, what stumps me is, why are ASML not charging more for these machines. Yes, they are absurdly expensive to begin with, but what would the likes of TSMC, Intel or any other fab do if ASML said they are $1 billion each? Those companies would have no choice to pay it, given how essential they are in the those companies producing chips in massive quantities, and the multiples of billions they make from it.

The relationship is more complex and interdependent than a simple vendor/client. The success of the fabs means literally everything to ASML. Also, the fabs contribute significantly more to ASML than the simple cash of a purchase.

 

[shady28]

Your right. But, what stumps me is, why are ASML not charging more for these machines. Yes, they are absurdly expensive to begin with, but what would the likes of TSMC, Intel or any other fab do if ASML said they are $1 billion each? Those companies would have no choice to pay it, given how essential they are in the those companies producing chips in massive quantities, and the multiples of billions they make from it.

Every product has a sweet spot of maximum profitability on the supply / demand / price curve. The higher the price, the lower demand will be. The lower the price, the higher the demand. Important point here is that if you can't afford something, you are not part of 'demand'.

Example of this would be, say you need and can afford at most a $30,000 new car. You would like to have a Toyota Land Cruiser, which starts ~$61,000. You are not part of demand. Now, if they lowered the price of the Land Cruiser to $30,000, you would be part of demand. Along with a lot of other people.

So back to the subject at hand, with a price tag of $350M I can say that companies like UMC, GloFlo, ONSemi, and probably HP are not going to be part of demand. Demand is pretty much just going to be Intel and TSMC.

So they won't be able to sell very many of these machines.

To use your example, if they increased it to $1B each, you create a downstream issue. TSMC and Intel might buy some still - but not as many. They will have to push the cost to their customers, and as a result their customers will decrease in number, hence demand will decrease too. Then it becomes a specialty item, for example just for military or high end research, and you lose economy of scale. Instead of selling 50 over the next 5 years to Intel and TSMC, you wind up selling 2 or 3.

Manufacturers have to know if there is a market for their product, how big it is and how valuable their product is. Most of these companies have people who do nothing but study this kind of stuff. I'd bet ASML uses a consultancy since they don't come out with new stuff all the time like say Sony or Ford.

 

[joeer77]

Good for IMEC. I've been there twice. They are research university. They have direct collaboration agreement with Intel. This move will benefit Intel too by accelerating EUV learnings. I'm sure ASML gave IMEC a discount on the tool if the Intel tool cost $500M.