ASML CEO: “We’ve structurally underestimated growth in the industry” – Bits&Chips (bits-chips.nl)
Their estimation sucks. I think the major chipmakers are themselves to blame for letting one company become a roadblock in their progress.
They are lucky someone was willing to invest in EUV, and that basically only happened because several foundries took a stake in ASML to provide them with the necessary funding. A lot of very knowledgeable people believed EUV would never be viable, and it took many years longer than its proponents envisioned before it became capable of reaching production volumes.
Compare it with 450mm wafers - over a decade ago there was talk about switching to 450mm wafers, but few of the companies that make fab equipment were willing to make the necessary investment. In this case it wouldn't be just ASML but every company (there are more than a few) that produces any equipment that handles wafers would need to make the investments to make 450mm wafers happen. Problem was, the efficiencies would be worth it only for a half dozen or so companies operating at the necessary volumes so they would have to produce special versions of their equipment to handle 450mm wafers but have a fraction of the customer base compared to their 300mm and 200mm based equipment. So despite companies like Intel and Micron REALLY wanting 450mm wafers, they never happened.
If the foundries wanted a second source for EUV they would have had to fund that second source up front like they did with ASML, and hope one didn't get out to such a big lead that the other couldn't compete. They didn't, because it would cost twice as much to do so - and perhaps they feared their competition might be the beneficiary of that extra supply.
Right now we have only three foundries pursuing lead edge design (I'm still counting Intel, we'll see how long they stay in the game - they will have to become a legitimate foundry or they simply won't have the volume to keep up) and one of them has a pretty significant lead on the other two as far as number of leading edge wafers manufactured at each generation. It is quite possible in a decade we could have only one left standing that still pursues the leading edge. Then I imagine you'll say "the major fabless designers are themselves to blame for letting one company become a roadblock".
While you wouldn't be wrong, if you (and Apple, Qualcomm, AMD, Intel, etc.) think this matters, the time to do something about it is now. Arguably Qualcomm is by patronizing Samsung's fab, though that may be more due to lower cost and better availability and may not be simply because they want to financially support them to keep them around. Intel is shooting themselves in the foot buying leading edge TSMC capacity, though at this point they may have no choice - but if they are still behind a few years from now and have got into the habit of buying leading edge share from TSMC for fear of losing market share to AMD it may become a self fulfilling cycle after a while.