True, but isn't this a "damned if you do, damned if you don't" kind of scenario?
Applied and Tokyo don't think it is possible to keep the pace of developing and delivering tools in a timely manner to enable semi manufacturing at the pace expected by their customers (I mean the fabs). So without merging, they are already saying "too expensive, we can't keep this pace up solo / fractured market share, we need the profits of practically the entire industry otherwise research and development here will slow down due to money."
Then of course the risk there is as you mention, there will be one less entity developing the tools, and the customers (fabs) will have have no choice but to effectively put their eggs in one basket.
What serious alternative from those two can you see? R&D is all about money, and they need more money to keep up the pace as the tech keeps on becoming more complex, and more money will only come from consolidating the profits to be had (unless the number of fabs worldwide double or triple soon, and that's even more unlikely to happen than a merger). If their backs are against a wall now, what else could they do?
This isn't rhetorical. I'm actually wondering what other possible alternatives they have. It looks to me like they either:
1.) accept that progress will slow down, and remain separate entities, or,
2.) attempt to continue the pace of R&D and new products as desired by their customers by merging and thereby consolidating all profits to effectively have more money to burn for R&D
Now, if you say their initial premise (they absolutely need more money because litho R&D is getting incredibly more and more expensive) is bunk, then that's a different matter, and not one I can evaluate for myself with any authority.
The model is broken, that if for sure. The problem is "risk management for joint development projects".
To understand why TEL and AMAT want to merge requires a historical perspective of the business they operate in, it is not about money and at the same time it is about money.
I'll speak from experience, the tools I was interested in cost $3m-$4m each and I spent three years evaluating three separate tools (one from DNS, one from Semitool, and one from SEZ).
Now the accounting went like this - the tool suppliers had to pay my employer the installation fees, the maintenance fees, and the removal fees as incurred by the tool while it was in our fab for the duration of the evaluation. Oh and the tool itself was free, zero cost to us for the tool itself. (first one is free, the next 19 we buy would be purchased but at a volume discount, naturally)
What did that look like to the bottom-line for the tool supplier? They had to front 100% of the costs to build the tool, front the costs to ship it and install it in our fab, hire/relocate a team of upwards of 6 field technicians to service the tool and support the JDP (joint development project) for a minimum of one year, and because only one of the three tool vendors was going to win the contract at the end of the day, two of the three were going to have to front the costs of removing the tool from the fab (and then try to sell it as a refurbished tool on the 2nd hand market).
That was a HUGE cost to swallow. But why did they agree to it?
Competition.
The SEZ guys only agreed to it because the DNS guys agreed to it, who only agreed to it because the Semitool guys agreed to it, who only agreed to it because we told them during negotiations that they either accepted our terms or they accepted they would lose out on a contract that could carry upwards of 20+ tools during the node build-out phase for production.
The customer (me) had all the odds in their favor, all the economic risk was borne on the shoulders of the vendor companies.
That is/was the part of the model that was broken. And it is because that part of the JDP model (which is a lopsided risk-sharing model) that has driven companies like TEL and AMAT to the point of merging into giant beasts just to be able to finally force some balance back into the negotiations of how JDP's and tool development transpires.
(more history - Semitool went on to be bought out by AMAT, and SEZ went on to be bought out by TEL)
If there was more risk sharing and cost sharing then the R&D aspects of tool development wouldn't become a "go bankrupt while trying" type situation.
But the equation is lopsided, so these guys have to try and grow themselves to become larger than the risk potential of a tool project turning out to land jelly side down.
The downside for the fab guys is that they have less diversity to rely on. It was great 4-5 yrs ago when you could have 3 or 4 tools from the competition to compare side-by-side. The odds were in your favor that someone somewhere had developed a tool that fit your needs perfectly.
Now you don't have those odds, so odds are higher now that you'll find yourself up a creek without a paddle and left with no choice but to reward incompetent project management (the guys who can't develop a usable toolset to save their life) with oodles of more cash in a crisis-mode effort to re-engineer their derp-tool to get the job done.
So this will push the risk balance in the other direction, now the Intel's and the TSMC's will have no choice but to basically develop the tool for the vendor or fail trying. The will have become too critically inter-dependent.