Altera went from a competitor to an also ran.
I am sure they will produce some valuable custom silicon to help intel milk their server line but gone is any market presence. But if intel silicon is so great why are they so late and how did Xilinx clean up when intel-altera was so good?
I am sure bozo has spun it differently on circuit for the koolaid drinkers
https://www.semiwiki.com/forum/content/6602-xilinx-vs-altera-update-2017-a.html?s=cf42f41518728fb7beab7b536113e444
The long game: Split off foundry from product just like AMD did, and merge foundry with TSMC. Therefore can't make foundry business too successful or government wouldn't allow.
Foundry was the biggest smoke and mirror job at TMG. I'd be embarrassed to be associated with the thousands in ICF that have been nothing but a money drain.
Foundry is about offering a rich set of IP and other collateral that offers a compelling alternative to what is out in industry.
So what is Intel's value proposition?
They did have leadership on 14nm high performance logic, but that went away with TSMC and Samsung now shipping 10nm. TSMC has a richer set of IP and easier to design in. Companies can turn designs far faster in TSMC and Samsung than intel. Also Samsung and TSMC are more a dedicated foundry with huge capacity while at intel you can never be sure if your silicon is high priority as x86 pays all the bills. At TSMC you are top dog. Also people who need and can afford the performance and cost of leading edge silicon are for the most part direct competitors of intel. Why would AMD, Apple, Nvidia, Qualcomm consider intel. A customer like Apple looks interesting but the capacity they need exceeds all of Intels total 14nm. They would demand a price of about 30 dollars as well which is far less than intel ASP for cpus.
The only s---ers for intel foundry are smalltime Far East companies who will try and get some freebies but don't take too much leading edge capacity. The fact they are small means they bring nothing to the table IP wise. When we hear about them it is all fluff with no meat. Lots of noise was made about customers on 22nm, don't hear much about any revenue there do we!
Foundry is an all in affair. If you aren't all in you won't get customers. The fact BK can still fool himself and Wall Street is most frightening for the intel employees.
From the analyst meeting transcript (pp38) : BK regarding foundry:
"Our approach is very strategically specific. Well certainly not positioning our Foundry
ambitions to be a general purpose motion Foundry." ...
"I think we're really looking at how we can strategically partner to improve the [depth] and breadth of our IP portfolio" ...
"we're looking at this very strategically and how our Foundry customers can provide indigenous benefit. So the products are also developing internally."
in other words - we are not going to share our design cell libraries with anyone really but we will gladly leave the door open to acquire others' IP blocks if they are stupid enough to fall for it. More for us to liquidate and shuffle around later when we need to distract wall street.
@LX3dPI2-lbm You make an interesting observation about intel process.
On the surface Intel was leaders holding up to a two to four year lead over the big foundries. What is little known, the cost of catering to the narrow x86 design community created a process and design infrastructure totally restrictive and unsuitable for wide spread use by anyone but intel x86 team. Even the effort of thousands of ICF failed to create a infrastructure that made intel foundry competitive even with that process lead. Says a lot how backwards intel infrastructure is.
Things were so bad they had to buy customers to get any real designs not x86.
Perhaps they have now opened their thinking but it is too late as TSMC and Samsung have now surpassed intel in capability, opportunity lost due to arrogance and incompetence!
I may still be a koolaid drinker - but I still believe Intel has the best process technology of anyone in the world.
What I don't understand is why Intel never leveraged their technological advantage there and at least made a real effort into the foundry business. Instead they did s--t like buy McAfee.
Remember Altera was supposed to be Intel's first "foundry" customer - instead they bought them. As someone who specs FPGAs now - Altera is not even a consideration anymore.
Sadly as the Moore's law curve is flattening out - Intel's competitive-advantage-gap in process and fabrication technology is also closing.