Breaking News
News & Analysis

Sondrel’s CEO Takes the Long View

IC design house keeps ‘global’ perspective as it staff up in Europe, says Curren
9/25/2017 00:01 AM EDT
More Related Links
View Comments: Newest First | Oldest First | Threaded View
User Rank
Re: Strong opinions - not necessarily correct ;-))
realjjj   9/25/2017 7:08:59 PM
Hi, maybe the timing is difficult for some of those.

While advanced packaging has its upsides, a lot more is needed and i see it as more of a stopgap solution.

Monolithic 3D should at least allow for costs to scale. Memory is going there already, NAND and SCMs (XPoint, 3D ReRAM) for now. A DRAM replacement will likely follow at some point.  AI kinda fits memory like 3D structures and its importance can only grow. So right there, with memory and AI, you have a huge chunk of the market going 3D- the probability is pretty high for this to happen in 5-10 years.

For general purpose, it's difficult to devise a 3D structure where costs scale well on the vertical, hard to get to the bottom electrode but what's the industry gonna do, get stuck with huge efforts for minimal scaling? It will likely start with a lesser approach like Leti's CoolCube and evolve from there.  Glasses likely require it for the form factor to become ubiquitous as much higher efficiency is needed and very cheap transistors would help with that.How long it takes, that is hard to be certain but it will take a bit longer. It's possible that something else solves the scaling problem but ,at this point in time ,it seems the most likely solution.

Glasses can offer (practically) a full field of view experience and an always on display and that's a lot more than what any other display offers. You have AR/MR/VR and that opens up all kind of cool opportunities. How about an AR chemistry app where kids can experiment with zero costs and risks. Simulators will be a huge opportunity in general - an example For media and gaming there is no contest. For work, you can use both physical and virtual peripherals - a rudimentary example . Imagine a workstation with 3-6 displays, would be great to replace all of those with a tiny pair of glasses with the added upside of a much lower cost while using a fraction of the power. The keyboard can be a normal wireless keyboard (tactile feedback = highest perf) , a virtual keyboard and ofc voice is always an option- a brain-computer interface comes later. For a virtual keyboard I think the threshold should be to at least match typing on a touchscreen. Won't be easy but 3D gesture tracking coupled with eye tracking and maybe some machine learning, should make it possible sooner rather than later. Facial recognition is just the start for 3D imaging. The form factor is very restrictive and progress in many areas is needed but the opportunities are astonishing.

PS: Eye tracking only keyboards are aimed today at people with certain disabilities but I wonder where such keyboards will be in 10 years from now.  In public places we might not want to type on a virtual keyboard or use voice.

User Rank
Strong opinions - not necessarily correct ;-))
sranje   9/25/2017 5:25:52 PM
Hi Realjjj -- perhaps your forecasted future is not all that certain.

Just one example - about 3D monolithic ICs.   SiP module approach is perhaps far more certain - including non-semiconductor components

Not at all sure about "everything being replaced by glasses" - including typing this response?

User Rank
realjjj   9/25/2017 8:53:01 AM
"automotive being one of the biggest ones, where the business that's driving it is really quite long term, and that's good. Industries like that aren't going to be overcome by investors jumping in and trying to make a mark. It's a much more stable, long-term industry ... There's an extremely long tail with automotive, and it's not susceptible to short product lifetimes.""  

That kind of thinking is what will lead to the end of many. Electrification will ramp much faster than people expect and will shuffle the ranks, destabilizing many. Autonomy brings car as a service, that destroys car ownership and the top 3-4 CaaS players will own the bulk of the global market.

The German (and global) auto industry will be obliterated and there is nothing anyone can do about it. VW is maybe ahead of others but they are severely underestimating the tsunami of change that's about to hit them. Maybe one German car maker is sold at a discount in 5 years from now and survives as a division of Apple or Amazon but the other two have a 3% chance of survival. Elsewhere in Europe, Fiat has 0% chances, in France maybe Renault has a 5% chance of survival while for PSA it's too late. This CaaS induced consolidation won't be kind on most suppliers either.

Server is in a bubble right now, it will last a while longer but the guys at the edge will always win.

Gaming might be relevant today as a PC niche but everything will be replaced by glasses - phones, PCs, TVs fade away- and that kind of consolidation is another bubble deflating.

ICs must go monolithic 3D and once there, embedded memory begins to replace discrete.

Not smooth sailing at all, it's much trickier than ever before and nobody is safe.

Like Us on Facebook
EE Times on Twitter
EE Times Twitter Feed