What is the value proposition?
Cutting cycle times down from days to hours, really improving designer productivity overall, letting them do multiple evaluations so that they can get a higher quality product. But it is also a resource management tool because where XtremePCB allows teams and managers to efficiently utilize, to have as many people as possible on a critical design and offloading so that people can pulse in and out of a project. With this tool you can have a designer using all the unused CPUs. For example, people working in the UK getting things done but the CPUs in the US are unused because folks are still sleeping. They can flip it over when the opposite is true. Or they can do it within a single location when everybody goes home in the evening and the jobs are automatically kicked off. It is really leveraging existing hardware in an organization. Because hardware can be cheap relative to the end product you are trying to make, we foresee some folks deciding to implement CPU farms and really leverage this technology.
We have examples of cutting run times from 5 hours to 1 hour and from 2 days to 18 hours. This is clearly a case where a layout designer had to sit and wait for a couple of days for a route. Now they can punch a button and come back the next day and see what the results are and make better choices and move on from there without having to wait too long before getting their results back.
I would note that there are classes of boards where this technology is not applicable. First of all if an autorouter can get a board done in one hour, applying multiple CPUs to get it done in 10 minutes probably doesn't have as much value. The guys who are routing multiple days tend to see the most value in this thing. The other cases are boards that may route in 10 minutes. If you apply this technology because it is distributing across the network, it may take 20 minutes to route. Not a good application for this technology. We are looking for the monster boards to apply this technology.
Any customer experience with this tool?
We have had the product out in controlled release for a few months so we could get some customers hammering on it and determining the value of it. In this case the customer applied it to a large optical switch card of 36 layers. The board is roughly 0.5mx0.4m, a major board with a huge number of nets and high speed content. When they routed this board with an autorouter, it took 78 days to crank through. If they had an expected cycle time of two weeks, they have only one iteration to run on that router. Then they have to run with the results. They don't have time to take a second pass to see if he can do any better with the results. That may be just end up pushing the cycle time out. With XtremeAR they leveraged all 15 CPUs and cranked it down to 18 hours. So they could do multiple iterations to determine optimized performance with optimal numbers of layers as well.
What is the packaging and pricing of XtremeAR?
The beauty of the method we have used to license XtremeAR is that designers with XtremePCB licenses and that have an autorouter at their disposal, can also use XtremeAR at no extra cost. They can choose which method they want to use. The example I talked about of a laptop designer can now do some prototype runs overnight to determine optimal layout before they start actually laying out by hand during the day it is very possible. Use XtremeAR overnight and then XtremePCB in the day to really optimize the utilization of resources in that environment.
What are you charging people who do not have XtremePCB?
We charge $50K for server and about $20K for each of the clients. That's why you are not going to get everybody to go for 15 licenses automatically. They are going to be determining roughly what they need. That's why we have done a lot of work to determine the optimal configuration. We don't want them buying up 15 clients when they really only need 3 unless they wanted to run three jobs simultaneously. It is really a function of determining the type of server that is out there, the typical circuit they are running and then recommending the appropriate number if clients based upon that data.
Are these perpetual or time based licenses?
These are perpetual floating license but time based licenses are also available.
What percentage of boards are you targeting?
The beauty of this thing is that you can apply it to interactive type designs or automatic type designs. It's not the type of board. It is more a question of industry and the impact on time that these industries are faced with. Certainly consumer market is the highest priority followed by telecommunication which is a close cousin of consumer. If you look at those two markets combined and maybe mix in computers, you are around 70% of the PCB market. It depends on how you slice and dice this thing. There are industries but becoming fewer and fewer where time is not of the essence. But even the military is being crunched by constrained performance to the extent that we are seeing some military taking a look at this technology.
Is there any other vendor offering anything that overlaps either XtremePCB or XtremeAR?
No! It is a patented technology. We will protect it at some level. It is unique. We have another product called TeamPCB which is a simpler methodology that allows people to partition designs manually, distribute them, and bring them back together. That's the closest thing we have that anybody out there. The folks at Cadence have some tools. They are talking about automating a little bit to compete with that solution. But nothing on the order of XtremePCB or XtremeAR where you can have multiple users simultaneously working on the same design.
What is Mentor's marketshare in the PCB space?
It depends upon the region but overall we have 40% share. The market has been growing 1% to 2% a year. Number 2 is Cadence with their Allegro product line and then Zuken. Between the three companies, we form an oligopoly that owns about 70% to 80% of the market.
Between the three firms, is one company stronger in one geographical area or in one end user industry?
Tough question! Where the company started has an impact. Cadence and Mentor are strong in the US. Zuken started in Japan and has enjoyed good marketshare there. It is somewhat diminishing just because of globalization of industries. They have not been able to keep a hold on that unique space. Cadence and Mentor have been making inroads into that space. At the same time Zuken has been making some inroads in Europe. In Europe we have over 50% of the market. That's the market where we have the largest share.
The overall market is growing only 1% or 2% per year?
Yes. If it pops over 5% compound, we are happy for it. At the same time, it is a question of how you measure that market. We are seeing crossover in technology between IC and PCB that will fuzz up that measurement in the future. That's where we are seeing a lot of growth at the moment.
Is it that the number of users is flat or that the prices are dropping?
It's a mix. Certainly we are seeing the industry shifting a bit. US jobs are definitely going over to China. We are seeing that happen. The prices of designs are transitioning as well.
If you look at the market Dataquest has the traditional pyramid model where you have at the bottom about 70% of the seats in what they call the late adopter market, low end type designers. Up to 30% lies in the mid range and 10% at the top. But the revenue is spread equally between all three. What that points to is an industry where at the bottom you have high volume low cost per seat, at the high end you have low volume and high dollar per seat. Because of typical industry consolidation, you see some shift in that. But I can't say by and large that all of the independents are getting bought out and everybody is filling up on that top part of the pyramid. There are still plenty of smaller designers out there where the barrier to entering the design market is low enough that individuals typically spinning out from larger corporations can form their own businesses and be fairly competitive.
Be the first to review this article