MyDesign: Final answers from IPextreme’s Warren Savage

Two things happened as a result of falling and breaking my right arm early on the morning of April 19th in Monterey: I instantly became a ‘Lefty’ for the first time in my life, and I missed Warren Savage’s presentation at EDPS later that day.

Warren is CEO and President of IPextreme, and I kid you not when I say that what he doesn’t know about the IP industry isn’t worth knowing. That’s why I wanted to hear Warren’s talk, and why I was very happy to talk to him this week about my Dick Tracy keychain project.

How do I learn to be a knowledgeable customer of the IP industry, I asked Warren, particularly when my hypothetical wearable is something I could really use right now: An SoC-based gadget, built with oodles of IP, to wear on my left wrist that’s got one small button to remotely unlock my car, one that will start my car, one that will open or close the garage door, one that will tell me if I’ve got enough milk in the fridge, one that will turn the heat up and down at home even if I’m not there, and prosaically, one that will show me the time.

Of course, now that I can’t use my right hand to push the buttons on the device strapped to my left wrist, I no longer want buttons. I want the thing to respond to voice commands – “Unlock.” “Ignition.” “Garage.” “Got milk?” “Set temp.” “Time?” – simple instructions that should only produce results when it’s my voice and nobody else’s.

Warren was extremely informative during our phone call. He understood I wasn’t looking for specific help with my design, but how to shop for the IP to go into my design. I started by telling him that my research into IP has so far included conversations with:

CAST – Hal Barbour, Nikos Zervas, and Paul Lindemann
Sonics – Grant Pierce, Raymond Brinks
Adapt IP – Mac McNamara
S3 – Dermot Barry, Darren Hobbs

To further clarify the information gleaned from these people, my questions for Warren were very succinct, as were his answers.

WWJD: What process node should I target for my SoC?

Warren: Selecting a target node depends on a number of factors. There is more IP [to choose from] at the larger nodes, because there is very slow movement to the more advanced nodes. There are fewer design starts at the advanced nodes, so [less IP has been developed for the market there].

At the recent GSA IP Summit in Silicon Valley, IP customers on a panel wouldn’t comment on their adoption and design starts at 14 nanometers versus 28. That’s why the IP vendors hold back [development at the advanced nodes].

Also, certain nodes tend to be specialized to industries. Auto guys, for instance, have stringent requirements [for their chips] with regards to temperature range, which means lower volumes and higher prices. So the auto industry, and aero with their rad-hard requirements, tend to go with internally developed IP, which [translates] to IP merchants having lower motivation [to produce IP useful to those industries].

WWJD: How do I know what a block of IP costs? My conversations so far have indicated that the cost always depends.

Warren: That question is equivalent to asking what a car costs; it depends on what’s loaded on the car. However, back in 2008 we were the first to publish our prices on the web, and we still honor those prices.

WWJD: Is it accurate to say then, like cars, the more expensive the IP, the less specific the price?

Warren: Yes, somewhat. Like EDA, making [the price of expensive tools] too public would encourage the commoditization of the products. And that would only create noise, which would confuse the sales process.

Having said that, every IP company has a pricing guide. Their sales guys have list prices that are discountable [depending on the situation]. Sales needs those types of guidelines when they’re facing the customers, but even then there’s a lot of horse trading involved.

Maybe one deal needs higher royalties and lower licensing fees, where a different deal requires something like a royalty buyout. In that case, the IP customer pays all of the royalties up front, so they can write off the fees as R&D costs. Whatever the deal is, it has to fit the needs of the customer.

It’s all very rational, but complex.

WWJD: Does the amount of free design services thrown in by the IP merchant depend on the price of the IP I’m buying?

Warren: Based on my previous experience at Synopsys and my current experience at IPextreme, design services have very little association with an IP sales transaction. Only 5-to-10 percent of transactions are done in conjunction with design services – although, we might be an outlier in that statistic.

Certainly, companies like ARM, Synopsys and Imagination are not in the business of design services for their IP customers. They want to be a product company. In the case of mixed-signal IP, however, where you might see more customization, there is more motivation [to provide design services along with the IP].

In fact, leading customers in some situations may actually help their IP merchant develop special IP at a new node, which the merchant can then license and use for the designs of their next customers at the next node, and the porting fee can be amortized.

WWJD: Does it annoy ‘leading’ customers that they’re bearing the brunt of the development effort?

Warren: Yes!

WWJD: Some estimate there are about 30 EDA companies today, but close to 500 IP companies. With so many IP providers, how do I know which one to use? Do companies who buy a lot of IP have dedicated staff on board whose sole job is to keep a look out for new providers, and to stay up-to-date on who sells what and at what price?

Warren: For large design houses like Faraday or Unichip, typically they have teams for developing specific IP and go-to vendors for [known, good IP].

WWJD: How can new IP companies emerge in that environment? Doesn’t that mean that a handful of well-known vendors get all the business?

Warren: Generally, that’s right, particularly [for commoditized IP]. The differentiators [between providers] might be around protocols, interfaces, or new technologies with better quality or lower power. But a USB is a USB – you don’t need something special for off-the-shelf IP.

However, for something special, a new company may become the provider, particularly when design services [are part of the equation] with the new technology. In that case, the new company becomes part of the Rolodex for the IP customer.

WWJD: So there are still opportunities for start-ups in the IP industry?

Warren: Yes, it’s still very much a Wild West in the industry. It’s moving so fast, with a dynamic group of players – especially with so much pricing and scheduling pressure on the companies who use IP in their designs.

There is lots of room for innovation for companies in the IP industry who can do things quickly. You will survive if you are big enough to ride the next wave technology, or small and agile enough [to respond to opportunities in the market] and get a foot in the door.

Being in the IP business is not for the faint of heart. Nonetheless, we are seeing an explosion of new companies in the industry.

WWJD: Final question, how do I know what IP is out there?

Warren: You can go to the well-known online catalogs, but these days most people start with a Google search and go from there.

*******************
Bio …

Warren Savage, President and CEO of IPextreme, Inc., is perhaps one of the most recognizable figures in the semiconductor industry. He has spent his entire career in Silicon Valley working with leading companies, including Fairchild Semiconductor, Tandem Computers, and Synopsys, where he focused on the problem of building a global, scalable semiconductor IP business.

In 2004, Warren founded IPextreme with the mission of unlocking and monetizing captive intellectual property held within semiconductor companies and making it available to customers all over the world. Warren has a BS in Computer Engineering from Santa Clara University and an MBA from Pepperdine University.

*******************
Editor’s note …

I have not enjoyed life so far as a Lefty, despite a keyless-ignition car, a mouse that’s had its right and left button functionality reversed, and Siri. Being left-handed, per my current experience, does not make one more creative, innovative, or interesting. But it does give one renewed respect for people who have survived permanent loss of limbs or physical abilities, and yet continue to live fully and without self-pity.

*******************

Tags: Adapt IP, ARM, CAST, EDPS, Faraday, GSA, Imagination Technologies, IPextreme, S3 Group, Sonics, Synopses, Unichip, Warren Savage

This entry was posted on Thursday, May 1st, 2014 at 10:58 pm. You can follow any responses to this entry through the RSS 2.0 feed. You can leave a response, or trackback from your own site.