Back to EDA

Joe Costello left Cadence Design Systems and the EDA business more than ten years ago, but now he's back with Orb Networks. E&T finds out what he has to say to today's design industry.

It is almost 12 years since Joe Costello left his job as president and CEO of Cadence Design Systems. In roughly a decade at the company's helm, he helped turn it into the largest supplier of semiconductor design tools and one of the world's most profitable commercial software vendors.

Along the way, he also garnered a number of personal plaudits. Not only did he know how to lead a company and exactly how its technology worked, but Costello also brought an enthusiasm and a pizzazz to electronic design automation (EDA) that reflected a sector 'on the move'.

But more recently, Cadence has been through interesting times: four CEOs since Costello, and more recently, both a failed bid for rival Mentor Graphics and the restatement of its financials. Another rival, Synopsys, has also stolen the crown for market leadership.

For his part, Costello has been seen primarily in three roles, all detached from EDA. He was briefly involved with the Knowledge Universe e-learning venture backed by financier Michael Milken and Oracle CEO Larry Ellison. Then, he spent almost a decade running a manufacturing CAD company, Think3. And since January 2007, he has been CEO of Orb Networks, one of several competitors offering services that allow you to access media on your home network from anywhere else in the world.

Even though Costello maintains many friendships at Cadence and elsewhere in the EDA business, you almost take him at his word when he says: "I'm happy to defer to just about anyone on what's happening [in the industry] right now. I've been out of things for a while."

The only thing that gives you pause, that suggests there may be just a smidgen of self-deprecation to the comment, is the fact that Costello has just got back into EDA - or at least his wallet has.

Oasys found

In the run-up to this year's Design Automation Conference (DAC), start-up Oasys Design Systems has come out of stealth (aka 'best-make-bloody-sure-that-this-works') mode with a tool for logic synthesis aimed at roadblocks seen for designs of 100 million gates and above (see box, p43).

Costello is named as both board member and investor, and while Orb remains his main priority, he is adding his weight to Oasys' pre-DAC publicity push. Meanwhile, Costello's involvement with Oasys comes alongside support from several other EDA veterans, including former Synopsys executive Sanjiv Kaul and Larry Yoshida of Japanese tools distributor Innotech.

Costello says his decision to get involved was largely based on gut instinct. "There are those times when someone says something or comes along with an idea and, inside, you just go, 'That's it. That's right'. That is a lot of what happened here."

However, there is also an established belief in EDA that the landscape will shift dramatically every 10 or 15 years, technologies will emerge that rejuvenate the business and, in the past, shake up the leadership rankings.

"It does feel like we are at one of those watershed moments," says Costello. "Synthesis for the size of designs today is difficult - it takes so long that the cost hurts. And with technologies like that, you reach a certain point where you need to look at things from a fresh point of view.

"I'm not saying that Oasys is the game change, the only one. But a whole lot of things are coming together and all of them are going to change things in different ways. And, if you look at how we first built the business at Cadence, we had exciting things that we were doing ourselves, but there were also a lot of elements that we added through acquisition. Success tends to be a mixture of things."

Rewriting the business rules

Costello also refers to game theory in explaining why he has put money into Oasys. "One of the fundamental mistakes that gets made when applying it to business is believing that there are explicit rules. There aren't - not in the way that there are laws of gravity or quantum theory or thermodynamics.

"When you play to the existing rules in business, the chances of winning are actually very low. The power is all with the companies that originally set them. So, the companies that are going to win in the longer-term will be those that set out to change and rewrite the rules. That's what I'm looking for."

The other notable aspect to Costello's thinking is that it puts him at odds with mainstream thinking about EDA during the current downturn. Traditional investors are pulling back from the market.

"And I can understand why that's happening. If you invest in a start-up now, it's very hard to see how one that might hit $10m or $20m in revenue is going to grow to one at $600m, $700m or $800m, but those are the kind of numbers being expected of companies going to IPO. And that's not just in EDA," says Costello.

"Before these guys spend a dime, they're thinking about the exit and there are all sorts of numbers they want from that. They also may not think that the sales numbers for acquisitions don't meet their targets. Then for Wall Street, EDA's a hard industry to follow - there's not as much market cap[italisation] to follow, there aren't that many big companies. None of that helps if your perspective is just all about selling shares. To me though, it's about the opportunities you can feel happening as the ground shifts."

Talk of EDA in the financial doldrums also inevitably leads to Cadence's situation today. Costello says bluntly that the company's board simply has not got the leadership right.

"When I left, we took completely opposite views about the succession. I said it had to be someone from outside who could take a look at where the company had got to, where it could go and do that at a remove from what had happened. Someone like that also wouldn't have the same baggage in being compared to me as the person who had taken the company through one stage as it went into another," says Costello.

"And they said, 'No'. At the time, there seemed to be more interest in selling the company than anything else, so at first they said we'll promote Ray Bingham - and I said, 'That's a terrible idea. Ray is a good, a really good, CFO but that is not what you need for a CEO in EDA'. The compromise was to go a couple of levels down and find someone who might be part of that next wave, and that's how Jack Harding got the job, but it wasn't fair on him and it wasn't how I felt things should have worked."

The right CEO

Indeed, both Harding first and then Bingham held the Cadence CEO position before the company did bring in an outsider, former Intel executive Mike Fister, whose tenure proved controversial in its own way as the company withdrew from the mainstream EDA world to go it alone at major events, saw rumours of private equity deals swirl around its ownership, and saw its strength in several technologies come under heavy assault from rivals.

"I'll say up front that I haven't met Mike Fister, but it just seems to me that he didn't have the right background or come from the right culture to provide what Cadence needed," says Costello, who does not discuss current incumbent Lip-Bu Tan, who is only just settling into to the role. Nevertheless, he retains a very clear idea of what is needed from an EDA leader.

"It's all the qualities that make a great CEO in any industry - put forward a compelling vision, build great teams, enable the execution, stress accountability and responsibility, delegate and all that. But, for EDA, you have to be able to articulate a very specific kind of vision. You're selling ideas to really smart software engineers who are working for you, and then you're selling the results to really smart design engineers who are your customers.

"The complexity of this business is incredible, and you really have to have a depth of knowledge. You can't just do the business school thing."

In the coming weeks, Costello will once more turn his entire focus towards Orb. Its MyCast technology allows users to connect to an 'always on' PC that stores all their media and which may even integrate a TV tuner via devices anywhere in the world - all you need is a broadband connection.

Its software has been available for a while but, Costello says, this autumn the company will launch a "full-on consumer product". I've always said that we needed a complete solution and it's going to be fun taking that out there.

Indeed, Costello does not think that he has ever swapped EDA for a series of challenges that are easier - just different. But, for him, that's the whole point. Changing the rules never should be that easy.

Recent articles

Info Message

Our sites use cookies to support some functionality, and to collect anonymous user data.

Learn more about IET cookies and how to control them