The August 1 completion of the acquisition of Apache Design Solutions by physical-modeling vendor Ansys brought a sudden end to a quick and surprising sequence of events. It also raised some fascination questions, not the least for Apache’s existing customers.
A brief recap might help. In mid-March, Apache registered its intent to make an Initial Public Offering. The idea of a successful small-cap IPOL in the EDA industry after so many years in the wilderness cheered everyone. Few mentioned that the S-1 filing described uninspiring GAAP earnings growth, a market in which Apache already had won the top 20 prospects, an a very pro-forma statement on the intended uses of funds.
Then on June 30, Apache announced Ansys’s intent to purchase it for $314 million in cash. On August 1, the deal was completed. “We only spent seven weeks from our first discussion to closing,” remarked Apache president—and now Ansys VP and GM—Andrew Yang. Apache will now operate as a wholly-owned subsidiary of Ansys.
But why? The motive was certainly not financial. Setting aside Apache’s $31 million in cash, Ansys paid about $280 million for what has been, during two good years in the highly-cyclic EDA industry, a relatively stable $6 million a year in GAAP pre-tax operating income. You can do better than that taking your money to the bank. The reasoning had to be strategic.
This is the point in an article where the word “synergy” is supposed to appear. But Ansys is an old company from a rust-belt heritage. Until the last three years, it has grown organically by finding new applications—in mechanical, structural, thermal, and fluid analyses—for finite-element modeling. The customers had all been one form or another of civil or mechanical engineer. Apache is a relatively young Silicon-Valley company that supports an elite group of design-analysis engineers inside IC design teams. There’s a synergy?
The short answer is no, but there will be. Ansys has, since the acquisition of electronics analysis vendor Ansoft in 2008, been building a suite of design entry and analysis tools for package/board/system-level analysis of high-frequency analog, RF, and microwave systems. In that sense there is a conceptual connection between the Ansys offerings at package level and Apache’s power-integrity tools at die level. Further, a small number of electronics engineers are finding themselves entangled in thermal and mechanical-stress analysis as they study the interactions between packages and their increasingly stress- and temperature-dependent circuits. But the two sets of tools—from Ansys and Apache—solve two different kinds of problems for different kinds of engineers and require quite different skills of their users. And, Yang conceded, these engineers are spread across a number of poorly-communicating silos today.
“Design problems often begin at the interfaces between the silos,” Yang said. Therein lies an opportunity. Today, Apache tools can generate compact models of on-die power networks and export them into the Ansys environment, allowing accurate modeling of die-package-board assemblies. Neither chip nor system designers have to settle for oversimplified models of the other’s work.
The concept becomes more exciting as 2.5D and 3D packaging erase the boundaries between die and package. The notion of die-level and package-level integrity engineers not just exchanging models but sharing a simulation cockpit suggests a real impact on both schedule and design quality. Still they are different kinds of professionals, but you can visualize them working together. “Integration has to bridge the different backgrounds of the engineers,” Yang said. “A common user interface could lower the barriers.”
The potential impact on 3DIC design is not lost on Apache’s major clients. Yang said some key customers have already approached the new Ansys subsidiary about teaming up to build capabilities in the 3D area.
But Yang also acknowledges that this pot of gold lies at the end of a difficult journey. Mergers often hemorrhage key managers and irreplaceable engineers, fail to find expected synergies, and cost shareholders dearly. Differences in culture can be an almost insurmountable problem to anything beyond financial integration. And the remaining executives in a merged organization must struggle to retain focus on existing customers while restaffing and reaching for those promised synergies. “The first priority is continued focus on our core business,” Yang emphasized. “Then we explore opportunistic investments. But you must balance focus and integration.”
Unsurprisingly, Yang is optimistic. He points out that the absence of market overlap is a strong positive in a merger, as is the potential for complementary areas of expertise, as certainly exist between Apache’s deep understanding on on-die networks and Ansys’s strength in electromagnetic, mechanical, and thermal simulation. And the Ansoft acquisition, successful by pretty much any measure, is still thriving three years in. “Ansoft may be an indicator of how good these guys are at mergers,” Yang suggested. Given Apache’s vital role in IC creation today, and the daunting challenges of 3D design tomorrow, we all need for Yang’s optimism to be correct.