(Image courtesy of Siemens).

Siemens Buys into Machine Learning Tools That Refine Chips

Siemens, to supplement its acquisition of Mentor Graphics, said that it had bought Solido Design Automation, whose software tools use machine learning to chisel rough edges off complex chip designs, optimizing power consumption and verifying that the chips are ready to be manufactured.

The acquisition is another smoke signal signifying that Siemens wants to expand into software tools for chips and circuit boards used in everything from factory equipment to airplanes to self-driving cars. Last year, the industrial giant doled out $4.5 billion for Mentor Graphics, one of the three major players in electronic design automation.

Solido, like Mentor Graphics, will be folded into the product life cycle management software business of Siemens’ digital factory division. The Plano, Texas-based group sells software to help manage the life cycle of products like electric vehicles and wind turbines, from design to production to service to disposal. The terms of the deal were not disclosed.

The Saskatoon, Canada-based Solido has raised around $10.2 million of funding since it was founded in 2005, according to venture capital database Crunchbase.

Buying Solido fits with the plans of Chuck Grindstaff, executive chairman of the product life cycle management unit. He is betting that manufacturers will start making chips that allow factory equipment and cars to monitor what is happening inside and around them. Siemens clearly wants to pair Solido's software with its own computer-aided design and analytics tools.

Solido exploits machine learning to make sure that the analog and digital circuits inside chips will be manufactured without flaws, which are getting easier to miss as the circuits carved onto silicon wafers become smaller and smaller. More than 40 companies including Nvidia and Broadcom use Solido’s tools for verification and characterization.

What Solido is not trying to do is replace engineers with software that can create chips from scratch without human intervention. It is not clear that machine learning algorithms will ever have more than a supporting role, taking tedious tasks out of the hands of humans and aiding engineers without decades of experience crafting chips etched with billions of advanced transistors.

Hide comments


  • Allowed HTML tags: <em> <strong> <blockquote> <br> <p>

Plain text

  • No HTML tags allowed.
  • Web page addresses and e-mail addresses turn into links automatically.
  • Lines and paragraphs break automatically.