NEW YORK What's the next acquisition target for MIPS Technologies?
While keeping his cards close to the vest, MIPS CEO John Bourgoin hinted in an interview with EE Times this week that a DSP core company may be next.
With its acquisition of Chipidea Microelectronica, MIPS now owns not only processor coresits bailiwickbut also analog and mixed-signal IP from Chipidea. All will be used to address the wireless, digital consumer and connectivity markets.
"MIPS should absolutely put a priority on developing complete MPU plus mixed signal subsystem IP products," said Gartner Group analyst Christian Heidarson. "If MIPS can shield mixed-signal integration issues from the SoC designer, this would be of tremendous value."
The deal brings MIPS more of the technology pieces it will need to become a supplier of virtual systems-on-chip, but the company has a ways to go yet.
MIPS CEO John Bourgoin
Bourgoin categorized SoC building blocks into four quadrants: the control plane (processing cores); intellectual property, including analog; embedded memory; and data planes such as digital signal processing. Bourgoin identified analog IP as the fastest growing segment, which prompted the acquisition of Chipidea.
Embedded memory is also an important area, but companies such as Taiwan Semiconductor Manufacturing Co. Ltd. are bringing out memory compilers that foundry customers can use for free.
The key missing piece, in Bourgoin's view, is the DSP core for data planes. Many currently available DSP cores are custom or specialized offerings. "We want something more generic," Bourgoin said. "We are beginning to see some interesting things out there," he added without elaborating.
A virtual SoC may be the Holy Grail for IP vendors, but most design teams are skeptical about licensing large, off-the-shelf subsystems, cautioned Gartner's Heidarson. "They would want the subsystem optimized for their application," he said.
For that, MIPS will need to provide optimization either as a service or through an EDA solution. That makes a tool vendor another potential acquisition target.
"The EDA solution would be more scalable but would require a huge R&D investment," said Heidarson.