MIPS still has a revenue stream. It's got multi-threading ARM doesn't have. To think of the semiconductor industry losing the only legitimate alternative CPU IP to ARM would be "a real shame," some say.
Imagination Technologies’ announcements last week — selling itself to Canyon Bridge while agreeing to sell its MIPS CPU business to Tallwood Venture Capital — have some industry observers scratching their heads.
They asked: “Who wants MIPS now?”
Mike Demler, senior analyst at the Linley Group, summed up the general sentiment: MIPS “is an OLD CPU architecture.”
My colleague Rick Merritt also observed, “MIPS doesn’t have much of an ecosystem. These days, it’s all ARM and x86.”
First, I wondered if things might be a little different in China.
When I asked how MIPS is doing there lately, one of my colleagues at EE Times China in Shenzhen described MIPS cores as “not the mainstream.” She acknowledged that Chinese fabless chip vendors like BLX IC Design Corp. (Loongson), Action and Ingenic are still in MIPSers. But she sees the MIPS ecosystem in China as limited to a small circle of companies.
The proliferation of ARM-based reference designs has made it easy for Chinese design engineers to go with ARM. “There are too many success stories” starring ARM, she added.
MIPS lacks market momentum, to put it mildly.
But MIPS was already suffering from battered “old CPU” syndrome five years ago.
In late 2012, when Ceva and Imagination Technologies were in a bidding war over MIPS, many industry types wondered what Imagination could possibly do with MIPS, a loss-making business worth maybe $60 million in annual revenue. However, when Imagination ended up snapping up MIPS with a knockout offer of $100 million, a kinder analysis then was that the price tag must indicate how important the CPU move was to Imagination’s future.
While Imagination management might have valued MIPS strategic importance then, they are culpable now of having neglected MIPS while mismanaging the CPU core and its roadmap.
So here we are, still wondering: “Whither MIPS?”
So I asked around: Does anyone even want MIPS anymore?
The Linley Group’s Demler noted that there’s “still a revenue stream, so that’s worth something.” He added, “But at this point, I don’t think anyone would buy MIPS for whatever patents that haven’t expired. It’s obviously a depreciating asset, dropping from $100 million to $65 million in just under five years, although Imagination got the bid up from their original $60 million offer.”
Indeed, some MIPS customers today include stalwarts like Mobileye (now owned by Intel) in automotive and Cavium and Broadcom in networking.
Apparently, MIPS is also going into high-volume smartphone modems. Just last month, Imagination revealed that MediaTek has adopted “multi-threaded MIPS I-class CPUs for smartphone LTE modems.” The first device from MediaTek featuring MIPS technology is the new flagship MT6799 Helio X30 processor, with MIPS in its Cat-10 LTE modem, it said.
While MIPS definitely has a “brand” problem and it’s clear the business has been mismanaged, some observers cling to optimism. They believe that MIPS still has potential. If done right, they say it can become a successful, profit-making business again.
Earlier this year, Kevin Krewell, principal analyst at Tirias Research, told us that “MIPS is still a classic CPU design with scalability and an established software ecosystem.” In his opinion, there's “still an opportunity for MIPS in the market, even though it's a diminished one.”
Meanwhile, others note that MIPS CPU architecture has certain performance and efficiency advantages — such as multi-threading technology. MIPS could shine in real-time, power sensitive applications such as LTE, artificial intelligence and IoT, they say.
In the future, though, Demler predicted, “Mobileye is likely to replace MIPS.” He is of an opinion that “the [MIPS] architecture has steadily lost market share to ARM and ARC.” But as Demler said, “There’s always a buyer somewhere, even in a fire sale.”
Asked about the value of MIPS today, one industry source who spoke on the condition of anonymity, told us, “There’s still a team of 200+ engineers developing CPU technologies, some of which are even more advanced than what ARM can offer — e.g. multithreading.”
Others are more reserved. It’s one thing for the industry to rally around the winning horse — ARM. It’s another to dismiss ARM’s competitor as a hopeless underdog. “To think of the semiconductor industry losing the only legitimate alternative CPU IP to ARM left would be a real shame,” they said.
So, if a company seeks to manage MIPS correctly, what’s the strategy?
“The value is in restoring the IP roadmap, the MIPS brand, and coming at the market in a way that doesn’t put MIPS head-to-head with ARM, but concentrates on the parts of the market where ARM is weak and there is good opportunity to license,” one executive said.
Citing Mediatek’s recent decision to use MIPS instead of ARM in its modem, he stressed, “ARM is in the modem by default, not for any particular suitability or technological advantage. The apps processor is where ARM has the real advantage, as all of Google Android is built on it, so you avoid this area completely.”
That strikes me as a sound strategy. Wouldn’t it be nice to read the story of the hopeless underdog’s miraculous recovery? Or am I just a hopeless romantic?
Maybe I should just admit that the horse is dead, and start flogging.
— Junko Yoshida, Chief International Correspondent, EE Times