Imagination Technologies Chief Executive change

[Ailuros]

https://imagination-technologies-cloudfront-assets.s3.amazonaws.com/presentations/Imagination Technologies - Chief Executive change, trading update and restructuring initiatives - 8 February 2016 .pdf

It's a sad day since it's true that IMG owes a lot to Sir Hossein Yassaie.

First signs of what I only can call poor wannabe journalism: http://www.fudzilla.com/news/39901-imagination-fires-ceo

Let's blame Apple since it's a coincidence that there are over a billion devices sold with IMG's IP and it's another coincidence that they're getting rid of PURE and plan to restructure the company. In all fairness Apple helped to the company's success for the past years; any problems they are facing lately are IMHO to some extend false management decisions and to some extend uncontrollable factors that have nothing to do directly with partner A or B.

But when you're prejudiced there's always the same old tired horse to beat....

 

[tangey]

Apple is the only reason IMG are the company they are/were today.

It's everything other than the apple relationship that was the problem, PURE being the most startling example. There was a possible case for it many years ago when IMG was trying to create a DAB market where none existed, and in fairness they did so. But in these days, selling commodity items such as digtial radios/wireless speakers ?...really...is that something an IP company should be doing? Well clearly not, as PURE has resulted in loss after loss after loss. Sir HY repeatedly said for years now that PURE was crucial to IMG, but now it's being sold. Guess it's not that crucial after all.

IMG have invested a lot of direct cash and indirect cash(in the form of IP), into a whole host of small "promising" startups, I've yet to see IMG get a good return from any of them, IMG's results over the years are littered with right-offs of such investments. Greenplug comes to mind, which was a small investment, but there have been many others. They also bought in a lot of IP, many of which hasn't exactly resulted in large queues forming to license it, such as the Hellosoft Voip/Volte IP. It appears IMG, desperate to find the "next big thing" after their great success with graphics/video, took a scatter-gun approach, in the hope that something might stick, and largely nothing has. Ensigma, the division that designs the connectivity IP has been IMO unbeleivably slow in getting design wins. It got exactly 1 license in the last reported 6 months period. 1 License. This is the division that we were told would blossom as people design Socs for IOT and wearable. I can understand that mobile graphics/video has peaked/matured, but connectivity?....isn't that EXACTLY what IOT is about.

The news today must put MIPS in jeopardy. IMG made a big issue of taking the first year to renew confidence in MIPS, bring stability to it, and ensure that there was an up-to-date and competitive roadmap going forward. Today means that effectively everything is on the table. Does that mean R&D in MIPS will be cut ? Does it mean that MIPS will be sold ? Those are the questions that existing/potential customers will be asking themselves. With that level of uncertainty, who in their right mind will want to license a MIPS core for a new design, when the ultra stable ARM IP family is available.

IMG IMO has made many poor management decisions, no fault of the engineering staff. It is clearly very sad if today's announcement means job losses, which looks likely.

 

[silent_guy]

IP cores are fundamentally a shitty business.

But it doesn't help if you waste your money on stuff that nobody wants. (MIPS, PURE, ray tracing, ...)

(Not a new discussion: https://forum.beyond3d.com/posts/1784415/ )

 

[Rodéric]

Sorry to hear the news, he was @ IMG the past 18 years...

MIPS is critical for IMG long term, IMG must be able to release proprietary SoC w/o paying royalties to anyone. That's what every CPU/GPU company understood, hence ATi selling to AMD, nVidia buying an ARM licence and Intel tremendously improving its GPU.

Selling MIPS would be a mistake IMO, the chips design is sound and very good, also there are plenty of MIPS used out there, just not in devices marketing their CPU.

As for PowerVR Wizard, it needs a console or Amazon or Apple partner to take off, the tech is very interesting, solves existing problems in a different way and is likely not a dead end, if VR takes off, having more believable gfx will be critical for immersion, and there's only so much/so long you can go with rasterizers bag of tricks, at some point you have to do things right

 

[function]

I've been trying to find some performance info on IMGs MIPS cores but can't. How competitive are they?

 

[Entropy]

I've been trying to find some performance info on IMGs MIPS cores but can't. How competitive are they?

In this space, it's all about performance/Watt, which in turn is strongly dependent on lithographic process, and where on the frequency/power curve you position yourself.
Your question is damn near impossible to answer from a technical perspective.

Questions you might be able to answer is what kind of power saving technologies their IP includes and make a comparison there, Or you could try to figure out the maturitiy and completeness of OS/software Tools. Or...

Being competitive carries requirements on several levels.

 

[Ailuros]

I don't personally see MIPS as problematic in any way; MIPS is an established low power design both on a hw and sw level and doesn't look like it's doing bad. It's more the other departments amongst them also PURE that have been mentioned above that might need a second thought.

Not only do I agree with most of Roderic's points, but there are markets like IoT or automotive that need low power SoCs but not necessarily a GPU in them and especially IoT is a market with low margin / high volume prospects.

 

[tangey]

I don't personally see MIPS as problematic in any way; MIPS is an established low power design both on a hw and sw level and doesn't look like it's doing bad. It's more the other departments amongst them also PURE that have been mentioned above that might need a second thought.

Not only do I agree with most of Roderic's points, but there are markets like IoT or automotive that need low power SoCs but not necessarily a GPU in them and especially IoT is a market with low margin / high volume prospects.

MIPS is a company that was sold once, and IMG said when they took it over they had to steady the ship. Existing customers were looking elsewhere, and new customers were most not looking. This was said to be due to under-investment, the lack of a credible roadmap and instability. If we take IMG's statements as fact, the ship was steadied (mind you MIPS units took a significant drop in the last 6 months). IMG have just announced drastic cost cutting measures which might well have impact across all divisions. Instability may be a thing again. A new customer isn't just looking at a processor, they are looking toolsets,support and roadmap. Whatever IMG might now say, if there is a suspicion that the roadmap may not come to pass, or that support might be cut, then of those that might have selected MIPS, many may now feel it is an additional gamble to do so, and prefer to go with the ultra certain ARM roadmap.

IMG said on taking over MIPS (around 3 years ago) "Our goal is to have 25% of the processor market within five years in terms of design elements".
http://www.engadget.com/2013/07/08/imagination-technologies-ceo-mips/

I note there is no mention of IOT (mid-2013 was probably before IOT was a buzzword). The great white knight for MIPS was google and android, as noted in the above article. The inference being that MIPS would/will make it in phones/tablets. You'd think that now that IOT is a "thing" MIPS fortunes would be soaring in licensing.

There are two years left to reach the above goal.

 

[tangey]

It looks like ARM was having a dig at IMG in their results today

"ARM is committed to maintaining a net cash balance in the medium term. This reflects the continued commitment to invest in R&D that is vital to the product development pipeline for ARM and its Partners.......
.......Regardless of the external environment, ARM can maintain this R&D commitment to deliver the next generation of ARM technology."

i.e. You are guaranteed a full roadmap with us. Those that are struggling citing external events may not be able to do similarly.

 

[function]

In this space, it's all about performance/Watt, which in turn is strongly dependent on lithographic process, and where on the frequency/power curve you position yourself.
Your question is damn near impossible to answer from a technical perspective.

Questions you might be able to answer is what kind of power saving technologies their IP includes and make a comparison there, Or you could try to figure out the maturitiy and completeness of OS/software Tools. Or...

Being competitive carries requirements on several levels.

I suppose it was a rather vague question ...

I was assuming the same processes and power envelopes. Although maybe that's still too vague. Some benches from the likes of the I6400 might be nice to see.

On a different note ... will this really be the year of MIPS?

http://uk.reuters.com/article/idUKnGNX2nT1S+e5+GNW20160106?type=companyNews

 

[tangey]

On a different note ... will this really be the year of MIPS?

http://www.eetimes.com/document.asp?doc_id=1328932

There is a long "venting" comment in the comments section, from a guy proporting to be an ex-MIPS employee (in the US) who was there when IMG bought it up. Impossible of course to verify he is who he says he is, but he does bring a certain view that would be consistent with someone on the inside. He makes some of the same points as myself, including the investment / buy-up of a lot of IP which ended up going nowhere or being stuff that no one wanted.

 

[Ailuros]

http://www.eetimes.com/document.asp?doc_id=1328932

There is a long "venting" comment in the comments section, from a guy proporting to be an ex-MIPS employee (in the US) who was there when IMG bought it up. Impossible of course to verify he is who he says he is, but he does bring a certain view that would be consistent with someone on the inside. He makes some of the same points as myself, including the investment / buy-up of a lot of IP which ended up going nowhere or being stuff that no one wanted.

I'll add the entire text in case his comment gets removed:

I was an employee of MIPS and lived through the aquisition by Imagination. We (MIPS) employees were excited by the aquisition since it seemed like a perfect fit between a CPU and graphics IP company. By that time MIPS was down to the bare minimum, experiencing tremendous competition from ARM. We looked forward to the resources and capabilities Imagination could bring.

To our shock Imagination was a disorganized mess of a company and proceeded to replace modern businesss processes MIPS had with their paper or wildly out of date software systems. The MIPS finance system was replaced with Quickbooks! IMG was only traded on the London exchange and as such had almost no analyst coverage. Any purchase over $5,000 had to be signed off by the CEO, it would take weeks/months/never for decidions to be made. Although the CEO seemed to be a genuinly honest person the "vision" he brought was not much more that one might read from IEEE Spectrum or by attending a GSA meeting.

Imagination would aquire companies and then let them atrify to a few otherwise un-employable or hopelessly optimistic employees, Nethra, Caustic, MIPS and so on. There were the very odd relationships they had with startups where Imagination exec's would be on the board of a startup that Imagination was investing in and recieve non GAAP comp arrangements (to be fair this might be considered normal in England). There was the Pure Digital division that was running a multi-hundred person group using spreadsheets. There was the "Flow" division doing who knows what which eventually was morphed into something IoT. There was the HelloSoft division that was pumping money into a sub-division in India doing development on VOIP software that nobody wanted. They built a palacial office park (usually mean death for any company) in the UK with an interior design well suited for a department store but horrible for a intellectual property company where engineers need to sit un-distracted for long periods of time doing RTL or software.

The CEO ran Imagination like a little boy with a pocket full of cash, that cash was the license fees from their last big customer, Apple. Sadly both Imagination and MIPS have some really great employees who work very hard and will never be properly rewarded for their many years of dedication. Bonfire of the vanities.

Shame on the board for taking soooo long to do something, perhaps too little too late!

Ouch! is all I can say.

 

[AlNom]

 

[Malo]

Wow, ouch indeed.

 

[tangey]

Apple is the only reason IMG are the company they are/were today.

IMG have invested a lot of direct cash and indirect cash(in the form of IP), into a whole host of small "promising" startups, I've yet to see IMG get a good return from any of them, IMG's results over the years are littered with right-offs of such investments. Greenplug comes to mind, which was a small investment, but there have been many others. They also bought in a lot of IP, many of which hasn't exactly resulted in large queues forming to license it, such as the Hellosoft Voip/Volte IP. It appears IMG, desperate to find the "next big thing" after their great success with graphics/video, took a scatter-gun approach, in the hope that something might stick, and largely nothing has. Ensigma, the division that designs the connectivity IP has been IMO unbeleivably slow in getting design wins. It got exactly 1 license in the last reported 6 months period. 1 License. This is the division that we were told would blossom as people design Socs for IOT and wearable. I can understand that mobile graphics/video has peaked/matured, but connectivity?....isn't that EXACTLY what IOT is about.

The news today must put MIPS in jeopardy. IMG made a big issue of taking the first year to renew confidence in MIPS, bring stability to it, and ensure that there was an up-to-date and competitive roadmap going forward. Today means that effectively everything is on the table. Does that mean R&D in MIPS will be cut ? Does it mean that MIPS will be sold ? Those are the questions that existing/potential customers will be asking themselves. With that level of uncertainty, who in their right mind will want to license a MIPS core for a new design, when the ultra stable ARM IP family is available.

Well the new board didn't take long to start disposing of stuff. They dumped their approx 10% holding of Toumaz for about £3.5M. This is the company that made DAB chips/modules and wireless health monitoring "plasters" that IMG gave IP to in return for shares. The SP they got for those TMZ shares is just about the lowest it's ever been. IMO, they've taken a massive loss on that multi-year investment. It also looks like they dumped their 10% holding in another company they supported, 7digital, and got the princely sum of £1M for them, imo having spent much MUCH more on that investment.

I'm all for them disposing of stuff that isn't core to the business, but frankly the way they have just dumped these shares at a massive loss doesn't bode well if/when they come to start disposing of PURE or some IP divisions. As an outsider I see a fire sale, anyone vaguely interested in some assests is going to think they can get they at knock down prices. And anyone looking to take for example, a MIPS license, would at this time be very wary of who is going to own MIPS a few months down the line.

I also note from the IMG website that they have drastically reduced their job vacancies, down almost 40% in 2 weeks, which is a prudent move.

I feel for the engineering staff, for various departments, some most have a gut feeling that their department will be closed/sold off. Also It's not entirely out of the realm of possiblity, that once IMG gets rid of non-core stuff, that the remaining company becomes attractive to a suitor. According to one of the UK national Sunday papers, Rambus was snffing around last year.

 

[Ailuros]

The last thing I'd ever want to see is a patent troll snatching them....

 

[Rodéric]

The last thing I'd ever want to see is a patent troll snatching them....

I doubt Apple (maybe Intel too) would let someone buy IMG other than itself. (So if Intel also doesn't want IMG sold it might joint venture with Apple or take as many shares as Apple.)

 

[tangey]

I doubt Apple (maybe Intel too) would let someone buy IMG other than itself. (So if Intel also doesn't want IMG sold it might joint venture with Apple or take as many shares as Apple.)

Intel signalled its disinterest in IMG last year when they sold their entire holding of IMG shares. I can recall but I think it was around 14%.

 

[Rodéric]

Intel signalled its disinterest in IMG last year when they sold their entire holding of IMG shares. I can recall but I think it was around 14%.

Ah I missed that, given they have a capable GPU and haven't used PVR hardware for a good while now it makes sense.
Apple is still dependent on IMG, and if its price drop it might just acquire it to secure its GPU designers.

 

[tangey]

IMG released a statement yesterday, confirming that they are divesting most of their "non-core" elements. "core" is defined as GPU,MIPS and Ensigma. Getting rid of all the sub-units that were needed to support the radio design business, and also reducing their SOC design capability. 200 redunancies, in additiona to the 150 already announced.

However statement confirms they are still looking to fill 50+ PowerVr vancancies.

To me, looks like a great piece of refocusing.