Imagination Technologies Agrees to Takeover by Canyon Bridge

[Daniel Nenni]

Don't you think $675M is a bit much? $65M for MIPS? Any idea what Tallwood is going to do with MIPS?

Imagination Technologies Group Plc agreed to be acquired by China-backed private equity firm Canyon Bridge Capital Partners. Canyon Bridge said it will pay 182 pence a share in cash, or more than 500 million pounds ($675 million), for the U.K. designer of graphics chips. That’s 42 percent more than Imagination’s closing share price on Friday.

As part of the deal, Imagination will sell its U.S.-based embedded processor unit MIPS to Tallwood MIPS, a company indirectly owned by California-based investment firm Tallwood Venture Capital, Canyon Bridge said.

Canyon Bridge was keen to structure a bid to avoid scrutiny from U.S. regulators, Bloomberg reported earlier this month

Imagination Technologies Agrees to Takeover by Canyon Bridge - Bloomberg

“We are investing in U.K. talent and expertise in order to accelerate the expansion of Imagination, particularly into Asia," Ray Bingham, partner at Canyon Bridge, said on Friday.

 

[JeffreyHF]

Canyon Bridge will surely have the funds, and perseverance,necessary to sue Apple for infringement of GPU IP, now that the purportedly internally developed Apple graphics are going commercial. Imagination joins the growing list of Apple suppliers, lethally cut by their dual edged sword.

 

[Daniel Nenni]

Canyon Bridge will surely have the funds, and perseverance,necessary to sue Apple for infringement of GPU IP, now that the purportedly internally developed Apple graphics are going commercial. Imagination joins the growing list of Apple suppliers, lethally cut by their dual edged sword.

It is a common cycle: Get too big for your britches when someone even bigger takes you down. For QCOM it was Apple, for Apple it may be the Chinese Government through Canyon Bridge. Huawei and the other Chinese phone makers have already demoted Samsung and Apple out of the mainstream Chinese market and India is next.

As I travel through Asia the term "American Arrogance" pops up much more frequently.....

 

[name99]

This assumes
(a) that Apple has created a GPU
(b) that this GPU infringes IMG IP

Both of these are hardly in evidence. Moreover Apple's experience with UWisc, regardless of whether its engineers thought the matter stupid or not, has surely awoken senior management to the importance of following patent law, not matter how irritating that might be.

To simply assert that Apple are so arrogant that they'll just ignore patent matters (even after having been recently served a $500 million judgement) is lazy ideology, not rational thought.
If Apple have gone down this path, they have good reasons for believing it makes sense, and that they are protected from patent lawsuits, either because they're making something very different from a traditional GPU, or they have licensed the full portfolio from some other player (Intel, ATI, nV, ARM).

 

[Daniel Payne]

This assumes
(a) that Apple has created a GPU
(b) that this GPU infringes IMG IP

Both of these are hardly in evidence. Moreover Apple's experience with UWisc, regardless of whether its engineers thought the matter stupid or not, has surely awoken senior management to the importance of following patent law, not matter how irritating that might be.

To simply assert that Apple are so arrogant that they'll just ignore patent matters (even after having been recently served a $500 million judgement) is lazy ideology, not rational thought.
If Apple have gone down this path, they have good reasons for believing it makes sense, and that they are protected from patent lawsuits, either because they're making something very different from a traditional GPU, or they have licensed the full portfolio from some other player (Intel, ATI, nV, ARM).

A few articles have made it quite clear that Apple has created their own GPU for the A11 SoC:

The new iPhone 8 has a custom GPU designed by Apple with its new A11 Bionic chip | TechCrunch

 

[Eric Esteve]

Apple has (most probably) implemented their own GPU IP in the A11 AP. As a matter of fact, Apple' engineers have integrated Imagination GPU since at least the A5 version (PowerVR SGX543MP2 GPU IP in 2011). In other words, during at least 5 years, Apple has integrated successive Imagination GPU IP in 6 generation of Application Processor (A5 to A10). That gives credibility to IMG, complying against Apple.

When a chip maker (Apple in this case) integrates a multi-core IP (not a single black box) into a complex chip, this integration process give a pretty good visibility, for example of the Test strategy (DFT), clocking scheme, and probably about the inside design as well...

Moreover, when this chip maker publicly announce (in advance) that he will not select this IP, with the immediate consequence that IMG stock fall by 60% (in January 2017) and use this opportunity to hire key engineers, building R&D center very close to the vendor to make the move easier... we can have a doubt: are these "fair business practices"?

Being focused around IP for more than 10 years, I have learnt 2 things:
- IP is a niche market ($3.5B serving $350B semi market)
- maintening an healthy IP ecosystem is essential: no more IP and 80% of the semi market will be at risk

Remark: Extracted from the Bloomberg article, above mentionned: "MIPS is a technology used by software to tell processors what to do. It was widely used for scaled-down processors in cars and home networking gear, and turned up in higher-end devices like Sony Corp.’s PlayStation, but its importance has faded."
Great to see that the same people who explain us the business don't even know what CPU IP is!

 

[name99]

You missed my point Daniel.
You (like IMG) are assuming that Apple has created a "traditional" GPU.
What if they have created something more like a generic throughput processor, which simply runs traditional 3D software as one of its applications? In that case, most of the precise patents that IMG holds become more or less irrelevant. The generic idea of TBDR is no longer under patent.

There are multiple ways one could create a relevant throughput processor, but one possibility is a laned processor, somewhat as I described (very schematically) here:

Real World Technologies - Forums - Thread: A11 GPU Compute numbers are out