Apple Q4

Apple announced Q4 last night:

http://seekingalpha.com/article/1780172-apples-ceo-discusses-f4q-2013-results-earnings-call-transcript

iPhone  33.8m
iPad 14.1m
iPod 3.5m
Apple TV not disclosed

All in that's another 48-50m volume for IMG (similar to Q3)
Given the strong iPhone 5s sales it looks like the royalty volume via Apple for IMG's H1 will be around 110-120m. I would also expect around 100m via Mediatek (including tablet and DTV SoCs), beyond that I don't really have a feel for the size of IMG H1 royalty volume (excluding MIPS). Last H1 was 237m vs. 298m in H2, so I would guess at around 300m (ex-MIPS) this time round, given developments/progress.

A couple of interesting charts courtesy of ZDNet, obviously we will be hoping for new peaks in Q1.

http://www.zdnet.com/apple-q4-13-hardware-sales-7000022536/

Apple unveils new iPads

As expected, Apple refreshed its iPad line-up yesterday with the iPad Air and iPad Mini with Retina Display.  Both utilise the A7 processor featuring PowerVR Rogue :

http://www.apple.com/pr/library/2013/10/23Apple-Announces-iPad-Air-Dramatically-Thinner-Lighter-More-Powerful-iPad.html

So Apple are now using 3 devices with Rogue graphics.

This has to be the best graph of the night !
That original iPad used SGX535 (A4 chip) of course, so we've come a long way.

I was hoping for some Apple TV news but perhaps too early for that.

In other news, Imagination announced the first MIPS Warrior cores P5600 (no lead partners announced though):

http://withimagination.imgtec.com/index.php/mips-processors/mips-p5600-series5-warrior-cpu

From discussions at the AGM and other research, the plan for MIPS  seems to be to catch a 'big fish' then hopefully the other semi players will follow suit. For my money that biggie will be Mediatek - given the recent license announcement with ARM that might sound odd, but watch this space.....

Feelings

Woh oh oh, feelings....

Ultimately this was a supremely professional and confident presentation & Q&A session.

'Without our graphics, the smartphone revolution would have been 5 years later'

Ballsy.

Many concerns assuaged here I think, some of the nuances ref the Mediatek relationship are hard to communicate and their decision to dual source (remember they took a MALI license BEFORE their IMG license IIRC) is one of these. It seemed like more of an incidental result of power politics and oversights internally during the mergers of various business divisions and apologies WERE MADE in relation to the manager's comments - it was not right and no-one should base any decisions on it.

The other thing that you had to see their eyes for was about the competition and the benchmarking questions. Not all test platforms are genuine reflections, not all benchmarks mean that much, power consumption will be more of a factor in future generations of tests, if you want to just make things bigger and run them faster you will run out of headroom.

If you run to the bottom and sell crappy tech as new tech then you get found out when the lower end moves up - I feel that this is why there is a Rogue 6100.

I am sure this report will not convert any views, but it is as accurate and honest as I can recall - I am here, listening, writing, asking questions, etc so a little overworked on an iPad !

If you take ANY negatives I really don't think you were reading it right or I wasn't writing it right.

Ultimately, if you are a worried shareholder or a happy shareholder, there is really nothing like being here. You get so much more out of it and will be more able to make comment on very important matters you have concerns over rather than sitting at a keyboard making comments that are ill informed by definition. If you have concerns there is no better place to be than in front of the people who can give you the answers - and if you don't buy their answers you can tell them, ask them for clarification or just plain tell them you don't believe them. I know it's easier to sit at home moaning but if you're serious about it then a small amount of effort is 100% worth it.

More to come later - either here or in another place following the casual chats later.

Thank you Kings Langley, you've been great.

More, MORE !

Ok, the new building is very nice, I'm sure the Californians feel quite at home when they're visiting.

Peace :-)

Q & A - my question about benchmarks and the competition

Said hello from Dave - wouldn't have been the same without his spirit raising its head. No tattoos though sorry, I'm not going there !

Ok my question was stupidly long as I hadn't prepared it and wanted it to be a general overall catch all about the competition and how things stand in relative performance terms.

Have the competition caught up ? Are they at the peak of where their solutions can go ? 
If you do not care about power usage you make the solution bigger and uglier and you burn your face when you make a call

Where is Rogue in terms of your expectations of it when in development. 
Rogue is doing exactly what we thought it would do - there is plenty of headroom we are still very early. 

Are you looking at ways of making benchmarks more effectively reflect the performance in relation ? 
yes the team is working very hard to try to get power consumption into more of these so that you get an actual real performance measure.
To get benchmarks to do what is claimed for them there is a lot of cheating going on - removing complexity, etc. We will be highlighting these cheats very soon

At v low end power is not as important so less good tech can get traction. Over time though as the high end becomes low end then rogue will again come into its own as low end will need OpenGL3, etc so companies will no longer be able to masquerade old tech as new. The day of reckoning will come. 

Per mm and per Mw we are 'by far' the best. QCOM second

Q and A

Atty G asks the MTK general manager comment question

Reality - that answer is not correct and senior MTKVguys have apologised. Their business is growing, was compartmentalised and some people said things. HY questions their logic in making the choice they made, but not that they made a decision.

May have been a function of them merging their feature and smartphone segments.

Ecosystem is important when you care about your customers (Apple) not when a tiny Chinese handset maker sells smaller numbers and moves on. They can choose any solution as they do not support an ecosystem. Sticky is important when you care basically

MTK is v strong relationship and more so on a corporate values level for the future

Finance guy

A lot more confident thankfully than at previous meeting.

Basically took the mick out of analysts previously complaining we were over dependent on Apple, now it's Apple, MTK and Intel. Looking foward to a few years time to seeing that reliance list getting ever longer.

Good confidence.

More blah on how positive people are about MIPS, etc.

PURE

Still very committed to the future and possibilities for PURE as was the venerable Paul Smith earlier - more from him later but he seems invigorated by the possibilities from the Jongo, Flow type dealio

Flow

Next step from Toumaz tests in USA is that the plaster tech will soon go onto A&E incoming as well as on the wards. Nice. Lots of chat about the possibilities, which are huge - thankfully no mention of bloody fridges - that's a tomorrow's world ref I hate !

Ensigma will be an increasingly important income generator of revenue and sales opportunities

MIPS/CPUs

CPUS usage

5-10 phone
Car - 30-50
House - 100's

Mentioned the need not to have a monopoly, DID NOT MENTION ANY SOFT DRINKS :-)

I don't mind that analogy, some of you lot do, but your CEO is listening to you.

Graphics trends

All recent trends will accelerate moving forward - but getting performance within power envelopes will be increasingly important.

Ray tracing - doing it 200 x more efficiently than current best of the rest. And it is easier, cheaper and faster to generate content.

Will ask questions later on more of this power/performance balance to stop the kids squabbling.

Industry trends - various stuff but concentrate on MIPS potential -slide 17

CPU ISA no longer a key in the near future thanks to LLVM, etc

Compile on demand will be big and this helps MIPS as you can compile when you need to run it and it compiles onto your platform.

Shipments - big news

Potentially very interesting development going back to something I've been forecasting for ages now.

1bn non MIPS target now moved to FY 2016 from previous CY2016. So my interpretation we are ahead of previous target date.

MIPS graph is as it is Ian because new IP is 1.5-2 yrs, customers have 1.5-2 yrs so the growth leg is down the line. Until then we have a steady rise, hockey stick comes later

"People underestimate MIPS, we will prove them wrong over the next few years."