Finally A very good reason for Apple to consider Nvidia
Application Processors iPods, iTV, iPads & iPhones - 300m units. LTE modems - 150m units.Total LTE + AP chips - 450m units
Savings using NVDA equivalent Chips ~ $15 average per chip. Apple Chips are typically very large - larger than QCOM's for any given performance, and QCOM modem is not cheap.
Total savings - 450m X $15 = $6.75B per year. Over 2 Years - $13.5B + $1B Profits from GPU related busineses = $14.5B
Approx 2X current price
Apple will benefit even more than the $14.5B due to opportunity loss for any further delays in bringing an T4/T4i class product to market! It is impossible for Apple to get the 28nm capacity immediately today from TSMC - Samsung's 28nm is far from mature! AND it's almost impossible for Apple to come up with an A15 class APU using Power VR without the associated power issues - this will be their biggest challenge in the next 6 months!
So NVDA's worth to Apple is far beyond pps - well of course they can get the milk for free by just buying Nvidia's APUs instead of buying the company. That is short sighted and shows a lack of understanding of the implications of the T4/T4i achievements and lack of understanding of the chip technology. Buying Nvidia locks everybody out of using the most advance APUs etc for a long time to come and provides a clear evolutionary path, a clear low risk path to the next two to three generations of market leading APUs with little to know risk.
Infact anybody who wants to be in the hardware business in a big way would be wise to consider the above implications. MSFT and Goog are not serious in the hardware space. Samsung & Apple are - HP may be.
' it's almost impossible for Apple to come up with an A15 class APU using Power VR without the associated power issues'
Err no, they could double the clockspeed of their A6 and it would perform the same as an A15 but still burn less power. Sorry but nobody needs Nvidia as anybody can do what Nvidia is doing and use generic ARM cores.
"Sorry but nobody needs Nvidia as anybody can do what Nvidia is doing and use generic ARM cores."
that's what brcm said in their investor's day last nov. The question they asked themselves was how could anybody differentiate and add values. For brcm, that was using single chip fully integrated solutions to attack the bulk of the market. brcm could not care less about the top of the market where #1 & #2 will use their own chip sets, #3 is cheap only, and #4 is into spreading the wealth. There is no money to be made, and tons of money to be lost, doing it the way nvda is doing.
nvda and nvda longs here have not accepted that above fact about the market segment that they are chasing. nvda needs LTE a whole lot less than they need 3G and even 2G for the next two years.
Ibx is right about JH not selling out unless he's at the helm. That is a very high bar to cross.
On the other hand - there would be a price where everything is possible - that's life - it would be some astronomical number however. JH and team certainly believe they can take the company to $25ps - so that would be a starting point for any discussion that would preclude him running a combined company.
It would be more interesting to see Nvidia morph into a system company like Apple /Samsung/HP. They are already half way there. Somebody mentioned Nvidia customizing chips for them - that certainly is a possibility - Apple had a lot of experience with that kind of relationship in the past. Apple right now is clearly behind - their GPU performance superiority in the past was achieved simply by using a large off the shelf GPU area - nothing elegant about it - such brute force approach has severe limitations in terms of power, cost and manufacturability - all of which is coming to a head at Samsung fabs.
It looks more probable that Apple will be using a Tegra variant sometime in the future to address the lower cost market, along with the high end more effectively. This also becomes more likely as they distance themselves from Samsung, and rely on TSMC. I'm sure Apple is well aware & concerned that they might not be able to effectively, and consistently compete on cutting edge performance and delegate a number of engineering teams to co-develop with Nvidia engineering teams on custom variants.