ARM’s move into the broad Tablet and PC space is based on lining up as many partners as possible to attack Intel from multiple angles. It’s a strategy not so different from what Intel employed in the early PC days. However, the strategy is unraveling as Apple and Samsung have reached market share domination without ARM’s merchant partners. The end game is still playing out as partnerships and alliances continue to form. The long term impact on ARM will be slowing revenue and earnings growth if it plays along the lines that I think.
But first, as always, an analogy that I think is relevant. 100 years ago, Henry Ford made two critical decisions that allowed him to dominate the auto market in the first 15 years. The first was the assembly line. The second – which I think is more commonly apropos, is the need to go vertical in the manufacturing supply chain building everything in house.
Back in the 1970s I took a factory tour of the Ford Mustang assembly line and the Steel Rolling Mill that is part of the River Rouge complex – the largest in the world when it was completed. I was more interested in the Mustang tour than the steel tour. However you get a good sense of what Ford could do and how he could sell cars for under $500 which in turn led to industry dominance until he forgot to refresh the Model T in the late 1920s.
ARM has a history of working with many chip companies to extend the architecture into every crevice. By having multiple suppliers it is able to have the upper hand in the pricing of its license and royalties. With the PC, tablet and phone market they have recruited a who’s who of semiconductor companies: Broadcom, nVidia, Marvell, Qualcomm, TI, Samsung, ST Ericsson, Freescale and others that I am sure I missed. Kind of like the 1927 Yankees Murderers Row lineup to beat Intel.
The business model in this space though is not going to hold up for one reason. The Tablet and Phone Markets are going Vertical. Apple and Samsung have their own internal developments that will shy away from the merchant market. They now control over 50% of the market in phones and Apple has 90% of the tablet market. Apple and Samsung are going vertical along the whole supply chain with DRAM, Flash, panels, processors etc.. Samsung is true vertical while Apple is a virtual vertical (partially paying ahead on capacity for lower component prices). Apple will especially need to squeeze pennies from every supplier as they face the likelihood of a subsidized Samsung business model.
If it hasn’t happened yet, Apple and Samsung will turn the tables on ARM and demand royalty re-adjustments and create most favored nations status for the two. ARM financial projections will have to come down and their sky high P/E will moderate. Given the market consolidation, I expect many of the above merchant suppliers to throw in the towel by the time of the 22nm generation.
Now here’s the interesting part. I believe Qualcomm and nVidia will survive and both will fight to win the Amazon, Nokia (and MSFT) and HTC Business – consider this the rest of the non-PC market. Qualcomm has an advantage with its Communications Technology but I am not certain that Snapdragon is meaningful to their overall future. NVidia will get closer to Nokia and MSFT because MSFT needs a hardware platform that combines processor and graphics to undercut Intel long term. Nvidia is rumored to be the lead on the Amazon tablet coming out this fall – which should have fairly good volume.
After Apple has marginalized or commoditized ARM it will make two more strategic decisions. The first is to enter Intel’s Fab – my guess is at the 14nm node. It could be with ARM but much more likely a multi-core configuration with one core x86 and the other ARM. The reason is that Apple is going to go to a dual O/S configuration on its Tablet and PC platforms in order to improve its customer experience and build on its Wall Gardens protecting the greater ecosystem.
In the internet and I’ll call it the consumer iCloud mode it will run off of iOS for lower power consumption and longer battery life. In business iCloud mode (running office apps) it will switch to x86 and run in MAC OS in order for the user to experience full performance and App compatibility. Intel’s ultrabook initiative is trying to replicate this – but it won’t feel as slick as Apple’s does and MSFT may try to block the implementation since Intel’s light mode will be based off of the Linux Meego O/S.
The final piece to Apple’s strategy is to arrange a nice marriage between Qualcomm and Intel for the benefit of just Apple. Apple will demand that Intel run Qualcomm silicon in their fabs to get maximum performance at lowest power and cost. The Qualcomm silicon will be used for iPhones and Tablets in some die-stacking package. Getting CPU and Communications silicon cost down is key to their long term battle with Samsnug. ARM wins in the Tablet and iPhone wars hanging on Apple’s coattails but it is less of a win than if the merchant players were all in the game.
Intel wins in Tablets and iPhone business through expanded foundry business for Apple that includes building ARM chips for iphone and ARM + x86 combo CPUs for Tablets and MAC Air notebooks. Add on to that the Qualcomm foundry business and it is a significant revenue upside. Qualcomm may have the option of increasing their business with Intel Foundry – for the purpose of selling chips into Samsung, HTC, and Nokia etc… As a long time participant in the mobile PC business, it is the shifting alliances that will be the most fun to watch over the next two years.