Get email delivery of the Cadence blog featured here
This year's Linley Mobile and Wearables Conference is coming up next week on July 26 and 27 in the Hyatt in Santa Clara. I'm sure Linley Gwenap is furiously rewriting his keynote on the mobile industry in the light of Softbank's acquisition of ARM, announced unofficially over the weekend and officially this morning. Linley's keynote, Mobile Market Evolution, opens the conference at 9am the first day.
The keynote on the second day is by Jim Morrison, the VP of Competitive Technical Intelligence at Chipworks. Chipworks is the company I think of when it comes to detailed information about chips. For example, when a new iPhone comes out, they immediately buy one, rip it apart, and then decap the chips and start to look at them under electron microscopes and other sophisticated equipment. They can then tell everyone who manufactured it, what process it is in, which ARM® processor is in it, how many cores it has, what is the GPU, and so on. You might guess that they are based in Silicon Valley, but in fact they are in Ottowa, Canada.
That keynote is immediately followed by a presentation by Cadence on Multi-Purpose, Low-Power DSPs for Mobile and Other Markets by David Heine. David designs DSP instruction set architectures (ISAs) and oversees DSP processor development for Cadence Tensilica high-performance, low-power multi-purpose DSP designs. He will announce a new Tensilica DSP, built on proven Xtensa technology, that meets the need for mobile for an energy-efficient DSP supporting computationally intensive workloads and multiple data types and precisions. Cadence will also have a table at the end of the first day when the traditional reception and exhibits takes place. Not quite sure how much we will be revealing there, since David's presentation is not until the following morning.
The mobile conference also includes sessions on:
I will be at the conference so expect a few blog posts on what is going on in mobile and wearables in the weeks afterwards.
So what about Softbank acquiring ARM? At one level, I think it might just be a financial transaction, as opposed to a major strategic move for Softbank to own their own technology in the space. After the Brexit vote, the British pound has declined 10% against the dollar. But it has declined 21% against the Japanese yen. Since interest rates in Japan are negative, Japanese companies are penalized for holding cash. Having said that, Softbank has $24B in cash, although I'm sure a lot of that is not in Japan. Simon Segars told the Wall Street Journal that the deal was negotiated in just the last two weeks. Softbank's CEO also told him that ARM would continue to operate as it does today. “There’s no intention for SoftBank to change our culture, our team," Simon said.
Softbank has interests in all sorts of businesses. They own a controlling interest in Sprint. When the acquired it a few years ago, everyone expected them to turn it around but Sprint has continued to lose money. They also own $55B of Chinese company Alibaba (roughly 28%). Yahoo owns $31B worth of Alibaba (almost the same as the market cap for the whole of Yahoo), which means Softbank and Yahoo are joined at the hip, especially since Softbank also own 36% of Yahoo Japan. Yahoo is basically up for sale, so they could end up aligned with a completely new partner. Since one of the suggested potential acquiring companies is Verizon, that would make for a very interesting dynamic, with the two companies working together on Alibaba and Yahoo Japan, and competing head-on in the US wireless market.
But ARM is "expensive" in the sense that they traded at at a PE of almost 50X. That means that with the 40% acquisition premium they are now valued at a PE of 70X. For comparison, Intel trades at 15X. Softbank/ARM will need to grow their revenue a lot to justify that price. SoftBank has already said that it planned to double the size of the UK team over the next five years. That would mean going from 1600 to 3200. They will also increase non-UK headcount.
If Softbank pumps a lot of investment into ARM, this could be positive for ARM licensees in creating a richer portfolio of processors. But otherwise I don't really see much upside for them, they would rather ARM remained independent. Indeed, there have been constant rumors of a semiconductor company acquiring ARM (often Intel is the suggested name) but a moment's thought makes it obvious that it doesn't make a lot of sense. Businesses that compete with their customers aggressively tend not to turn out well.
For more information, watch the 8-minute video interview with Simon Segars.
Next: Economic Uncertainty, the Global Economy and Semiconductors
Previous: Softbank Buys ARM for $32B