You are currently viewing SemiWiki as a guest which gives you limited access to the site. To view blog comments and experience other SemiWiki features you must be a registered member. Registration is fast, simple, and absolutely free so please, join our community today!
The right price point, the right scale, and the right, mutually complementary synergy with QCT business. It would look like Qualcomm is buying into the IoT/IoE business; but what's wrong with that? Doesn't Qualcomm badly need to show a plausible growth prospect?
When it comes to Makers and gadgeteering, if any established semiconductor vendor could be said to have "street cred", it gotta be Atmel: the ecosystem, the channels, the many, many prototypes already out there with the Makers, etc. Foot in the door for Qualcomm in a much hyped growth industry that, with the combined oomph of the ailing giant and Atmel, could just turn hype into something resembling reality.
Now I am putting my hard hat on; let the rotten tomatoes fly ...
Lee Courtney: Atmel would complement QC in edge IoT and Wearable devices. But, given QC's expertise with power mgt, comms, and ARM IP why not plug any holes using a quick and dirty spin of an ARM M0 with integrated BLE radio and push thru existing QC sales channels and (much stronger) developer programs?
Qualocomm has been probably the biggest beneficiary of the mobile wave that came more than a decade ago. Apparently, IoT/wearables are the next wave and Atmel is one of the best positioned companies in this realm, if not the best positioned company outright.
It may certainly make a good match because technology opportunities come to companies who are prepared way before these opportunities descend. Like Qualcomm was ready for the mobile wave when it shifted from 2G to 3G.
There are many companies who began spending R&D to catch up on a wave, and in most cases, they weren't successful despite heavy resources and deep pockets. Remember Intel's mobile undertakings? So if Qualcomm wants to make it in the IoT era, it may consider not to reinvent the wheel and go out shopping.
Atmel was selling some of its RF ICs to Qualcomm way back in 2000, so there does appear to be some synergy between the two companies. There are some high-level questions about acquisitions, like:
Does Atmel want to be acquired and give up its autonomy?
Can Qualcomm just continue buying chips from Atmel and call it good?
Is there a similar work culture at both Atmel and Qualcomm?
There was a story in EE Times a couple of weeks ago titled "Atmel on the Block" and the title of the story is self explaining. Though Atmel has refused to comment on this.
In the early stage of IoT I think that most of the designs will be like the Fitbit in the sense that they will be built out of a handful of discrete parts (I accidentally typed discreet parts, I guess we need those in IoT too, security will be so important) and Atmel is very well positioned with its enormous portfolio of microcontrollers. Qualcomm could obviously design a microcontroller of their own, but a family of hundreds? Plus Qualcomm has all the wrong channels for that. The big question to me is whether much money is to be made in the low end IoT market.
The industrial internet of things (IIoT) seems to be where the money is. Especially if you include medical and automotive where you need the same sort of mature qualification processes that Qualcomm has in place for modems and mobile application processors. Parts need to be qualified on the networks. But it is not clear to me whether Atmel would help in that.
And the new device ecosystem. Over the years, both through the Arduino community and via its own marketing platform, Atmel has generated a huge amount of enthusiasm amongst the Makers of the world. While maybe 96-98% of all the Kickstarter gadgeteering may not lead to a viable commercial product, the remaining 2-4% could still speak for the future of smart devices: Pebble, Occulus, Ouya, etc, etc. And it's in this space that Atmel rules almost unchallenged. I doubt Qualcomm will be able to develop this kind of ecosystem from scratch. Look at Broadcom: its chip was used in the very popular Raspberry Pi, but Broadcom failed to leverage that into much of a business gain. Amongst established IC vendors, Atmel is unique in still being vibrant enough to ride the Maker wave, and its strong position there may lead to some real commercial success in the IoT hype. If Qualcomm emerges from the ongoing turmoil still interested in IOE/IOT, Atmel acquisition could very well get it there.
Great point from Dan re cultures of the two companies. In technical terms, I am particularly curious about the SW part: the IoT and Maker space are both calling for a platform play, and I suspect much of the differentiation comes in system SW, intuitive driver installation, API, etc. My (perhaps erroneous) impression is that Atmel leaves most of SW dev to its ecosystem, and Qualcomm - although it has made strides in recent years learning from MTK - may find smart device entrepreneurs to be more demanding customers than phone vendors. Look forward to comments from people who know more about this?
Paul: agreed re auto and med - lots of up-seeling potential for Qualcomm. Re "division of labor", perhaps customers can do sensing hub and thin client with Atmel, and use Qualcomm parts for aggregators? That being said, I am personally a believer in going horizontal with IoT (Nest-wise) rather than hierarchical IoT stacks. Perhaps Qualcomm can leverage its own expertise to offer easy self-organizing networks of Atmel nodes?
In IoT space, Qualcomm is well placed with its next generation 4G networks with MuLTEfire technology that allows running 4G on an unlicensed spectrum. Add to it the technology from CSR acquisition - Mesh protocol that adds security to Bluetooth standard. Qualcomm is working on increasing mobile network capacity with its MuLTEfire technology. It it is able to 4G ubiquitous like WiFi, it could be a winner in IoT standard space.
So, as far as Atmel is considered, even if Qualcomm remains with the same arrangement with Atmel as of now, it could be in advantageous position going forward. However, if Qualcomm can find enough money to buy Atmel, it could add to it in wearable portfolio. But money will be a question. Btw, TI is also eyeing at Atmel, so Atmel definitely would have a high price!
In processors, that would find their place in automotive, industrial, home,... Qualcomm would be self sufficient.
I am personally a believer in going horizontal with IoT (Nest-wise) rather than hierarchical IoT stacks. Perhaps Qualcomm can leverage its own expertise to offer easy self-organizing networks of Atmel nodes?
Vertical and horizontal, both are required for IoT. Verticals will need separate standards according to their specific requirements. Horizontal needs to cut through each of them for say, security, connectivity, capacity,... And there Qualcomm will be coming strong with its MuLTEfire technology with 4G and Mesh protocol that can provide secure connectivity.
Yes, TI has big ambitions for the embedded space, which is an Atmel forte. But right now, TI has a large MCU portfolio, which means a big product overlap with Atmel.
Qualcomm is really searching for its post-mobile identity. The IoT play for them could include MCUs at the edge, but the real strength would be to make a push in the gateway and infrastructure tiers. They don't need Atmel to do that.
What an Atmel play could do for them is provide end to end optimization of the software stack, particularly Thread, on any device. The recent news Microsoft wants to run AllJoyn over Thread could be really interesting. But frankly, *any* device would mean *any* MCU (or at least ARM-based MCUs), and Qualcomm might be better off staying MCU agnostic and building a bigger ecosystem.
As pointed out earlier, these are two very different companies, and they have two very different types of customers and channels. Atmel is the Arduino community, Qualcomm is carriers. One has no rules, the other has thousands of pages of rules. The good news is they don't conflict, but tying them together in a cohesive strategy could be a challenge. Software would be the key.
I'd also wonder about margins. Qualcomm's operating margin is like 30%, Atmel's is like 9%. Having run an acquisition strategy team for a division of Motorola, my experience says don't chase lower ASPs and lower margins unless there is some very unique capability that can't be obtained otherwise.
Atmel is a very different proposition for Qualcomm compared to Atheros and CSR.
If Cavium is on the block, I'd see Qualcomm headed there first.
And for IoT, Qualcomm's game is Gateway and Infra as pointed out by Don. Qualcomm is exactly doing that with their CSR acquision. Do they need another acquisition at this time, that too for the edge devices? In my humble opinion, they should wait, use 4G for secure WiFi like connectivity for IoT.
ATMEL's lead in the Arduino community is only sustainable with price. There isn't any specific IP there to protect Atmel.
In fact, ATMEL does not control anything at all in that platform. Every MCU manufacturer and even Intel have Arduino compatible kits. The value of IOT is not in the MCU silicon where price pressure will be intense. There is nothing in this business that would be interesting for Qualcomm, I believe. The value in IOT is somewhat in the devices and much more in the services. This might be of great interest for Qualcomm, of course, but they won't find any of that in Atmel.
Mollenkopf: Qualcomm plans to take part in chip industry M&A
Jul 29 2015, 13:27 ET | About: Qualcomm Inc. (QCOM)
"Qualcomm (QCOM +0.2%) is very likely to be some form of actor in the consolidation of the semiconductor industry," CEO Steve Mollenkopf tells Bloomberg. “The timing of which is always the debate.”
The remarks come a week after Qualcomm said it would (following pressure to do so from Jana Partners) consider a breakup, along with other strategic options. There has been some speculation Qualcomm's chip unit (QCT) could merge with Intel (has been seeing heavy mobile division losses) following a spinoff; antitrust regulators would closely vet such a deal. Others have argued Qualcomm could by an RF chipmaker such as Skyworks or Qorvo, or (with the goal of expanding into the data center) acquire an ARM server CPU developer.
Qualcomm hasn't exactly been averse to M&A in recent years. Noteworthy acquisitions since 2011 include Atheros (Wi-Fi chips), Wilocity (WiGig chips), and CSR (Bluetooth/Wi-Fi chips).
Update: Some other Qualcomm news of interest: Mollenkopf discloses he bought 15,815 shares yesterday at $63.31, and CFO George S. Davis discloses he bought 8,100 shares yesterday at $62.34.
Speaking of merger w/ Intel: sometimes rumor runs by proxy. In recent years, to expand market shares in China, Intel partnered up with a company called Rockchip, and Qualcomm with a chip company called Allwinner. In recent months there has been a lot of rumors about Allwinner+Rockchip consolidation, "mirroring" the on-and-off Intel+QGT speculation. Reminds me of the Cold War era, when USSR and USA would fight these wars of proxies in Africa, Asia, Latin America, etc.
Just wondering out loud: why did Qualcomm get such a heavy beating lately in the press and in the market? No single event could be interpreted as a fiasco, be it the EU legal trouble or the Samsung - Snapdragon 810 upset. Compare it to Intel: the gaping hole left by the PC decline is here to stay and impossible to plug - and Intel pretty much ran through the short list of "filler" candidates, without a plausible solution: digital home, tablet/phone, foundry business, etc, etc. And according to analysts its Data Center biz is showing weakness, as well. Now THERE is a Doomsday scenario brewing. While Qualcomm's problems are real, the decline is not as bad - and irreversible - as Intel's. My guess is that it's a matter of perception: Intel management has been more proactive spinning new tactics, while Qualcomm seems less so. Not much said since announcing the CSR acquisition. Could it be that this perceived lethargy, amplified by Jana, made the market uneasy?
Along the line of Atmel helping Qualcomm "get" IoT (or Maker) software: right now Qualcomm has no finger on the pulse of new programming languages beyond Java and C - Python and such. Neither through its network side of business nor Snapdragon-based higher-end IoT solutions. Looking at Qualcomm's SW track record (Brew, and phone turnkey) to date, perhaps a shot of Maker community's raw energy is needed to keep it on the edge. Qualcomm's AllJoyn and its works on standards/protocols, plus mindshare in the "Pro-Maker" developer community, can be a strong combination to both stimulate enthusiasts and calm the nerves of more traditional customers.
