There's a ridiculous amount of consolidation going on the semiconductor industry these days.
Freescale (formerly Motorola) was purchased by NXP (formerly Phillips) in 2015. Qualcomm is in the process of buying NXP, although apparently there's been some difficulty getting it past the European regulators. Now it sounds like Qualcomm is going to get swallowed up by Broadcom. That's going to be a huge company.
Last year, Microchip bought Atmel, ADI bought Linear Tech, ON Semiconductor bought Fairchild, and Renesas bought Intersil. The year before, Intel bought Altera.
Any bets on the last two semiconductor conglomerates standing? I'm thinking Intel vs. Samsung. I hope whoever swallows up ON Semiconductor rebrands to the Fairchild name - there's almost something poetic about a reverse-Fairchildren split.
EDIT: I should also point out that "Broadcom" isn't really Broadcom anymore - Avago Technologies purchased Broadcom last year and took on the Broadcom name for itself. Avago was previously the semiconductor division of Agilent which in turn was a spinoff of Hewlett-Packard. Avago traces its lineage back to the semiconductor division of HP which was formed way back in 1961.
I've heard that mergers aren't (legally) supposed to be anticompetitive, but I've seen the banking and semiconductor industries undergo a merger-fueled competition implosion in the last few decades, so obviously whatever mechanisms we have in place are laughably insufficient. As a voter, what keywords should I pick up on to fight the good fight?
Chinese are just glad to have one big fish to swallow, instead of 20 small.
Qualcomm and Broadcomm both spent enormous funds trying to enter Chinese market. Both ended up with Pyrrhic victories
Qualcomm sold distribution deal for China to Allwinner guys where they chew into their already small margins - most Chinese smartphone makers simply do not pay "Qualcomm tax"
Broadcomm bought few Chinese wifi fablesses, yanked up prices. Most Chinese soc fablesses just switched to own internal IP
Infineon tried to buy Wolfspeed (Cree’s RF spin-off). Triquint merged with RFMD to form Quorvo. Who the hell comes up with these names? At least Triquint was a cool name.
Analog device bought Hittite and destroyed a good web site.
Avago isn’t even that clear cut. Avago sold SSD assets to Seagate, and then a bunch of other networking stuff as LSI/Axxia/Infineon to Intel. Then they picked up PLX and Emulex along the way.
Broadcom had also acquired Netlogic/RMI before that, so it is consolidation all the way down.
This might be a bit in the weeds but this is related to broadcom moving their head quarters back to the US. If you work in the merger arbitrage space you'll know about CIFUS, the US committee on foreign investment.
Since the silicon space has had so many deals recently and is looking at more consolidation, moving their headquarters back to the US removes one possible legal obstacle to buying up other semi conductors chip makers as CIFUS won't review any deals where a US company buys another US company.
I mean, broadcom's sweetheart deal with Singapore is almost over as well, so the financial benefits are probably over. That probably plays into their move as well:)
If this ploy works, look for several Chinese companies to buy US companies and looking to do reverse takeovers to get a US address for their company to attempt to avoid the CIFUS review.
> If this ploy works, look for several Chinese companies to buy US companies and looking to do reverse takeovers to get a US address for their company to attempt to avoid the CIFUS review.
That's not how review works. They look at the ownership of the company in question. It's not enough to just pretend to have a US shell doing the buying, while it hides a Chinese government parent. For reference, according to the CEO of Broadcom, his company is ~90% owned by US investors.
Just having a Chinese entity (whether Alibaba or a government corporation) holding a large ownership position in the company doing the acquiring, is enough to trigger review and a possible rejection depending on what's being acquired.
Beyond high level corporate jobs, would moving headquarters back to the US lead to more lower-class and/or manufacturing jobs? Or would the company just be based in the US on paper?
So Broadcom wants to be your one stop shop for everything networking related? Server, Switches Controller, Bluetooth, LTE, WiFi etc. On the WiFi chip there is basically only three players left in the consumer market, Broadcom, Qualcomm ( Atheros ) and Mediatek ( Mediatek is actually doing surprising well )
Another point is Broadcom seems to have a very good relationship with Apple so far. If you look at all the iPhone taken apart there are lots of Broadcom chips in it.
Off Topic: How do these Acquisition works? I get that it is a Share + Cash. The Share part is basically Qualcomm shareholder getting shares at the new company, but cash, we are talking about near 100 billion acquisition, where do the loans come from? Banks? Or could it be interest free coming from Apple?
You mostly answered it yourself. If it's an all-cash deal, of this size, they pay for it out of on-hand cash / short-term securities, and or raise debt from major banks or private equity types (eg Silver Lake Partners, as in the Dell/private & Dell/EMC deals). Depending on the cost of the interest on the debt, a company may choose to make the deal more or less cash, more or less debt, more or less stock.
Sometimes corporate partners do throw in on the deal. Microsoft for example helped finance the deal to take Dell private. It was obviously in their interest to have a healthy Dell corporation. I believe Microsoft put $2 billion into that.
It wouldn't be interest free coming from Apple. Zero chance of that. They also won't risk $100 billion. They could put in $10 or $20 billion however, with standard loan conditions including interest.
It's going to be difficult (expensive) for Broadcom to buy Qualcomm. QCOM is in a stronger financial position, despite the lower multiple on their stock presently. Qualcomm's quarterly earnings are about 50% (or more) higher typically than where Broadcom is at today. Qualcomm has $21 billion in cash, Broadcom has a mere $5.4 billion in cash.
It should really be Qualcomm attempting to eat Broadcom, rather than the other way around. In this case, Broadcom is being opportunistic when it comes to timing. The result of that will (if it goes forward) be very expensive for Broadcom shareholders. The sole reason Broadcom is able to attempt this, is due to the stock market bubble driving Broadcom's valuation extremely high (50+ times earnings) over the last year (the stock has climbed from ~$172 to ~$273).
There are ways to get these deals done without introducing PE though. Look at the recent SABMiller acquisition by AB Inbev (different sector, same thing -- close to $100bn).
That's why I mentioned raising financing via major banks. For a deal like that, JP Morgan, Goldman Sachs, (a dozen major Asian banks) et al. will happily step in and help arrange the debt necessary.
Broadcom is relatively cash poor when it comes to a deal this size, and their balance sheet is already a mess with negative $16b in net tangible assets. Annually they're burning nearly the equivalent of a quarter worth of net income on debt interest payments.
Given their size and the strong cash & income position of Qualcomm, there's no doubt major banks will be very willing to help despite the weak balance sheet. $50 billion in debt would nearly cost them all their current income. $100 billion would further chew into about 1/3 of Qualcomm's income. It's won't be an easy debt arrangement unless they lean heavily on their bubbly valuation.
Not to say, they have short term debt deals ready to be signed:
What Hock Tan did his whole career - he used patently massive short term debt leverage to buy something big and then sell less tasty parts quickly to get rid of that debt, no matter what harm it did to the company being bought.
Intel has respectable market share among WiFi client adapters for PCs, but they're missing from the embedded/mobile client device space and from the router/AP space. Overall, I doubt they matter any more than eg. Marvell.
It probably makes more business sense to do this mega merger/acquisition rather than continue to sue each other into oblivion over IP. The last settlement between QCOM and Broadcom was almost a billion dollar payment from QCOM to Broadcom... while that may not necessarily kill off Qualcomm, can't say it's not a major dent in their pocket after 4+ years of litigation.
The only reason I would want this to happen is if Qualcomm's obviously terrible and anti-competitive leadership would get the boot. And then the whole company's abusive culture needs to be cleaned-up.
Also, I wonder how serious Broadcom is about competing with Intel, and whether or not they would consider purchasing AMD, too, to aid in them in that effort. AMD's market cap is now only 1/10 of the price they intend to pay for Qualcomm, so it's almost a matter of "throwing some extra money on the table" at this point.
Broadcom could use Qualcomm primarily for mobile and IoT, and AMD for desktop, servers, machine learning, and automotive.
> Also, I wonder how serious Broadcom is about competing with Intel, and whether or not they would consider purchasing AMD, too, to aid in them in that effort. AMD's market cap is now only 1/10 of the price they intend to pay for Qualcomm, so it's almost a matter of "throwing some extra money on the table" at this point.
AMD effectively cannot be bought. There is a termination clause in the cross licensing of patents between AMD and Intel for x86/x86_64 in the event either company is sold. [0]
Anyone who bought AMD would immediately cause a huge issue for any company manufacturing x86 CPUs. AMD wouldn't be allowed to manufacture x86-based CPUs anymore, and Intel wouldn't be allowed to manufacture x86_64 CPUs. The only people who would conceivably benefit from such a situation would be patent lawyers who are probably already salivating at the mere thought of litigating that.
Intel already tried to raise this concern when AMD spun-out GlobalFoundries. I assume immediately after that one of their lawyers noticed they'd lose all rights to x86-64 if they tried to invoke the termination clause because Intel never mentioned it again.
IMHO this is effectively a non-issue and has been from the moment Intel was forced to abandon Itanium in favor of x86-64.
The other aspect is anti-trust. If x86 really does become single-source (Intel) then there is no evading monopoly status. The current US administration won't do anything about it but the rest of the world probably will. Intel would be forced to openly license x86 anyway.
Is there anything preventing a re-negotiation of that? presumably Intel wouldn't be too enthused about losing the x86_64 licence, so would be fairly motivated to sign a new deal with the acquirer.
looks like someone who did not wanted this deal leaked it to media. why would one try to acquire it now when the nxp deal is still in limbo? would it not be better to complete nxp deal and then make acquisition offer? doing it now will create more regulatory hurdles.
Broadcom is awful for the open source world, it is worse than Oracle on the software side I think. TI/Freescale and to some extent Qualcomm are open to OSS. I hope this acquisition to fail.
What are you talking about? Qualcomm is still hostile to free software, forcing binary blobs down your throat if you want to use an 802.11ac chipset. Ath10k might be open source, but everything interesting is hidden away by the proprietary firmware, like on Intel and Broadcom chips.
This would be my concern. Qualcomm is trying to buy NXP, who make a whole load of really great chips and microcontrollers (including everything made by Freescale). If the Broadcom attitude eventually makes it down to the NXP divisions, it could really ruin a good range of products.
Broadcom are employing Eric Anholt to build a fully open source, upstream DRM/Mesa Gallium driver for the VideoCore 4 (https://anholt.github.io/twivc4/) and assorted Raspberry Pi hardware.
Now that's hardly a big focus of Broadcoms business but you can pretty much run a fully open source GPU accelerated Linux userland on the Raspberry Pis with those efforts.
Bluetooth and WiFi is proprietary though, if I remember correctly. And that is what Broadcom makes money with.
RPI is an exception instead of norm. Other than that you got nothing, not just software, you got nothing on its chips either. I avoided Broadcom on all my designs as much as I can.
Freescale would love to open its GPU etc, however it is licensing the IP from other vendors that Freescale has no right to open source it, yes it would be nice if Freescale can get that part work. Otherwise, Freescale is fairly open on everything it owns and makes.
The Raspberry Pi having a non-closed bootchain, let alone a free driver is a feat of tens of thousands of people pressuring Broadcom, and a few people doing the hard work to free as much of the RPI as they could.
Having to run Raspbian just to boot the board was a dark time!
Freescale (formerly Motorola) was purchased by NXP (formerly Phillips) in 2015. Qualcomm is in the process of buying NXP, although apparently there's been some difficulty getting it past the European regulators. Now it sounds like Qualcomm is going to get swallowed up by Broadcom. That's going to be a huge company.
Last year, Microchip bought Atmel, ADI bought Linear Tech, ON Semiconductor bought Fairchild, and Renesas bought Intersil. The year before, Intel bought Altera.
Any bets on the last two semiconductor conglomerates standing? I'm thinking Intel vs. Samsung. I hope whoever swallows up ON Semiconductor rebrands to the Fairchild name - there's almost something poetic about a reverse-Fairchildren split.
EDIT: I should also point out that "Broadcom" isn't really Broadcom anymore - Avago Technologies purchased Broadcom last year and took on the Broadcom name for itself. Avago was previously the semiconductor division of Agilent which in turn was a spinoff of Hewlett-Packard. Avago traces its lineage back to the semiconductor division of HP which was formed way back in 1961.