Did "China" Just Buy The Most Important Company In The World?

Tyler Durden's picture

In the aftermath of last night stunning announcement that Japan's Internet giant SoftBank would acquire UK-based ARM Holdings, a company which makes chips present in virtually every mobile and "connected" device, for $32 billion, sending the semiconductor sector surging, questions emerged why the company is doing this.

On one hand, even the founder of ARM Holdings himself, Hermann Hauser said, told the BBC he believes its imminent sale to Japanese technology giant Softbank is "a sad day for technology in Britain". Hauser said the result of the Softbank deal meant the "determination of what comes next for technology will not be decided in Britain any more, but in Japan".

On the other, the stock of SoftBank has tumbled now that the Japanese market has reopened after a one-day holiday.

Bloomberg gave the trivial answer first thing this morning in a piece titled 'Why SoftBank Is Spending $32 Billion on U.K. Chip Designer ARM", which concluded the following: "Softbank Chairman Masayoshi Son sees ARM’s future in being inside the legion of products that are becoming internet-connected, from street lamps to air conditioners, washing machines to drones -- so-called “internet of things" devices."

Perhaps. However, a more provocative explanation has emerged courtesy of SouthBay Research, which when looking at the same deal, asks if China (yes China) "just acquired the most important company in the world? "

Here is SouthBay's explanation why:

ARM Holdings (ARMH) holds the keys to the future of electronics  That's not hyberbole.

Not only does ARMH dominate the world of mobile devices, it is rapidly penetrating all electronics: from consumer electronics to the computer network.

ARMH designs and licenses semiconductors.  Their designs are the core of the critical components of consumer electronics: smartphones, tablets, TVs, and so on.  For example, most of today's tablets and phones run on Qualcomm chips: they did $26B in sales last year.  These chips re-package ARMH designs.

As electronics continue to penetrate everything from cars to refrigerators, they use ARM designs.  The Internet of Things (IOT) uses ARMH technology.

Like a spider in the web, ARMH sits firmly at the heart of the future of all electronics.

I have reviewed ARMH two different times.

First, as the life preserver to save Intel.  Intel is unable to survive in the post-PC world and ARMH was a way to buy their way back in.  Unfortunately, anti-trust was probably a factor in preventing the buy.

Another time I mentioned ARMH was when discussing the future of China's semiconductor efforts.

After airplanes, semiconductors are a major capital outflow.  The South Korean and Taiwanese economies are driven by semiconductors.  Beyond economics, semiconductor strength enables national security strength (super computers).   For this and other reasons, the Chinese government has made a domestic semiconductor industry a major strategic goal.  Even going so far as to earmark $10B for intellectual property development.

The fastest path is acquisition, and ARMH embodies the future.  For China that presents two problems: price (ARMH was $19B last week) and politics (China buying the crown jewels of the internet touches some nerves in the Western world).

Today Softbank announced an offer to buy ARM Holdings.

Softbank = China

Softbank is a Japanese company best known for owning Yahoo Japan and Sprint.
With their background in telecommunications and the internet, why would they want to buy a major semiconductor company?  And why, with $89B in debt, is Softbank adding another $31B?

The answer: Softbank is not what they appear.  What isn't as well known is that Softbank is actually a major player in China's internet economy.
For starters, they bankrolled Alibaba.  They control 32% of Alibaba, and through Alibaba, they dominate the Chinese internet economy because Alibaba has invested in the top internet companies in China: Weibo, for example.

Although based in Japan, Softbank is very much a Chinese company.

The fact remains that, despite the $10B budget, China has yet to land any major companies.

Perhaps Softbank on its own is front-running a bigger China budget.

Or perhaps Softbank was tapped to be the buyer, but not the ultimate buyer.  After all, who is lending them to $31B to close the deal?  And Softbank will earn a lot of political credits for doing a favor for the

Chinese government effort to ramp-up a semiconductor company.

Don't be surprised if we see an announcement in a year or two that ARMH is up for sale and the buyer is a major Chinese company.

This was a brilliant move.

Comment viewing options

Select your preferred way to display the comments and click "Save settings" to activate your changes.
4shzl's picture

A chink in our armor?

NidStyles's picture

Naw, and protectionism is "EBIL"....

santafe's picture
santafe (not verified) NidStyles Jul 18, 2016 9:10 PM

China= busy making money.
America= busy setting the Middle East on fire.

erkme73's picture

And the spammer is back. 

NidStyles's picture

Hard to make money when you don't suck Jew cock.

Stuck on Zero's picture

A bit of an exageration:

Like a spider in the web, ARMH sits firmly at the heart of the future of all electronics.

ARM is so 1980s. China is developing its own architecture. LLVM gives any architecture software capabilities. MPMD architectures and cloud on a chip offerings are the future for powering IoT devices and machine intelligence.  ARM is not the future.

Automatic Choke's picture

I still design new product w ARM7 cpus....the perfect compromise between tiny embedded and massive power. It is a great architecture

Stainless Steel Rat's picture
Stainless Steel Rat (not verified) Automatic Choke Jul 19, 2016 12:29 AM


"Orbotech makes inspection and repair tools that help factories avoid errors as they pump out increasingly complex chips for smartphones and tablets."

Here, let me fix that random number generator for you...

cheech_wizard's picture

A great architecture? 

ABSOLUTE BULLSHIT. Would you care to take the discussion offline or have me embarrass you in front of everyone?



Barney Fife's picture

I'll bite on that cheech. Let's have a friendly debate. 

You have seen my other comments so you know my angle. For what it does in the size, cost, and power range that it exists in, ARM is a great architecture. 

Rule of engagement: You have to compare architectures of similar constrained design criteria. Fair warning. I know these devices inside and out and have even done a few dozen PCB layouts on them personally. 

Automatic Choke's picture

go ahead cheech, embarrass me.   i'm a physicist, so i only slum as a controls designer, but i do a lot of slumming.   i like them because i can design extremely robust tiny systems for embedded applications -- no OS, no high-level language compilers -- everything in bare assembly and all code belongs to my clients.  my clients tend to be folks who have been bitten on the ass by hidden OS latencies, weird compiler interpretations of simple instructions, etc etc.   The ARM7 is not the latest and greatest technology, but it is a solid platform that really meets the tiny embedded constraints while having impressive credentials, albeit no fpu. 

i also like older technology because most the errata have already been dug out, and because EVERY platform has weird quirks and contortions you must know.....so staying conversant with one has a lot of benefits.   i can whip out a new design using the NXP LPC21xx or LPC23xx family extremely fast, having lots of templates.....both board design and fully debugged code.    works for me, and works for my clients.

your turn.


cheech_wizard's picture

And the US defense companies design with 20 year old FPGA's... Want to ask me how I know this?

(Tunnel vision is a horrible thing)

Freddie's picture

What about the Russian defense stuff?   Their stuff used to be weak but they are great in software and very good software can help fix middle hardware.

Barney Fife's picture

ARM is so 1980's? I don't think so at all but never would I say that ARM will always remain king of the hill. Hell, less than 10 years ago TI, Freescale, and others along with ARM carved up the mobile devices scene but now ARM seems to be in everything. Last I heard ARM dominated with 90+% share of the market. 

So in that regards, it has no where to go but down. 

On another note, LLVM is a compiler. So your statement regarding that really makes no sense other than being ultra-obvious. Did you mean something else by it. 

If you want to see a badass SoC, take a look at the ARM, A72 processors that are out. Absolutely unbelievable how much firepower can fit into a device that might be 11x11 mm at best. 

Stuck on Zero's picture

LLVM is not a compiler. It is a transition mechanism (HAL) between an architecture and a compiler. It enables almost any CPU design to handle standard software.  This means that China can now produce its own designs and abandon ARM and licensing fees. Next will come Korea, Japan, and the rest of the world.

spqrusa's picture

Actually LLVM is both a compiler front-end and back-end. iOS and OSX software is now mostly built on LLVM.

Mobile CPU micro-architectures have been stale for the last decade. ARM's latest micro-architectures are not much different from Intel CISC designs of two decades ago.

booboo's picture

That bridge troll at the eccles building should buy it, she's got the roll.

giovanni_f's picture

The bridge troll is interested in assets that carry REAL value only: mbs, government bonds, car & student loans.

Badsamm's picture

Where we are going, will it matter?
"I don't know how World War III will be fought, but World War IV will be fought with sticks and stones"

putaipan's picture

uh oh.  for all of you guys cheering on the brexit .... you'd better check out webster tarpley's perpective on-  http://gunsandbutter.org/

then come  back and read this article again. add a little murrdoch engeneered may coup and voila- wercome to the new new red shield menace.

striped-pad's picture

Here's a weird couple of things I heard on "The World Tonight" (the "serious" radio news programme on the "serious" BBC Radio 4) last night:

  1. Mention of the phrase "deep state" (although strangely in relation to a hidden power in opposition to the visible state in Turkey rather than being the real power behind the scenes)
  2. Casual but serious mention of the possible political influence of freemasonry in the UK, by analogy with the divided loyalties of some people in public office in Turkey.

I can't help feeling something's afoot...

Crash Overide's picture

Remember when Malaysian flight 370 went down and along with it a bunch of Freescale Semiconductor employees that in their death their patent rights ended up in Rothschild hands?

Just saying...

Malaysian plane: 20 passengers worked for ELECTRONIC WARFARE and MILITARY RADAR firm


Rothschild owned Blackstone Group benefits from missing flight 370, becoming primary patent holder of new technology, reports say



Itgoestoeleven's picture

name change comming. ARM to CHIN

HenryHall's picture

So why on earth is the British government allowing this sale?!

ARM is far more important than military adventures.

TradingIsLifeBrah's picture
TradingIsLifeBrah (not verified) HenryHall Jul 18, 2016 9:20 PM

lol, ARM isn't that important (see their revenue, don't even need to go down to net income to see their irrelevance).  Softbank also said they plan to keep the headquarters where it is and will double employment over the next few years to keep the politicians at bay.

HenryHall's picture

ARM may not be all that important (more exactly may not be large) as a profit making machine. But ARM is hugely important as a means of converting specifications for digital processors to semiconductor designs (at least netlists and usually masks).

And the importance of that is that control of ARM means control of ensuring that it is possible to build complex systems with many (20 or so) layers of architecture AND be certain that there are no backdoors that let your system be compromised by malware. If you want to be 100% certain that nation state intelligent agencies cannot get into your systems then you need to own and control ARM (or own Intel).

20 layers is no exaggeration. For example quantum device physics, semiconductor physics, 3-D transistor design, doping plans, metastable units, clocked circuits, logic elements, microinstruction set, pipelined and out of order processor, macro instruction set, chip initialization, firmware, I-O structures, kernel, superstructure, middleware, database, applications. Those a not a complete or reliable guide, I'm too long out of the industry, merely an indication of how complex it can become. Like most things in life :-) the closer you look into it the more complex you understand it to be.

Urban Roman's picture

Yup. Exactly.

This irons out all those pesky "intellectual property rights" for the Asian companies that were already making all these parts.

Now they will have control of the masks and all those opaque 'blobs' in the cold-boot procedures.

cheech_wizard's picture

>If you want to be 100% certain that nation state intelligent agencies cannot get into your systems then you need to own and control ARM (or own Intel).

Keep your delusional remarks to yourself. I did verification work as a contractor for Qualcomm for their first 64 bit chip. Worked for Intergraph's Advanced Processor Division, then SUN Microsystems, then AMD, then Qualcomm... I hate to tell you this but ARM is woefully dated. The stuff they are doing with ARM now was done by SUN Microsystems 20+ years ago.

Care to state your credentials?

Standard Disclaimer: Whether you care to believe it or not in 2014 I showed Qualcomm an UltraSPARC-I architecture manual from the 90's on how SUN did trap tables. Their lead assembly coder took one look and immediately adopted it for their verification environment because of it's sheer simplicity and elegance. Yes, ARM basically sucks. But like Microsoft, which also sucks, happened to be widely adopted. (I don't know why the world is ruled by the lowest common denominator, it just is...)

cheech_wizard's picture

Let me expound on this...

ARM now adds cryptography instructions supporting AES and SHA-1/SHA-256.

(Starting with the designs first laid out back staring around the 2013 - 2014 time frame...)

Now compare this to SUN Microsystems 6 years earlier...

With the introduction of the UltraSPARC T2 in 2007, the processor’s on-chip cryptographic support was further ex- tended. Support for Elliptic Curve Cryptography (ECC) – both prime and binary polynomial fields – was added to the MAU. Additionally, a per-core unit to accelerate symmet- ric ciphers (RC4, DES, 3DES, AES) and hash operations (MD5, SHA-1, SHA-256) was also introduced. The two sub-units can operate in parallel, such that each core’s ac- celerator can concurrently perform, for example, a public- key operation and a bulk cipher operation. These operations also happen in parallel to instruction processing on the re- mainder of the core. In other words, each core’s accelerator can concurrently perform an RSA operation and an AES op- eration, in addition to the 8 hardware threads per core that

can continue to execute unimpeded

Standard Disclaimer: At the risk of repeating myself... ARM architecture - dated and essentially crap.

Freddie's picture

Was the Sun design Andy von Bechtolsheim stuff?  I forget the CPU he came up with before Sun was folded up.

Who has good stuff today?

Barney Fife's picture

This is a totally unfair assessment of ARM. What ARM does isn't to make the most advanced cores or SoCs on the market. ARM makes small and cheap devices. I have been involved either directly or indirectly with instantiating at least a dozen ARM processors in my time, and when we pick a uP, and you should know this, we pick a system that has "just enough plus margin" with the lowest cost and smallest footprint possible. 

We're not looking for top of the line due to cost, size, heat, and reliability. ARM is exceptional in their ability to spin systems that have a high value to cost and size ratio.  Remember, the design margin for a cell phone is extremely small, even compared to many PCs, so why instantiate overkill in to the device if the result is a hot, large, and expensive phone that requires a heatsink and a 1 kg battery?

So in my opinion while you are correct in the purest sense, you are discounting ARMs unique ability to pack so much into a device so small. THAT is what makes their products so incredibly valuable. 

Give an engineering team a massive budget and no constraints on footprint or thermal profile and we can come up with some mind boggling systems with performance metrics that will bake you noodle. THAT is not impressive. 

The real high tech comes from the ability to pack a ton of functionality into a low-cost, small footprint. So absolute performance is never the sole benchmark of high tech. Value to performance is. 

cheech_wizard's picture

Here's my point...(and thank you for making my argument for me as well) Let's not call it a "great architecture" again. Because it isn't. Never was, never will be.

Unless of course you are one of those people that believe Microsoft has a great operating system.

>you are discounting ARMs unique ability 

Not discounting it at all. It's a lot of shit (and I do mean shit) packed into a small die... bravo...

Here's one of those thought provoking questions you should run around in the back of your head.

Why was Apple able to deliver an ARM 64bit processor (namely the Apple A7 starting from September 20, 2013  to now) so far ahead of Qualcomm's 64 bit processor?

Proof: http://www.imore.com/apples-64-bit-a7-chip-hit-us-gut-says-qualcomm-empl...

Also: http://appleinsider.com/articles/15/01/21/bugs-in-qualcomms-64-bit-snapd...

Standard Disclaimer: Let me know when you come up with the answer, but it's a lot simpler than most people realize. If you need a clue, let me know, and I'll explain it to you. And when you do come up with the correct answer, you'll agree with me that the ARM architecture is indeed a piece of shit.

Urban Roman's picture


If you have a wafer fab, and want to make Sparc chips, that architecture is essentially public domain now.

Knock yourself out:

Overview of OpenSPARC Resources

Freddie's picture

Any of that great stuff done by the German guy who was a founder of Sun, left and later came back?  He is a great enegineer.  Sun was a cool company.   Andreas von Bechtolsheim.  Bill Joy was another one.

Sad they had to be sold off to the See Eye Aye pirate.

cheech_wizard's picture

Blame it all on that pony-tailed fucker who stepped in as Scott McNealy's replacement, Johnathan Schwartz (once voted as worse CEO in the coountry).

Schwartz was the complete douchebag who claimed "SUN is a sofware company"... Yes, and I am the King of the Unicorns...

Standard Disclaimer: I loved my time at SUN, because I got a chance to work with people far smarter than myself. At least until they hit the PhD level, because then the egos got in the way.

WTFUD's picture

It's a 360* reversal, the Chinese are offering Opium in exchange.

mkkby's picture

What a fucking joke.  Arm can go away tomorrow and few would notice.  Internet of things is something very few want or need.  Apple uses it's own chips.  If there were profit in it, many semi-conductor compaies in US or Korea could compete very quickly.

Barney Fife's picture

Uhh, Apple processors use the ARMv8 core. 

poeg's picture

No one ever reads the fine print.

phatfawzi's picture

this article is based on the assumption that technology is constant which it isnt. You can also argue they just over paid for outdated technology. in the near future chip based systems will be a thing of the past. 

Wild E Coyote's picture

Damn right.

If this story is True, then China is getting another IBM type outdated technology. Whatever listick you put on it, Lenovo is just a pig. 

TradingIsLifeBrah's picture
TradingIsLifeBrah (not verified) phatfawzi Jul 18, 2016 9:17 PM

You are exactly right.  ARM "controlled" the smart phone market but then profits in the market went to nil.  Intel ended up backing out of its push for smartphones because it saw there was no value there.  In truth there will be no value in the Internet of Things either (like how Brazil was the new hope, then China, then India, now the US, etc), its the latest fad but there will be a lot of competition.  


Softbank paid a huge value upfront for old tech.  Any new tech is going to need new dollars of investment (R&D) so what did they really get.  Some patents and a workforce for a 23x revenue multiple.  This screams overpayment and will likely end horribly for Softbank but Softbank is riding its Alibaba win down to its grave.  When the fraud of Alibaba is exposed there will be massive writeoffs all over Softbank if it doesn't just go out of business completely (not sure if they are levered up, my guess would be they are but I don't care enough to look it up)

ACES FULL's picture

You can buy whatever you want if you have access to enough fiat. Just digits on a screen. Its difficult to comprehend having the ability to conjure fiat out of thin air and be able to buy REAL stuff with that fiat. Its such a simple plan that I guess the difficult part is knowing that people are willing to let it happen. Nevermind,whats on TV?

HenryHall's picture

The hard part is not making or buying a machine to do whatever it is you want done.

The hard part is making (buying not an option) a machine that does whatever you want AND being certain that the machine is not capable of spying on you and revealing your secrets to your enemies.

You can bet that Iran's centrifuges were not controlled by Microsoft Windows after they learned about stuxnet. Unfortunately however the spyware/malware can be in the CPU itself if you buy the CPU or a computerized device that includes the CPU.

The solution is to buy ARM itself or or buy from ARM and trust that ARM is not controlled by NSA/GCHQ. And build your own system using only people you trust.

Automatic Choke's picture

um...what I remember reading (and I could be wrong here, i'm not an expert) was that the centrifuges weren't controlled by windows (and why would they be?).....they were controlled by Allen Bradley PLCs.     The delivery vehicle was Windows.


Barney Fife's picture

TI exited the smarphone processor business about 6 or so years ago due to lack of profitability. My company was using TI processors for what seemed like forever so when this happened we were shocked. Now we use the QC Snapdragon series.