Better than Qualcomm, Intel or Apple.
As rumored this week, chip designer and supplier Avago is buying chip designer Broadcom in a cash-and-stock deal worth $37bn (£24bn). Broadcom is probably best known to Reg readers as the biz behind the BCM system-on-chips found in the Raspberry Pi and various other gizmos. The combined biz will rename itself Broadcom Ltd as …
Yes, it's a pity ex Philips NXP is doing so little and Texas Instruments gave up doing phone chips and then swallowed National Semiconductor.
Motorola Semi is split now into two also rans.
As fab building costs rise and markets become commoditised (Apple original iPhone almost stock parts) and fewer gadget makers, now just using bought in chips and not designing them, such consolidation is inevitable.
There are a tiny fraction of the companies making Semis today as there were in 1960s to 1980s.
Sanyo, Toshiba, Sony, Panasonic (Mitsushiba) all used to develop loads of chips.
The UK (Inmos, Ferranti, Plessey) all gone
Intermetal, Telefunken, Siemens ..?
I think this ex HP Avago purchase saves Broadcom.
On Panasonic I think you meant "Matsushita Electric Industrial"
Ferranti microelectronics fell to Plessey in 1988.
Plessey ultimately ended up as part of MicroSemi (Plessey to GPS to Mitel to Mitel Semi to Zarlink to MicroSemi).
The Plessey name has risen again and now owns the Fabrication plants in Swindon and Plymouth that were part of the former Plessey Semiconductors (Plessey to GPS to Mitel to Mitel Semi to Zarlink to X-Fab to Plessey)
Siemens Semiconductor became Infineon.
Chip cost is a big reason, I think, why these companies get together. The cost of making a modern SoC from scratch is MASSIVE. Many $10 of Millions if you want anything other than ARM standard designs (and they are not cheap)
And more to the point, I sold all my Brcm shares last week, just before they went up 20%, so a bit pissed. off.
I'm sure that Intel will look at it's low power portfolio now and re-consider ARM. Intel can afford it. It would be shame to lose our UK sweetheart but it is not large enough to resist a takeover from a giant like Intel unless AMD, already a customer, decides it wants it first. In a bidding war (still time to buy ARM shares) Intel would win for sure. Ironically, Broadcom (soon to be Avago) is an ARM customer. Boy this business is incestuous...
Intel used to have a big ARM range (obtained from DEC), they still have one ARM comms processor I think, the rest sold to Marvell.
There is no reason for Intel or anyone else to buy ARM, they are not a chip maker.
Intel, I think, still has a full ARM licence, Their biggest strengths are in marketing and Chip Manufacturing.
Perhaps Intel may make ARM chips again more generally, they seem though have a "Not invented here" ego issue as Intel processed and Intel Peripherals on SoC Arm makes far more sense than their vast expenditure in Mobile Sector, making them less x86 sales every year.
The SOCs I've seen inside recently have lots of CPU cores. ARM's advantage is that it can run Android apps but you only need one core to do that and it can be a cheap one. The other cores can be Tensilica or ARC or whatever as they tend to be running custom firmware for the SOC and don't need a third party dev ecosystem to succeed. As others have said Intel already has an ARM license and will always be able to have one.
Who owns the fabs is more interesting.
"Intel['s] biggest strengths are in marketing and Chip Manufacturing."
Chip manufacturing, yes.
Marketing? Not so much. Marketing couldn't save Itanium, for example. Intel said x86-64 couldn't be done. AMD proved it could be done, and here we are a couple of decades later, it was IA64 that was the loser in that particular race, despite the efforts of Intel marketing and others. Intel have continued the x86 success during that time, courtesy of the folks in Redmond. What else have Intel achieved?
Intel's biggest strength: the default chip where you need Windows. Where you find Windows, you find Intel. Where Windows is of no significance, Intel is of no significance.
Intel: the x86 company.
"there are no Apple and Linux boxes running on x86, right?"
There are no Windows-independent boxes I know of that are running on x86. Correction welcome.
Apple have become almost as much a member of the Wintel/IT department world as Microsoft, hence Parallels, Bootcamp, etc. How well would that have worked on PowerPC or ARM? It wouldn't have been much fun, therefore there's a Windows dependency.
Linux etc? A box with an x86 in it is typically that way because it needs to be able to run Windows (even if the end user in question chooses not to do so, e.g. with many x86 servers).
If you're Dell HP Lenovo etc you need your boxes to run Windows and there's no motivation in general to not use x86. Even HP are allegedly looking at ARM for servers, though obviously not for Windows boxes. Dell? They're so dependent on Intel that they daren't even thinl about that kind of thing.
If a particular product design doesn't need to run Windows, it generally doesn't need an x86. E.g. routers, set top boxes, random consumer electronics and any professional kit powered by a computer. Basically anything that isn't an IT Department box.
If x86 is so brilliant, where are the Windows-indepent x86 boxes?
Feel free to provide examples to the contrary. I've not seen any recently (I've got to go back to the days of Intel iRMX and things like x86-powered automation devices from Cegelec and Siemens. Maybe someone has more recent examples?)
Anyway, here's a slightly different question in case the previous one's too hard: what *have* Intel achieved in the last decade or two, outside the x86 world? (Hint: highlights may include selling most of their ARM business, abandoning their 'industry standard 64bit' (IA64) business when AMD made IA64 look technically and financially silly, and buying McAfee).
Apple didn't choose x86 to let its users run Windows. It did because PowerPC started to lag behind, and ARM is not a choice for laptops/desktop yet but for very low-end ones. It chooses x86 exactly to be able to keep on competing with Windows, and avoid to lose customers. It worked.
Most servers run x86 not because of Windows, just because all the other CPUs, PowerPC, Alpha, Mips, Itanium, SPARC, eventually had a far worse price/performance ratio, for most server needs and loads.
HP and others sold and sell their Unixes as well, not only Windows.
Without the 'cheap' x86 server available, Linux would have gone nowhere. Now, I guess it's the main platform is run up on.
After all, the first x86 server didn't run Windows at all, they run Netware and LAN Manager. For a long time, the unavailability of 64 bit Windows unable to easy use more than 4GB of RAM was a big handicap for heavy workloads.
The device that don't use x86 are those that don't need much power, but instead a cheaper, less 'hot' CPU. The main drawback of x86 CPUs are power needs and heat generation. Anyway, there are a big group of embedded devices 'Intel inside', just not the cheap low-end ones.
AMD 64 bit solution won over Itanium for the very same reason Windows won over OS/2 and other OSes: it was cheaper and fully backwards compatible, not better.
Wrong arguement/comparison. Intel makes most of its money from manufacturing chips, whereas ARM makes its money from licensing cpu/chip architectures, which any licensee can maunfacture in whatever way they feel fit.
The only thing that Intel could gain from buying ARM are the license revenues, but I don't think that the deal would get through anti-monopolisation legislation; apart from X86/AMD64 & ARM, plus perhaps ppc, the only other architctures that might make any degree of impact are the ones being developed in the East. But they're never going to get a real foothold in the West, so if Intel acquired ARM then Intel would have a complete monopoly.
Biting the hand that feeds IT © 1998–2021