Atmel is Microchip???

[Harrkev]

Presented without further comment....

Exclusive: Microchip is undisclosed Atmel bidder challenging Dialog - source

http://finance.yahoo.com/news/exclusive-microchip-undisclosed-atmel-bidder-201207693.html

[ajb]

Assuming the story is legit, I wonder what such a purchase would look like.  Presumable the memory and other non-MCU parts would get rolled into Microchip's portfolio, but what would they do with the MCU lines?  I guess they'd probably keep the Atmel ARM parts, since they don't have any ARM offerings, but would it be worth it to keep producing AVRs, if only to keep existing Atmel customers happy, and if so for how long?

[FireFlower]

Google definition: A microchip is a term that covers all electronic circuits (like microprocessors) that are printed on chips of silicon and reduced photographically so that the circuits are microscopic. These circuits cover microprocessors, memory chips, display drivers amplifiers and so on."

[Ian.M]

*IF* their bid is accepted, Microchip already have some 8051 core devices in their lineup that they got from their SST acquisition, so keeping existing lines going for AVRs would be highly probable.   However they seem to be determined to make everyone suffer the suckitude that is MPLAB X, so I would expect a push to support the ex-Atmel parts under that and retire Atmel Studio.

[Wilksey]

Different toolchains, I would think it better for them to keep the Atmel Studio line, the MS VS platform seems more stable than the Netbeans platform they are using for MPLABX, but I guess limits it to Windows users (or emulation).

There is the ARM vs MIPS scenario, but they would be foolish to retire the ARM processors.

[Smokey]

I wonder if by Microchip standards the AVR hobbyist market (Arduino, etc) is even worth purchasing?  Some hobbyist might buy like a total of 100 AVRs over his lifetime, but one professional commercial production widget might use hundreds of thousands of PICs over that product life.

[EEVblog]

There is the ARM vs MIPS scenario, but they would be foolish to retire the ARM processors.

Microchip CEO Steve Sanghi told me that there is no money in ARM processors and it's a bad business to be in. Whether or not he still thinks that I don't know.

[EEVblog]

I wonder if by Microchip standards the AVR hobbyist market (Arduino, etc) is even worth purchasing?  Some hobbyist might buy like a total of 100 AVRs over his lifetime, but one professional commercial production widget might use hundreds of thousands of PICs over that product life.

The AVR commercial market is just as big as PIC's. And the AVR hobbyist market is just as small and insignificant as PIC's.

[retrolefty]

I wonder if by Microchip standards the AVR hobbyist market (Arduino, etc) is even worth purchasing?  Some hobbyist might buy like a total of 100 AVRs over his lifetime, but one professional commercial production widget might use hundreds of thousands of PICs over that product life.

The AVR commercial market is just as big as PIC's. And the AVR hobbyist market is just as small and insignificant as PIC's.

 Yep, the tail does not wag the dog in those big corporate mergers and acquisitions. It's all about debt levels and leveraged growth and investment capital availability.

[coppice]

There is the ARM vs MIPS scenario, but they would be foolish to retire the ARM processors.

Microchip CEO Steve Sanghi told me that there is no money in ARM processors and it's a bad business to be in. Whether or not he still thinks that I don't know.

If there is no money in ARM processors, are Microchip getting out of the MCU business? Having any other core doesn't add value, unless its a specialist core, like a C2000 with its CLA addons, which can do things well that current ARM cores do poorly. If ARM MCUs can't make money, the Microchip MIPS based MCUs are also doomed.

[coppice]

Were the reasons for Microchip's original bid for Atmel in 2008, and its subsequent demise, ever properly explored in the press? It was canned as the world's economy worsened, but was that the true cause?

[ajb]

There is the ARM vs MIPS scenario, but they would be foolish to retire the ARM processors.

Microchip CEO Steve Sanghi told me that there is no money in ARM processors and it's a bad business to be in. Whether or not he still thinks that I don't know.

Sounds like the kind of thing he'll keep saying right up until they unveil their new line of ARM MCUs.  I can see it being a bad business for them to get into--what with having to bring up an entire line of ARM parts from scratch and there already being so much competition--but if they buy Atmel they suddenly have an ARM portfolio that spans from M0 to A5 @ 600MHz.  I can't see any sense in throwing that away.

If anything, I could see maybe trimming the XMega series.  With Atmel's newest M0+ parts the XMegas fill an increasingly narrow niche with some really nice peripherals bolted onto the same ol' AVR8 core.  And I doubt they have the demand that the lower cost ATMega and ATTiny parts do.

[EEVblog]

Sounds like the kind of thing he'll keep saying right up until they unveil their new line of ARM MCUs.  I can see it being a bad business for them to get into--what with having to bring up an entire line of ARM parts from scratch and there already being so much competition--but if they buy Atmel they suddenly have an ARM portfolio that spans from M0 to A5 @ 600MHz.  I can't see any sense in throwing that away.

Possibly. Depends upon the financials.
Having inspected Atmels books they will know the score on that.

If anything, I could see maybe trimming the XMega series. 

Two things that are most likely:
a) They'll trim anything that doesn't make money.
b) There will be much pressure from other Microchip departments to not harm them with competing groups. Whether not that makes financial sense. Doesn't mean they'll win, but there will be a lot of pressure.

[Stonent]

Well I suppose he would say that since Microchip is a MIPS house.  But even if he's correct, ARM seems to rule the embedded applications processor market.

[nctnico]

Well I suppose he would say that since Microchip is a MIPS house.  But even if he's correct, ARM seems to rule the embedded applications processor market.

Actually MIPS is/was used a lot by Broadcom but since Broadcom usually only deals with very large volume customers the number of products is limited. I wouldn't be surprised if the MIPS cpus have a double digit market share but ARM definitely has the main stream traction which MIPS lacks.

[coppice]

Well I suppose he would say that since Microchip is a MIPS house.  But even if he's correct, ARM seems to rule the embedded applications processor market.

Actually MIPS is/was used a lot by Broadcom but since Broadcom usually only deals with very large volume customers the number of products is limited. I wouldn't be surprised if the MIPS cpus have a double digit market share but ARM definitely has the main stream traction which MIPS lacks.

Broadcom and others have been high volume producers of full scale MIPS machines, with MMUs, running full scale operating systems like Linux. A large percentage of WiFi access points and routers, for example, are based on MIPS cores. However, I think Microchip is the only licencee MIPS has so far achieved for their MCU level cores.

[Wilksey]

From what I can see the Microchip MIPS core is a very cut down version?

Nearly everything had a PIC or an AVR in a few years ago, these days, nearly everything has an ARM in it!  I can't see a valid argument for there is no money in it, even if you take into account licensing the core, the peripherals and everything they add around the core make the devices much more sellable, not only to professionals, but to hobbyists as well, it won't be long before we see an Arduino with an M0 on it (or is there already one?!)

[Ian.M]

One thing's for sure: There will be a great sound of gnashing of teeth from the direction of the AVR/ATmega fanbois and  the eternal "Which is better?" flamewar will become even more stupid.

[nfmax]

A goodly selection of reasonably powerful ARM SoC devices, with Microchip's track record of keeping silicon in production for as long as the market wants it, would be a strong contender for new designs. At least round these parts

[nctnico]

Let's not fool ourselves by saying Microchip is the only manufacturer who keeps devices available for a long time. Any manufacturer who is in the lower volume industrial business (for example: NXP/Freescale, TI, etc) keep their products active for a long time. For example NXP's earliest ARM device (LPC2106) can still be bought after more than a decade.

edit: changed to LPC2106 which was introduced in 2003

[tom66]

I think what Steve Sanghi is getting at is that he doesn't believe Microchip can compete well with ARM offerings available from TI, NXP, Freescale, etc. What would Microchip's ARM chips offer that the competition doesn't?

[free_electron]

MicroMel
AtChip
AtMicro
MelChip

[mariush]

I'm wondering what such merger would mean from a chip manufacturing point of view.

As far as I know Atmel has several fabs but I don't know how they compare feature wise with the microchip fabs. Microchip has a bunch of fabs and as far as I know they're doing pretty good at keeping pics available at distributors.

I'm thinking, would they adapt atmel designs (takes time and money for validation i guess) to manufacture them on the lines where microchip makes their mips microcontrollers (to keep their fabs busier, be more efficient, whatever) and sell/kill some of those atmel fabs?

[retrolefty]

I'm wondering what such merger would mean from a chip manufacturing point of view.

As far as I know Atmel has several fabs but I don't know how they compare feature wise with the microchip fabs. Microchip has a bunch of fabs and as far as I know they're doing pretty good at keeping pics available at distributors.

I'm thinking, would they adapt atmel designs (takes time and money for validation i guess) to manufacture them on the lines where microchip makes their mips microcontrollers (to keep their fabs busier, be more efficient, whatever) and sell/kill some of those atmel fabs?

 It's been a while but last I recall Atmel got rid of there own fabs several years ago during a down cycle.

[Wilksey]

Microchip have only ever discontinued a handful of devices over the years as far as I am aware?

Of course at present, everybody and their dog are doing ARM, so a different processor core (MIPS) is sometimes a refreshing change, but if Microchip took over Atmel their product line would grow stronger, so I don't think they would discontinue the ARM line, as it just doesn't make sense, they are more feature rich than their PIC32 offering, and can run Linux on some of the higher end Atmel ARM processors.

Microchip AVR...Has a ring to it 

I've never understood the "fanboy" arguments, if you use a particular part whether it is because you were introduced to it by someone/something and continued using it, then you will naturally stick to it, you won't be able to jump to another chip as easily as sticking with the "devil you know", I use both, and I use whichever better suits my application and what other engineers are used to, makes it easier for them to do some debugging rather than all on me!