Is Atmel on a decline or rise?

[FenderBender]

Well I've always been an analog guy, but after playing around with my new Arduino, I'm pretty hooked on this MCU stuff. I'm not saying Arduino is the end all solution to everything, but it has a good friendly IDE that is easy enough to learn and to get good at.

Arduinos use AVR MCUs. I was wondering if anyone has any expectations or thoughts about the direction of Atmel. Is Atmel still going to be around in 10 years? I'm fairly certain there are other companies like Microchip who might have an edge on them, and others like TI which practically have monopolies. Should I try learning on a PIC instead? I guess it all becomes relative once you learn the basic logic and ideas behind MCU programming, but I'm still curious.

I know this is a strange question haha. So if you have any ideas. Let's not try to turn this into PIC vs Atmel. Just the direction of Atmel.

[PeteInTexas]

From what perspective?

I know Radio Shack just started carrying all manner of Arduino.  That's good for sales I guess.

[gregariz]

Theres no doubt that Atmel has had problems with profitability. Microchip on the other hand has done pretty well in recent years so is in a strong financial and market position.

I wouldn't be concerned at all using Atmel however. They have a successful product line if they can keep their costs under control so I would expect even if they don't survive as an independent company the architecture will, even if under another name.

[EEVblog]

I know Radio Shack just started carrying all manner of Arduino.  That's good for sales I guess.

Arduino sales are down in the noise compared to the rest of the industry.

Dave.

[EEVblog]

Atmel were doing poorly for a long time financially.
In fact, when I spoke to Steve Sanghi (Microchip CEO) he pointed out very matter of factly that Atmel hadn't actually turned a profit since they started.
Whether or not that is true and/or how it relates to their overall financial situation (which does seem to show a yearly profit), I can't say. But considering that Microchip tried to buy Atmel, I figured he'd know all the details.
He also mentioned that Microchip were basically the only one turning a profit a the time (for equivalent microcontroller divisions). It might have been exaggerating, but this did come direct from the Microchip CEO.

Atmel also famously had major supply problems. Not only at the top level for the big customers, but also clearly visible down at the Digikey stock level. The Digikey level now seems resolved.

If you were to bet either Microchip or Atmel as being around in 20 years and supporting your chosen silicon, you'd certainly have to pick Microchip.

I would not be surprised is Microchip tried to buy Atmel again in the future.

But as for learning now, it doesn't matter. Pick one you like and run with it, or both, or four...

Dave.

[amspire]

How well Atmel do over then next few years will probably depend more on their maxTouch touch screen ICs then their processor ICs. They currently have some very good technology, but the question is how long before other foundries are making similar chips at half the cost, and can Atmel stay competitive?

[PeteInTexas]

I know Radio Shack just started carrying all manner of Arduino.  That's good for sales I guess.

Arduino sales are down in the noise compared to the rest of the industry.

Dave.

Perhaps it is but the old "get 'em while they're young" strategy can payoff later for Atmel as it has for others in the past?

[EEVblog]

Perhaps it is but the old "get 'em while they're young" strategy can payoff later for Atmel as it has for others in the past?

That's an interesting question actually.
Given that the Arduino environment is so highly removed from the actual use of and reliance on the AVR (that's why there are PIC Arduino's too), is someone raised on Arduino going to naturally chose AVR in future professional designs, as much as the early PIC starter naturally chose and stuck with PIC?
I think likely it's not the same, just given the investment in architecture knowledge that early PIC hobbyists had to go through for example.
It would likley be just as easy to move from Arduino to PIC or Ardunio to TI, than Ardunio to AVR.
But still will have a large effect in the future I'm sure.

Dave.

[gregariz]

I'd be inclined to place arduino more in the category of the Basic Stamp or Picaxe. More of a hardware intro for software guys.

[Psi]

It wouldn't surprise me if a new company, or division of existing company, manged to produce a new class of micro controller using an ARM core with plenty of speed, ram, flash and eeprom that was priced to meet the pic/avr. There are a few already such as the NXP LPC range but it doesn't seem to have caught on yet.

Ram is so cheap now that producing a micro with only 8KB of flash is a bit pointless.
They should be able to produce them with ~1MB of flash without much effect on the price.

[Bored@Work]

From a professional point of view Atmel IMHO went bonkers in recent years. Their XMega stuff was delayed and is broken (they are about to introduce a new revision where "everything" is supposed to be fixed - famous last words), their AVR32 MCUs are between a rock (cheap 8 bit MCUs) and a hard place (cheap 32 bit ARMs), they completely lost the plot with their new Studio 5 development environment and their ASF software "framework", and they have serious supply issues (home made, because they decided to go fab-less and now depend on booking in advance and being ready in time for manufacturing time slots at some chip foundries).

They do, however, have a winner with their touch screen technology which does go into a lot of smart phones, pads and the like.

But for micro controllers we now avoid Atmel. They are just not a reliable partner and only care for the super large volume buyers and otherwise have no problems to lie to customers as if there is no tomorrow.

From a hobbyist point of view, however, most of this doesn't matter. Many of their mega and tiny AVR micro controllers are cheap and easy to get. For comparison, try to get something like a Renesas H8SX or ST STM8L micro, especially if you live in a country where unfriendly distributors like Farnell and RS refuse to sell to hobbyists. And you can get, albeit limited, open source tools to work with Atmel AVRs. Atmel dev tools are unfortunately expensive and less reliable compared to other offerings, and they really hate supporting open source tools.

[mikeselectricstuff]

MCUs are a small part of Atmel's business, which is one reason they have historically had poor availability when demand increases on their other lines.

There was a time when they beat Microchip signifcantly on cost for similar capabilities but that is rarely the case nowadays.

[FenderBender]

Yeah some good points made. Atmel does have some good offerings. Their success probably depends on how good their R&D is and how quickly they can adapt to new technologies (touch screens etc.).

[Zad]

Less than a year ago Arduino were having BIG problems sourcing Atmel chips. We might think Arduino are big, but they are peanuts in comparison to most of the industry players. It pushes Arduino a long way down the pecking order.

For what it is worth, Wikipedia states that Atmel had an operating income in 2010 of +US$212M on a net income of US$423M

I think the bloke from Microchip was being rather economical with the truth, conveniently forgetting some facts.

http://ir.atmel.com/financials-statements.cfm

[FenderBender]

Less than a year ago Arduino were having BIG problems sourcing Atmel chips. We might think Arduino are big, but they are peanuts in comparison to most of the industry players. It pushes Arduino a long way down the pecking order.

For what it is worth, Wikipedia states that Atmel had an operating income in 2010 of +US$212M on a net income of US$423M

I think the bloke from Microchip was being rather economical with the truth, conveniently forgetting some facts.

http://ir.atmel.com/financials-statements.cfm

Yeah I forgot what video I heard Dave say that Atmel hasn't made a cent since day one...I think it was the Arduino vs PIC video. But I looked up the Wiki article and saw bigtime profits so I wasn't quite sure what the whole "no profit" thing meant.

And you're right about Arduino not being as big of an Atmel customer as we might think. I mean they do buy a lot of chips, but probably only a fraction of say...what Samsung buys.

[EEVblog]

I think the bloke from Microchip was being rather economical with the truth, conveniently forgetting some facts.
http://ir.atmel.com/financials-statements.cfm

Possibly, but "that bloke" is the head honcho CEO himself. He seemed very adamant about it and talked about it at length.
Maybe he was referring to just their microcontroller division or something?, or maybe they have some hidden debt that they haven't bought forward into their financials somehow? (talking out my arse, I know nothing about big company financials)

Dave.

[amyk]

That explains why it's so difficult to get samples from Atmel but Microchip almost gives them away in comparison.

[dfnr2]

If you're just trying to learn, there's nothing wrong with Atmel or Arduino.  The AVR instruction set is easy to follow, and there are good compilers and development environments available. 

However, I would not put any Atmel part in a real design unless it were second sourced.  About 10 years ago, I was starting up and looking for a controller for a design, and I really wanted to use the Atmel part, but finding them in any quantity was a nightmare.  So I went with Microchip, bought a good compiler, and repressed my distaste for the architecture and learned it well.  In retrospect, it was a great decision.  Microchip parts have long life cycles, and even devices based on obsolete technology (the "C" Eprom parts) are available easily.  They have excellent support for large and small customers; ordering preprogrammed parts is easy and robust. 

In the course of keeping up to date, I regularly check in on Atmel, and sure enough they continue to fall short.  However, the arduino is so easy to use, I picked up several of them just to bang out one-off test rigs, simulators, etc.

If you are designing a product, one thing that is nice to do when deciding on parts is to get on Digikey, Mouser, etc., and do a parametric search, making sure "in stock" and "RoHS" are checked, check your parameters, and sort by descending stock quantity first, then by increasing price.  Check on a couple of sites, and on octopart.

You will want to pick the right tools for the job, but ultimately, if you learn a small range of tools to cover the bases, and buy the tools, and learn them, then you will end up going back to just a few architectures, such as (PIC12/16, PIC24, and NXP), or (PIC12/MSP430/ST) so you can hit the ground running every time. 

Dave

[gregariz]

I think the bloke from Microchip was being rather economical with the truth, conveniently forgetting some facts.
http://ir.atmel.com/financials-statements.cfm

Possibly, but "that bloke" is the head honcho CEO himself. He seemed very adamant about it and talked about it at length.
Maybe he was referring to just their microcontroller division or something?, or maybe they have some hidden debt that they haven't bought forward into their financials somehow? (talking out my arse, I know nothing about big company financials)

Dave.

Roughly 40% of all CEO's are psychopaths. A significant number of the rest are malignant narcisists.

[vis1-0n]

I think the bloke from Microchip was being rather economical with the truth, conveniently forgetting some facts.
http://ir.atmel.com/financials-statements.cfm

Possibly, but "that bloke" is the head honcho CEO himself. He seemed very adamant about it and talked about it at length.
Maybe he was referring to just their microcontroller division or something?, or maybe they have some hidden debt that they haven't bought forward into their financials somehow? (talking out my arse, I know nothing about big company financials)

Dave.

Mostly out of date data, I think from around the time of the start of the recession (2009). Atmel have since sold their fabs, and not invested much (low R&D & acquisitions - can't find the charts right now). Atmel have therefore been on the rise (with the industry actually) and pleased shareholders who bought low, selling high - these don't care about actual profitability.

[dfnr2]

Mostly out of date data, I think from around the time of the start of the recession (2009). Atmel have since sold their fabs, and not invested much (low R&D & acquisitions - can't find the charts right now). Atmel have therefore been on the rise (with the industry actually) and pleased shareholders who bought low, selling high - these don't care about actual profitability.

Ugh.  Sounds like shareholders first, customers last.  Not that I'd base such decisions on a single random forum post from a (to me) unknown source, but if true, another reason to avoid atmel for production use (unless reliably second sourced).

Dave

[bullet308]

Best of all worlds?

http://pinguino.cc/

Right now I am playing with various Atmel-based -duinos, but this Microchip-based -duinoid will probably be the next frontier for me.

[westfw]

Atmel had a a long run where they didn't turn a profit, or not much of a profit.
http://ycharts.com/companies/ATML/earning_yield#zoom=10&startDate=&endDate=&format=real
More recently, they've done pretty well.  They did have a boardroom/exec shakeup, and I imagine that a hostile takeover attempt by your arch-rival can cause a re-thinking of priorities.  The Microchip takeover bid was $5/share, and they're up to almost $10 now.  I even made about $1200 trading their stock this month!

While I wouldn't rate Atmel as the most trustworthy of companies, they've been around long enough and weathered enough difficulties that I would hesitate to learn the AVR architecture as a learning experience.  Ditto for the AVR architecture itself.  Also, the architecture is reasonably standard; most of what you'd learn would be useful somewhere else even if Atmel disappeared the day you finished studying.  It's not like it's an iAPX432 or something.

(hmm.  It's vaguely "interesting" that most of the famous "failed" computer chips I can think of came from pretty substantial vendors.  Intel 432, i960.  National (well, any of them.)  TI's TMS9900, TMS7000...  It almost makes you believe that the survivors have some inherent value!  (bad micros from big companies disappear.   Good micros from small companies end up owned by big companies!)

[vis1-0n]

Ugh.  Sounds like shareholders first, customers last.  Not that I'd base such decisions on a single random forum post from a (to me) unknown source, but if true, another reason to avoid atmel for production use (unless reliably second sourced).

Dave

I didn't mean to give you the impression, I am no reliable source - just commentating out of some discussions I remember on AVRFreaks. What westfw says above is true. I hardly think the AVR will go away, and even if it did stagnate under another company it would still be pretty useful and fit for purpose. Ala Java (Sun now Oracle), Yahoo!, Symbian (Nokia)... Even Delphi Pascal has developed over the years, sold from Borland to Embarcardo.

[dfnr2]

I hardly think the AVR will go away, and even if it did stagnate under another company it would still be pretty useful and fit for purpose. Ala Java (Sun now Oracle), Yahoo!, Symbian (Nokia)... Even Delphi Pascal has developed over the years, sold from Borland to Embarcardo.

I am sure of it.  However, I don't think I'd bet my small business on Atmel getting me chips when and how I need them over the next decade, unless they had something I absolutely required and just couldn't get from anyone else.  A cute development environment wouldn't change my mind, although I might use it for some one-off quickies if it's really convenient.  I still think their chips are pretty nifty, and I love the arduinos.  To venture a metaphor, Atmel is a fun date, but Microchip will still love you and take care of you when you're over the hill.