AMD offers to settle Llano litigation for $29.5 million

Back in 2011, AMD launched the Llano series of APUs. These chips combined CPU cores derived from the Phenom II line with an IGP based on the Radeon HD 5000 series GPUs. At the time, AMD was excited about the first steps towards its Fusion concept and talked up Llano perhaps a little too much, as the chips ended up a year and a half late and had underwhelming performance. In 2014, a group of disgruntled AMD shareholders filed a class-action suit against the company claiming that AMD overpromised and underdelivered with its first-generation APU hardware. Now, AMD has made an offer to settle that suit to the tune of $29.5 million.

The suit alleges that AMD made "false and misleading" statements about its ability to manufacture and supply Llano APUs to customers, as well as about the chips' competitiveness in the market. Despite AMD's own admission in 2011 that it was having supply issues due to poor chip yields, the suit holds that the company claimed it would have "ample product available at launch" for its first-generation APUs. The suit also says that AMD reported it was "well positioned" to get Llano into PCs for the back-to-school season at the time.

The settlement is only an offer at this point, as Judge Yvonne Gonzalez Rogers still has to give her approval. If the offer is approved, AMD can finally be rid of this whole mess. As a bonus, the company doesn't have to admit any wrongdoing, either. Supposedly we'll hear more information from both sides of the case on October 9.

Llano was originally set to be the very first AMD APU. The chip series was supposed to release in 2009 and go head-to-head with Intel's first-generation Core hardware, against which it might have been a credible competitor. Instead, it released in August 2011, eight months ater the Bobcat-based Zacate series. Against the very impressive second-generation Core hardware—perhaps better known as Sandy Bridge—it was crushed in CPU performance benchmarks. Llano chips like the A8-3850 offered impressive integrated graphics performance for the time, but that horsepower was easily outpaced by even cheap discrete graphics cards. Furthermore, by the time Llano was ready for market, the next-generation Trinity chips weren't far behind, and ultimately AMD had to write down $100m in unsold Llano chips.

Comments closed
    • ronch
    • 2 years ago

    I wonder just how much all those Llano chips ‘made’ for AMD. How much money did it actually take in, minus the $100M write off and this settlement including litigation fees and who knows what else.

    • HisDivineOrder
    • 2 years ago

    If this were about nVidia or Intel, this place would be LIT UP. Doesn’t matter how old or how far gone the technology involved is now. The flames would be raging completely out of control.

    Thank God it’s only AMD, right? 😉

      • chuckula
      • 2 years ago

      If you want I can go dredge up the usual 15 year old conspiracy theory about how all of AMD ills are due to Intel giving a discount to Dell because AMD would rule the world right now if they had been the ones giving Dell steep discounts to put AMD chips into those nasty 1999 – 2002 era plastic boxes.

        • Spunjji
        • 2 years ago

        It’s not exactly a theory, though – that Intel were keeping Dell away from AMD was pretty widely known info at the time and has long since been confirmed.

        The only theory part is about what AMD had done assuming they:
        a) Were able to get decent terms with major manufacturers for their own chips, and
        b) Had somehow magicked up the manufacturing capacity to supply them in the required Intel-matching volumes

          • K-L-Waster
          • 2 years ago

          [quote<]Intel were keeping Dell away from AMD was pretty widely known info at the time and has long since been confirmed.[/quote<] That part isn't being questioned. What's being questioned is the idea that that occurrence was solely and exclusively responsible for all of AMD's woes ever since. AMD's own missteps couldn't have had anything to do with it: it was aaaaalllllll due to the Intel / Dell deal.

          • Kretschmer
          • 2 years ago

          AMD has always been fab constrained, though. I’m not sure if their marketshare would have gone anywhere with a more level playing field. Intel’s marketing reign’s supreme, and most Dell buyers don’t do so based on benchmarks and detailed technical assessments.

          • badpool
          • 2 years ago

          Come on man, please learn what “theory” means and use it properly.

    • Wildchild
    • 2 years ago

    Misleading share holders is a pretty big deal and this was a stupid thing to let happen regardless if it was or wasn’t completely AMD’s fault.

    It kinda sheds some light on how desperate they were at the time though. Oh, how times have changed. Let’s just be glad that AMD has, for the most part, licked their wounds and finally started moving on.

      • flip-mode
      • 2 years ago

      That graphics card wound is still festering. A card that cannot beat the competitor’s even when launched a year later and fairly worse power consumption? Call the doctor.

        • Lore
        • 2 years ago

        I guess that would be a big problem if AMD were able to produce more Vegas than they can sell at jacked up prices.

          • Beahmont
          • 2 years ago

          True, but they still aren’t getting all that much money out of them because they just don’t have enough of them in the first place. Yields from GF are on the low end of acceptable for such a massive die, but they are losing more dies to bad assembly with the HBM2 than from yields. Add in that the just don’t have enough HBM2 and it’s just bad times all around.

          Apparently it’s stupidly hard to get a massive chip like Vega and 4 HBM2 stacks to join properly on an interposer without one of the pieces somewhere breaking. It’s hear-say, but last I heard that because of the nature of how you have to join parts to the interposer, it’s very hard to start over if something goes wrong with one of the parts.

            • psuedonymous
            • 2 years ago

            [quote<]but they are losing more dies to bad assembly with the HBM2 than from yields.[/quote<]That's the case with everyone's interposer-based designs (hence the GP100 PCIe version with one of the HBM dies disabled). Though with Vega using a pair of high-capacity dies rather than 4 lower-capacity dies, they don't really have that option as a fallback binning. The tradeoff for being able to use a smaller interposer.

        • shank15217
        • 2 years ago

        I would agree, except the deficit wasn’t nearly as bad as lets say an Opteron vs Xeon back in 2016. AMD graphics have always performed well within their price brackets and have provided compelling features.

    • NTMBK
    • 2 years ago

    So AMD is stuck in a contract to only use chips from GloFo, are required to buy a certain number of them no matter what, and then when GloFo can’t deliver on time AMD ends up having to write down thousands of those chips as a loss… and now they’re paying even more on top of that to apologise to shareholders for those delays.

    Thanks, Ruiz!

      • chuckula
      • 2 years ago

      Only now can you see the true power of Ruiz’s genius in moving AMD to the vastly superior [s<]profit[/s<] [u<]asset[/u<]-light business model!

        • ronch
        • 2 years ago

        But Profit-Light does seem accurate now, doesn’t it?

        • NTMBK
        • 2 years ago

        It’s not the basic business model that’s the problem- it’s clearly worked out pretty well for NVidia and Qualcomm. The problem is the insane terms that Ruiz struck, chaining AMD to GlobalFloundering for 15 years. What the hell is the point of spinning off your fabs if you’re stuck using them anyway?

        • Peldor
        • 2 years ago

        I always felt like Ruiz sailed off on the saucer section leaving the rest of the suckers on the battle bridge to fight off the Borg (Intel).

      • cynan
      • 2 years ago

      Well, it’s obviously because they weren’t forthcoming about the delays, not that there were delays. And they might have had a shot at beating this in court, but obviously did not want to take the risk and waste more resources.

      • strangerguy
      • 2 years ago

      Yeah, it sucks so hard that the underdog x86 company is cursed with such an terribly incompetent egomanical CEO who collectively set the whole industry back for a decade. If you think NV is strong now, imagined if they managed to secure the x86 license 10 years ago…

      • ronch
      • 2 years ago

      Hmm… Wasn’t Dirk the CEO back in 2008 – 2011? But yes I agree Hector really brought AMD down. AMD was rockin’ when Sanders handed him the reins. Both Barcelona and Bulldozer specs were frozen during his watch and look how they both made AMD fell behind. Phenom II may have fared OK in the market but it’s just a lazy tweak of a previous generation and Bulldozer, well.. And then there was how they paid too much for ATI and sold their fabs with those frickin’ terms attached that haunt AMD to this very day.

        • Pettytheft
        • 2 years ago

        And Dirk was the guy who sold the handset tech to Qualcomm for a whopping $65 million. Another huge blunder.

      • Kretschmer
      • 2 years ago

      AMD didn’t have the revenue to maintain in-house fabs and couldn’t sell their fabs without some sort of guarantee. IBM paid GlobalFoundries $1.5 billion to take over its fabs, and it’s probable that AMD chose onerous terms instead of paying out cash they didn’t have.

      • Jtaylor1986
      • 2 years ago

      What value does a fab have that has 0 revenue? There was no viable spin off plan that did not involve AMD buying from them for a long time until they could create a pure play foundry business from scratch. AMD probably could have gotten a better deal out of it in the end but getting people to pay you for taking losses off your hands is a pretty tricky thing to do.

    • chuckula
    • 2 years ago

    You can see the original complaint here: [url<]http://securities.stanford.edu/filings-documents/1051/AMD00_01/2014115_f01c_14CV00226.pdf[/url<] In particular, good ol' Rory and ex-CFO Tom Seiffert made a whole bunch of very positive sounding statements about both the demand and the supply of Llano that we later found out to be less than accurate. Especially when they made lots of statements about how they had put Llano supply issues behind them when this was not the case.

      • AMDisDEC
      • 2 years ago

      True, you can’t pin this malfunction on Ruiz, which is like pinning the US deficit on Clinton when it in fact lied with the incompetence of Dixiecrat Bush.

      This is due to the incompetence of country club hires, Meyers and Rory and the terrible helmsman-ship both displayed as CEOs.
      Besides the APU marketing screwup, under these goons the company wasted time and resources on the doomed to fail ARM Server strategy, which was a really insane decision.
      Not that Ruiz didn’t commit any sins when he got caught up in that industry wide inside trading scam, but that only affected him professionally, and not AMD’s bottom line.

      The bright side is, people can learn and not repeat bad habits as shown with the hire of ultra competent D. Lisa Su.

        • NoOne ButMe
        • 2 years ago

        While failing at nearly everything else, Rory Read did cut costs a ton, which I recall as being the reason why he was brought on.

        The list of short screw ups he made in his short tenure is incredible, but the core responsibility was done.

          • AMDisDEC
          • 2 years ago

          Those cost cutting measures closely resemble a well used and tired practice among CEOs of publicly traded companies.
          For example, the outsourcing to India trend was always prompted less to make processes more efficient, and more to do with blowing smoke and mirrors to appease Wall street analysts and investors. The justification was always, to pass cost savings down to customers. You know that is impossible with Wall street. Any savings would be given to shareholders.
          Companies missed projections and responded with plans to cut costs by outsourcing which narrow minded Wall street applauded, Two quarters later those cost cuts were dwarfed by budgets needed to correct poorly designed code and hardware developed by the outsourced companies.
          I’ve actually observed companies lay off all their senior personnel, outsource the tasks they were doing, only to be forced to hire them back at twice the pay to complete and correct the work done by the companies outsourced to.
          Wall street is an intrusive and destructive non-sustainable model that is detrimental to any innovation within tech companies and CEOs who play this losing game are just incompetent.

            • Spunjji
            • 2 years ago

            I’d like to actually back this up from experience, albeit in the UK. I am very good friends with a database contractor whose job with British Telecom was shipped out to a team of 5 in India. He was paid to train them, then let go.

            They hired him back 18 months later because the code was late and didn’t work properly. He spent 6 months fixing it at twice his previous salary.

            So… yeah. Cost-cutting exercises are rarely straightforward and often cause more wounds.

Pin It on Pinterest

Share This