One lawyer I talked to about it said RegFD is basically ignored by the SEC now, a sad state of affairs. [...] Of all the companies that I talk to, only NV consistently violates this practice, others may slip, but NV seems to do it as practice, playing favorites with analysts. If the IR department ever gets subpoenaed, they are in deep deep deep shit.
Heh, yeah, doesn't surprise me. SEC is a joke right now anyway, so I wouldn't hold my breath. FWIW, I wasn't just talking about stock analysts - I also know some definitely non-public things which individuals investors managed to get from NV IR. Pretty retarded behavior...
NV _IS_ out of the MCP business, mainly because chipsets are not going to exist in a year or so. Stop and think about it, what is a chipset? It is, classically, a memory controller, GPU, SATA, USB, PCIe, low speed IO, and a boot rom.
Of that list, memory controller and GPU are the difficult ones, the rest are trivially licensable IP. MC and GPU are gone in a year, and will never come back, PCIe, probably the next hardest one, is going on die shortly. So, do you think NV can make margins necessary to build a SATA, USB and boot ROM chip? Even if they call it XRP or something stupid? Chipsets are dead dead dead. The end.
Uhm, do I really need to remind you of what you said yourself? Here's the link:
http://www.theinquirer.net/inquirer/news/1021993/nvidia-chipsets-dead along with two choice quotes: "the division was officially killed, and the teams will be rolled into GPU projects" and "So in the end, Nvidia chipsets are history. You will see a few more trickle out mainly because they are already done and dusted."
That doesn't contradict MCP79 being used for the Ion platform. But it completely contradicts the existence of the 40nm Ion2 aimed at ~4Q09 for Socket 775. Remember transition times are always long, and NV should easily be able to get a lot of business with that part in 2010 and through 2011. And that's before we even consider the fact it supports VIA Nano, which has a very nice roadmap. Don't get me wrong, everyone makes mistakes and there's nothing abnormal there, but it doesn't mean they have to be denied later though.
After that, I agree it gets cloudier because of the obvious lock-in mechanisms. There are some things they can do on the MSI socket that involve a GPU, believe it or not, but they'd definitely be at a disadvantage and I'm skeptical it'd/it'll be worth the R&D. Which is why NV said they'd want to integrate the CPU eventually. My favorite strategy for that is still an agreement with VIA for 28nm where they cooperate to put the two on the same die and legally speaking it's manufactured by VIA with them giving out large 'IP' royalties (proportional to die size, not like real IP in the handheld world) to NV. This would completely bypass the x86 license issue, although it might be a problem that VIA's license expires in a few years and will have to be renegotiated AFAIK...
Would you care to define "notebook marketshare based on design wins"? Does that mean notebook designs or chip sales, the two are not mutually inclusive. Dell is worth more than Eurocomm, but both are a design win.
That said, given that notebook design cycles are much longer than desktops, I think notebooks will lag desktops by about a year in any changes. Also, desktops change every quarter, notebooks once a year. I don't expect the changes to be public until the 2010 round of notebooks.
FWIW, Hara explicitly said it would apply to both the number of design wins and chip sales. But okay, that makes sense - so you'd basically predict NV to lose significant desktop share in the back-to-school and winter cycles?
Hara is trying to skillfully bend the science, basically if you stay below Tg, you are OK, but no GPU in use will, fan or no fan. If you go above the Tg (80 C or so) it doesn't get noticeably weaker, so it isn't WORSE. He is trying to explain that while pretending they don't understand, and science hasn't caught up with 'melting' yet. (Yes, technically it isn't melting, it is glassification, but still....)
Ahhh, okay. So that's the side of the story I was looking for - that'd certainly be very very disingenuous, but it definitely is relatively plausible. I'd love to see NV's reply to specifically to that, although for a number of reasons it's pretty obvious they'll never comment...
If that is $150/unit, how much do you think it would cost to do a recall, and replace those notebooks. Think $1000/unit average is fair? Plus $200 for costs/handling/phone support/logistics? (this isn't a high number from what I am told, phone tech support calls are $35 or so in cost each). That would mean 8 * 100M there.
While your initial reasoning makes good sense, you suddenly jump to calculating the cost of a recall (which I do realize would be enormous) while my question was whether it would be legally necessary to do one...
Plus, do you really think HP and Dell want to recall such a large number of notebooks anyway? Is it really in their best interests, and at such a late stage?
Then there are the suits. Keep in mind, there are product liability suits with damage multipliers, shareholder/SEC violation suits, there will inevitably be more shareholder suits when discovery happens (Trust me, I know where the bodies are buried here, I just don't have the docs in my hands). This will add up, not to mention money and executive time sucked up by stupidities and depositions.
The lawsuits are the big question mark here, yes. If they go really badly, then I agree NV is potentially in very deep trouble. However, these things do tend to be hard to predict no matter the facts, so I won't try to speculate on that myself.
NV is covering up tons of stuff. They are mismanaged. They are arrogant pricks. They are on the bottom. The line stretches to the horizon, and they won't get up from this.
Well at least you're making specific/verifiable claims now!
We'll see how it goes, although (excluding the lawsuits) I suspect that even if OEMs were pissed off, the most important factor remains the quality of the product line-up and that's an entirely separate question.
