The Geforce chip is made of copper instead of aluminium, which means it can run faster - Spencer Kelly, BBC Click Online
ALMOST FOUR YEARS AGO, the first antitrust shots were fired between AMD and Intel. At that time, I said to wait and see how this evolves, knowing that the final outcome could be a long time in coming.
Today marks what I would consider to be nearing the end of that long time. The full 542 page phone book of a ruling is not published yet, but there is a large and growing body of evidence in the public view. A lot of the questions that were open in 2005 are now answered.
Stepping back a bit in time, shortly after the first antitrust actions were announced, I did some digging on the whole rebate, bundling, blocking controversy. I had been told about the practices again and again for years, so I was pretty sure they were real. Calls and emails dug up what I would consider ample evidence that the practices were quite real and widespread. That said, it was all hearsay, but there sure was a lot of hearsay if you so much as scratched the surface.
Step one was easy enough, finding cases of alleged wrongdoings. Step two was significantly trickier, so tricky in fact that unless you were an antitrust lawyer up to date on geopolitical happenings and leanings, you couldn't address it. The underlying question is simple enough, though. Assuming that Intel did do the things alleged, are they actionable? The answer is far from simple.
It brings up questions like, is Intel a monopoly? Is a monopoly in x86 chips enough, or do you need a monopoly in CPUs, semiconductors or something else? Are the actions, unpalatable and unethical as they seem, against the law? How about if it is done across international borders? Those coupled with the usual caveat of hearing only one side of the story, or in this case, one-tenth of one side of the story, make things a mess.
So we sat and listened. Patiently, paying attention to the details, and watching things unfold. They did, with all the speed of a governmental hearing run by octogenarians on sleeping pills. Three of them so far, with three more still in process.
One thing most people don't recall is that Intel to date has lost in Japan, lost in Korea, and now lost in the EU. It's batting zero for three. AMD has not lost a single ruling yet. From a legal perspective, this rather odd order was done for a very, very good reason. The lawyers at AMD are earning their money.
Why the proceedings were approached in this way is a long story. The short version, much of which is speculation on my part, is that most of the alleged wrongdoing by Intel was done with an intention to not document the acts, or at least to not document the illegal parts. In some cases, you get two people, lets just use a hypothetical meeting between Mike and Craig.
How can you prove that, if it happened at all? You can't, it is your word against Craig and Mike's, and they won't likely stand up under oath and admit, "Yes, we colluded to violate antitrust laws and defraud the consumer." Well, almost everyone wouldn't admit to something like that, and both of these guys are likely to have not only a willingness to break the antitrust laws but also the desire not to get caught at it, along with above average intelligence to get to their respective positions.
So, you look for secondary sources, things like emails and memos. If Mike, again hypothetically, sends an email an hour later to purchasing saying, "I just negotiated a 10 per cent discount on everything Intel, so drop that AMD order," that would be a good place to start. The problem is that if you are smart enough to cut deals like this, you are also usually smart enough to not blab about it. Instead you have lunch with the purchasing head the next day and chat about the weather.
If you want to find incriminating evidence, you have to dig and dig through mountains of emails and documents from multiple Fortune 500 companies. To narrow your search, you start small. You not only start small, but you start small in cultures that are much more regimented in their business dealings and require more documentation.
That is why things went from Japan to Korea, then to the EU. Japan documents dealings much more thoroughly, and Korea does so almost as well. The EU is more tenuous, and the US is so hit or miss that anything goes. The trials mirrored this path of most to least evidence, using what they found in previous cases as stepping stones.
Basically, Intel got slapped down in Japan and then also in Korea. Intel disputed this, and says that it will appeal in both countries. Intel admits nothing and will loudly proclaim that it has been wronged. Three times. Under three different justice systems. With three different sets of evidence. All of which are wrong, unjust, and will undoubtedly be overturned eventually, but haven't yet. Just you wait though....
The EU ruling was likely augmented with some evidence found in the other two actions that established patterns of behavior. In its conference call yesterday morning, Intel was quick to point out that there was a lack of evidence about some things. I believe that is exactly why the three cases were brought in the order that they were.
If you can document practices in nine circumstances that resulted in a certain behavior, and the tenth time you see that same behavior but don't have paperwork, and the behavior is not a common industry practice, it doesn't take much to draw a conclusion. While I am not sure that there isn't plenty of actual documentation here, the more you have, the better off you are. Anecdotally speaking, AMD sure seems to have compiled a lot of evidence.
The hardest jurisdiction of all in which to pursue an antitrust case is the US. Business practices here are much more lax, our government is blatantly purchasable - ask Microsoft about its antitrust lawsuit - and most regulators have had better things to do lately, like keeping the financial system from totally collapsing.
Well, that was the case until January, when things may have changed. Earlier this week, there were serious rumblings emanating from Washington about tougher enforcement of antitrust laws, but as usual, it may very well be more hot air and political posturing. In any case, the wind is shifting away from Intel toward AMD as the next three cases start heating up.
Those cases are a US antitrust action, a similar case in New York, and a civil lawsuit brought by AMD against Intel. Things are going to get ugly for Intel when you can document behaviors in several jurisdictions, but not in others. When you subpoena certain things and you get the answer, "I'm sorry your honor, the dog ate my email," things get bad.
Spinning the answer, while expected, doesn't usually sway the courts. One other bit to think about is that email has two sides, and if the emails went out of Intel, those might be discoverable in a subpoena to a third party or company. While I have no idea if such things exist in these cases, given the sheer volume of documents I would be shocked if there were not such things, on both sides.
If Intel can't find them, and some turn up via some OEMs, Intel will have a lot of hot-shoe tap dancing to do in front of a guy in black robes. Given the number of people who were jumping up and down waving their hands at AMD lawyers begging to be subpoenaed, I would expect at least a few fireworks in the US. Yes, IBM Opteron server engineering teams, I am looking at you.
Given what has come out, I was more than a little surprised at the Intel press conference yesterday. Most people focus on Intel's almost farcical denials of any wrongdoing. As we noted above, how many courts need to slap it down before it admits to anything? The answer is all of them, because if Intel admits anything, no matter how obvious it is, it hamstrings its appeals, so that part is understandable.
The parts that got me were Otellini's claim that AMD was "more vibrant than ever," that Intel had evidence that was excluded after nine years of the process, and most ludicrously that there were no customers that complained.
AMD vibrance isn't claimed by anyone who doesn't talk to their shoes and act as if the shoes talk back vigorously. To say that at a press conference is insulting to the intelligence of the audience. I honestly expected better from Intel, it was a sad sight.
The evidence exclusion claim is curious as well. The EU action was go around number three, with four and five in process as well. What mystery evidence popped up this time? Why was it not relevant to the Japanese and Korean actions? How does it destroy the preponderance of evidence that the EU Commission presumably found? As we said in 2005, let's wait and see, but I will be officially skeptical until proven wrong here.
Lastly, we have the bit about no one complaining. The EU lists a bunch of third party OEM complainants, Acer, Dell, HP and Lenovo, all of whom spoke up. The Media Saturn group also chimed in, hardly a corner store. These were the companies that were willing to stand up. There are probably others that didn't for one reason or other, but I think OEMs representing slightly over half of worldwide PC sales make up a fairly representative and persuasive sample, don't you?
We understand that neither Intel nor AMD got a heads up from the EU about what the decision would be, or at least they didn't get enough of one to organize press briefings beforehand as is the norm for events like this. We also don't doubt that neither side had seen, much less read or analysed, the ruling before the press conferences were held.
Once again, Intel looked really silly when it claimed that it didn't know of any customer complaints. It is the third paragraph of the press release, here. I will cut Otellini some slack for having had a horrible night. He was likely woken up at two in the morning and force fed bad news, then told to smile for the cameras. I will cut Intel some more slack for a hastily scripted press conference filled with both annoying piranhas and slow English majors, that takes bravery to do. I won't cut Intel any slack for failing to read the press release before standing up and taking questions from the press.
After four years of watching the cases unfold, all the open questions have been answered to my satisfaction. I can now say that, until proven otherwise, I think what Intel did was wrong, unethical, and illegal. It is zero for three now with three more to go, with a seemingly steep uphill climb from here.
Could Intel be right? Sure. Could three different justice systems have been duped by a conniving AMD? Sure. Could Intel power to victory in the next three antitrust cases? Sure. These, and thousands of other scenarios are possible, but at this point it is up to Intel to prove its claims. Until then, AMD's claims seem to be the correct ones. This should all be settled some time before the year 2059. µ
Sign up for INQbot – a weekly roundup of the best from the INQ