OK, my turn on Rambus.....(long post, warning)
Apologies to the thread for being out of touch the past few days. Friday was a brutal day for Intel/Rambus investors.
The post at the bottom from the INTEL thread sums it up. But first, my take as a gorilla gamer....
1. I am disappointed in comments by Freeus and Pala; their comments indicate they do not understand Rambus, its product or the industry. It's OK not to know things, but like Uncle pointed out, we have waxed long and clear on Rambus for a long time now. For the 1-millionth time, I'll say again that Rambus is not a Gorilla, but a "gorilla candidate".
2. Yesterday, it became evident that the INTEL chipset, Camino, which allows the CPU and memory chip to communicate will be delayed.....again. This comes before the heretofore announced Camino launch Sept. 27th. This lays entirely at the feet of Intel, which has slipped up 3 times in less than a year on this issue. First, it announced what was perceived to be a delay for Camino earlier this year; then, it reversed course and said it would support PC 133 SDRAM with a chipset (this was no big deal to Rambus longs since PC 133 would fill the low-end and mid-range markets, which were not a target for DRDRAM until late 2000, and 2001); now, just a few weeks after confirming all was on target at the Intel Developer's Forum, Intel admits some sort of Camino chipset or motherboard glitch. Freeus and Pala, there is nothing wrong with Rambus, but there sure does seem to be something wrong at Intel. As an Intel investor, it pains me to say this.
3. I agree with comments before on Rambus stock volatility. Whew!!! it leaves me breathless. What isn't volatile is the company and its management. From what I can see, Rambus as a company is solid, profitable, and very consistent in what it claims its memory architecture design can do. Rambus management is steady, very conservative, and usually understates rather than hypes. Rambus has not done anything wrong here; however, what has happened is a series of glitches in the value chain partnerships. Memory makers have been slow to come on board, and Intel has screwed up the Camino chipset launch in spectacular fashion. Because of the delays this will encumber, and because of past delays, I expect a major confidence hit and a ripple effect throughout the whole value chain. I think the stock price will suffer again, but its share appreciation that ultimately comes will be slower, more guarded, and at a lower slope than would have occurred had Camino been launched and Intel followed thru on its promises.
4. Rambus memory works, is scalable and provides more bandwidth than any other present or perceived memory option. Period. Bottlenecks to the CPU occur at the level of the memory, and Intel's future partly rests on improving this situation. More globally, anything that increases data transmission and bandwidth is generally in keeping with the themes of the next decade related to the wild and relentless growth of the Internet. The Camino chipset or MOBOs will be redesigned, because Intel needs Rambus, and because there are no better long-term alternatives to DRDRAM, and also I suspect because Intel will lose great face (and profits) if they have to abandon this path and start over. Therefore, I think they will make it work, come hell or high water. Also, other elements in the value chain have already invested hundreds of millions or billions, and they therefore have a vested interest to see it work.
5. In video games, Sony has openly declared DRDRAM to be its choice for the Playstation II; Nintendo has all but declared it for Dolphin; tech specs for Dolphin resemble DRDRAM, and DRDRAM is in present generation Nintendos.
6. So where to from here? I think Rambus dram will become a dominant dram, but its ramp up will be slower than anticipated. Intel's web site shows the ramp up they expect for Rambus DRAM; however, I notice thru '99, the web page has changed in that the ramp was recently projected to be slower. I think that after today, it will be slower still. Unc's and Merlin's comments about the size of the value chain and the quality of the players also supports the notion that Rambus will happen. I remain a Rambus long, but am severely disappointed at Intel. With Camino's launch next week, and availability of workstations and corporate PCs from Dell, Compaq and HP in October, Rambus would have CROSSED THE CHASM. But now, the chasm was made wider; crossing will be delayed by months.
7. When Intel announced a couple months their interest in PC 133, I indicated then that I lightened up on Rambus to buy Qcom. I sold more of my Rambus position on Friday, and reinforced positions in Qcom & Gmst. I haven't lost money on Rambus, but the profits got erased on Friday. I indicated before that my equities portfolio was mostly gorillas and kings, but that with Rambus as a gorilla candidate (~ 15% position), I was hoping to hit a home run. My present position in RMBS is now much smaller, and I think it will take a longer time to see the fruits of all of the research in '99 on this company. This latter has been frustrating for me; I have invested mucho time into really understanding Rambus as a company, and as a gorilla candidate. As a physician and intensive care unit director, I suddenly know alot more about computers, DRAM and the PC industry than I would ever have imagined, as well about gorilla game principles. So it is disappointing that tho' the company is very steady, the stock is so fickle and wild, and may not return what I had hoped given all the invested time. Believe me when I tell you that restoring a patient's airway when asphyxiating or achieving return of spontaneous circulation from cardiac arrest is child's play by comparison.
8. So what about the gorilla game. what are the lessons? How could I have seen this coming? Well, the answer is, I don't know. As Tony Viola posted here in the summer, Intel management is "the greatest on earth" and normally engenders great confidence; its word tends to be golden; so its missteps this year on this issue is a confidence shaker. Others on this thread did warn about the risks of investing in pre-gorillas, and these difficulties with Rambus's value chain and partners (NOT with Rambus itself) do point to this. I said before that if nothing else, investing in Rambus would be a great case-study on market forces and gorilla-stalking, and the evolution of the technology adoption life cycle. And so it is; for me, the lesson from friday is that even if you discover a company with a great product, good management and which has real earnings, things can still go astray if the partners or value chain stumble.
Have I learned anything else? Well, every time I lighten up on Rambus, I reinvest in Q the gorilla. So now my portfolio is approaching ~ 60% Q. This board has been enthusiastic about Gemstar, which has crossed or soon will cross the chasm. It'll be interesting to see, based on recent experience with Rambus, what if any surprises await us early investors in gemstar between now and when the proprosed tornado takes place in the next 12-24 months.
I invite all threadsters to comment on this take on Rambus, especially in terms of dissecting out the gorilla game principles and lessons here. I especially would like to hear from Tony Viola on his take on Intel's missteps here; how does this compare historically, etc.
Apollo
To: Ali Chen who wrote (88856)
From: Burt Masnick Friday, Sep 24 1999 4:25PM ET
Respond to Post # 88858 of 88901
This has been the week from hell as far as Intel longs are concerned
- Taiwon earthquake leading to possible MOBO, RAM, Chip set shortage
- Camino who-knows what problem with Rambus
- Xeon 8 way problem
The earthquake is not under Intel's control. The other two seem to fall at the feet of Intel. I'm fairly startled at the AMD-like inability to get out of their own way. I believe that they will address the problems straight up early next week with public statements about all three situations and where they stand on corrective plans. They are looking more than a little out of control.
The possibilities are 1) random events in close proximity 2) systemic problems that are development process related. My WAG is that, while Intel is superb in manufacturing standards, discipline and procedure, the development process is more loosely controlled. Both device/board problems should have been surfaced in development test phases.
I suspect that a lot of Excedrin will be consumed this weekend in Santa Clara. Can't come up with a way to feel good about these events. |
