This is my response to a post on AOL. It is again based on the information I received from a good friend with ties to one of the institutional investors who attended the AXC presentation last week.
"......Based on that history one has to conclude that there is an industry reluctance to embrace this new technology. Why is that? Does anyone know? At this point, I am not sure that Ampex will ever benefit from this new technology because enough time has passed whereby more favorable news should have been forthcoming...."
Fair point. I question, however, your sense of how the disk drive industry works and to what extent that informs your sense of timing. There is a great degree of interdependence between the heads (about 40% of the cost), the media (next most expensive item), and the other components of disk drives that are assembled in a semi-automated assembly line. There are also ongoing developments in the architecture of the computer (Firewire, USB, Fiber channel) that the drive guys have to consider in their product development efforts. What you see as industry reluctance is really an ongoing and often protracted process of matching heads with media with components in line with where the really big guys (INTC, MSFT) are going.
At the end of the day, you really have to look at the technology and economic advantages KM offers AND use your business instincts to determine its relevance and to anticipate the value of this solution.
As you know, I question those who see AXC solely in terms of keepered media. These are the people who get spooked the most by every positive news regarding MR, seeing it as THE technology that will eventually render KM by the wayside. These are probably the weak hands (Salim is Exhibit 3) being peeled off everyday that this stock drifts down. But hey, it is not unusual for weak hands to sell to strong hands! Anyway, let's tackle this MR bogeyman below the knees, shall we?
MR involves seperating the read and write functions. The write function is performed by a specialized inductive head while the read function is performed by head with a precisely positioned MR element. There are obvious efficiencies gained by the simple separation (inductive recording uses the same head for read/write functions), but where the MR head really shines is in its ability to read, using the magnetoresistive effect measured by the MR element, essentially weaker signals. This allows MR to achieve higher density because it is able to read more densely packed magnetic bits far beyond the capabilities of the inductive head.
AXC has publicly stated that, in a test conducted with a drive maker and a read maker, it has been able to activate a keeper layer using a MR head of common design. What does that mean? It is useful to refer to the graphical interface of the defragment utility of your OS where the recording area of the disk drive is divided into small boxes that are either being R (read) or W (write) or blank, reflecting no activity. In inductive recording, activating a keeper layer means that the head introduces a bias charge to the "box" and basically allows the signal to pass through between the magnetic layer and head. While the relevant 'box' is being 'opened' up or saturated with the bias charge, the keeper layer SHIELDS the other boxes from leaking magnetic signals. This shielding effect increases track density. While I do not really understand it, there are verified tests that show that there is also a reduction in spacing loss that also contributes to the increased capacity. There are also yield improvements. Bottomline: 20-40% increase in platter capacity.
Because MR recording relies on the MR element in the read head to measure the magnetoresistive effect of the bits written by the inductive head, MR/KM requires a lot of additional work especially since a lot of the advanced MR heads and media are in the prototype stage.. Pennyslvania Merchant Group, with probable guidance from AXC, anticipates that it would take 12 months to commercialize KM/MR. But there is progress beyond AXC's public statements. No guarantees, of course, just encouraging signs.
These are the areas where Ampex is making some progress in making KM work with MR. Obviously, there are variations in the composition of the keeper layer and the thickness of the layers that are being tested.:
1) The activation of the keeper layer results in similar shielding effect and a similar increase in linear density. Shielding also increases track density.
2) A fair amount of work is being done to determine whether the use of keeper layer could possibly reduce the cost of the MR head design.
The current work, which again has been estimated to take up to 12 months, is also focused on even more leading edge MR designs like Spin Valve, Giant MR, and Colossal MR. The areas being investigated includes the increase in linear and track densities AND the issue of self-demagnetization. This is not a real problem in the lower density area of inductive recording and even the early generations of MR, but at the 5 Gb/in2 - 20 Gb/in2 levels, the even tightly packed magnetic bits tend to lose their magnetic qualities. As the MR heads become more powerful, these require more sophisticated architecture AND more densely packed magnetic bits. Packing even more bits together increases the occurence of this type of phenomenon. There is work being done with inefficient prototype heads and media that indicate that the keeper layer acts to stabilize the magnetic bits. This MIGHT be even more important than capacity which is not really an issue anymore with the powerful leading edge MR heads!
At any rate, below you will find a comparison of the road for MR and TFI heads for the next few product cycles. Note that TFI are still ahead of MR heads in terms of density:
"...There are at least two cycles foreseen for the inductive planar -- one at 1.1 Gb/square inch, which is equivalent to 1.6-1.7 GB/platter and another design for 1.5 Gb/square inch.,equivalent to 2+GB/platter.
The company hasn't seen any new big issues on the media side of MR.
MR is gaining momentum, but not at the expense of advanced inductive. Both the Fremont facility and the Japanese facility are producing substantial quantities of MR heads.
MR densities right now are at 700 Mb/square inch, or just over 1 GB/platter, to 1.2 Gb/square inch, or 1.7 GB/platter...."
Source: Motley Fool summary of Read Rite conference call 1/23
That's it for now. Given the milestones reached and the roadmap ahead, at the end of the day you have to figure out for yourself whether there is real value here and how long it will take for that value to reflect itself in the stock. Those who think of AXC SOLELY in terms of keepered media will undoubtedly suffer the most each day the stock continues to drift. Weak hands selling to strong hands that's the way I see it until the volume picks up or something else comes and changes that equation.
Oh, and this is third hand and in the highly subjective and FWIW category, one investor who has dealt with Bramson over the past 2 years made the comment after going to one of the AXC presentations last week that he seemed more confident than ever. As you know, Bramson was born with a disclaimer stamped on his forehead. Useful fact to remember if ever we need to do an autopsy on this one. Unlikely, tho. |
