HOST INTRO

Welcome to the LTO show, the premier podcast for leaders in the tape and storage industries. I’m your host, Pete Paisley. In each episode, I’ll bring you conversations with industry leaders who are contributing in special ways to the tape storage community. Our goal is to deliver fresh insights into the business case for LTO tape storage. And today I’m excited to welcome back Nathan Thompson, founder and CEO of Spectralogic.

In the late 1990s, Nathan Thompson faced a crossroads. His company was building both tape automation hardware and backup software, competing with everyone in the market. The easy move? Sell the hardware. Keep the high margin software business that employed much of his team. Instead, Nathan did the opposite. He sold the software division to Sterling Software and bet everything on tape libraries. That decision, choosing focus over diversification,

transform Spectrologic into the market leader it is today, building the world’s largest tape libraries and maintaining customer relationships that span decades.

INTERVIEW WITH NATHAN THOMPSON

PETE:

So Nathan, during this time at Western, before the Spectrologic acquisition, how did you balance the various requirements of the business? Administrative, technical, operations, sales? Were you a madman wearing every hat in the business, doing it all? Or did you feel even then that you had particular strengths and weaknesses in certain areas that you liked to focus on or avoid? I would have to say that

NATHAN:

Not having a lot of formal management training. I probably did a few more things than I should yet. I was able to hire people that I could put into roles. I had a very early finance accounting person. had a, I, after designing that memory board, I mentioned I, I had to spend a lot of time making it work.

correctly. And I came to realize that although I might have the skills to do it, I might not have the endurance to make something like that work. And so hired engineers. Several of them are still with the company. Probably went through a few cycles on the sales force, but hired sales managers and sales persons that could lead sales and had manufacturing people that could build these components.

Again, we were a pretty small company, but a goodly number of them were able to progress quite a ways. And there’s still a number of them sort of in the Boulder community that maybe laugh a little bit, but look back fondly at those times.

PETE:

Nathan, tell us about how the opportunity to acquire Spectralogic came up.

NATHAN:

Our company at the time, Western Automation, was building add-on memory boards and we started working on some concepts of controller boards. At the time, there was a company, which probably people have been in the computer industry.

a long time, remember called Cypher data products. And they built a, instead of a tape drive that was vertical where you put a tape on one side and have it thread to the other, they came up with a pretty neat concept where you would take a tape and a tape at the time was one of these open reel half inch tapes. And they’re probably doing nine and 12 inches in diameter. And so they would look, you would open the door and you would load it in

and close the door, would thread and then it could read, the tape. And if I recall correctly, it had an interface, which was not an analog interface. was kind of a per tech interface. These were used a lot by many computers, often for backup, but also for interchange. And the one mini computer that we had some pretty good depth in was

the Texas Instruments 990, but there’s probably others that people will recall that were around at the time. This company made interface boards for Texas Instruments, but they also built for the digital equipment, VAX and PDP 11s and Data General, Nova and Eclipse. the ⁓ Perkin Elmer Cipher had bought this company because it was integral for them attaching tape to ⁓

uh, these to these mini computers and the, the mini computer companies, in many cases OEM, their interface board, but there was also a pretty sizable aftermarket. And so we had become aware that this owner of this company that we had known and respected Spectralogic wanted to sell the asset. And Spectralogic was in Mountain View and there were, oh, probably at the time, maybe 25 people.

with it. And so we approached them and there were a couple others that approached them. And you never know, it wasn’t probably as competitive as it might have seemed at the time, but we might’ve been the only viable buyer. But we had to go get a bank loan. And this was at the time, I think it was United Bank. And we’ve had this banking relationship since.

you know, two months after I graduated from college and I had to go get a loan. remember I had to go pitch them on this loan. Right. You know, they ran me through the coffee grinder on that and it was good training, but that the, we had to come up with projections and, and, ⁓ show them that it was worthwhile. Of course, the concern is, are they going to get paid back? And I remember a guy saying, well, what if your projections are completely wrong? And you know, I,

Probably knew it before, but I responded, well, if we never saw another board, we’re going to be okay because there’s so many in the field. We’ll keep repairing them. so we acquired the business. sold some more, not a huge number, probably met our forecasts, but that repair business was extremely strong. so, Nathan, were they under maintenance or just a repair business that was healthy? was a repair bit business too.

the maintainers. So I had mentioned my wife had run this company in Florida that did TI contract maintenance on Texas Instruments computers, how I met her, but they maintained several hundred, I’m kind of guessing, trying to recall of these many computers around Florida and some of the areas and on Air Force bases around the world. So one of those boards would fail. People use

talk about it’s size of a pizza box. They were all like the size of a pizza box, you know, as early bipolar logic and a chip would burn out or something. And so we would repair the boards. Yeah. And so, ⁓ we did that for, ⁓ you know, I probably had a tail of about 10 years, ⁓ but probably the most significant thing was the, we got a brand, there were a series of resellers around the world who sold these.

that we could have a relationship. then we started moving our products through ⁓ that brand. And so we picked up the name Spectralogic. We reincorporated in Delaware. mean, was not a, it’s not like we went out of business or anything, but we, took Western Automation, ⁓ which owns Spectralogic and we had the rights to the name.

reincorporated as Spectralogic. And I think it was about 89 or about. you know, when I was trying to name this company, when I was in college and running it, I went around to my friends and said, oh, I’ll buy you a six pack of beer, which was probably illegal because I was underage at that point. But if you gave me a good names and so I got a number of names. I remember logical system design, which

Ha ha ha means LSD or mad dog systems or some of these. but I’d come up with a name, but I wanted to, I wanted a better name and a better known brand, but it was the same company. And, and how long after that did you decide, well, let’s get into building tape automation ourselves. Sure. So we had, ⁓ this concept and around that time,

⁓ Most tape was this ⁓ was called 800 BPI 1600 BPI was nine track tape. We had the concept of ⁓ what we saw a local company called Exabyte build a 2.2, I guess 2.2 gigabyte ⁓ drive. It’s the Exabyte tape drive and it used a Sony helical scan deck from a camcorder.

And, that deck slash drive set the world on fire for going to tape to, instead of using reels and reels of tape to, ⁓ a cartridge, ⁓ it, we had this idea that we would take that old per tech interface and we built a per tech to SCSI to exabyte. And we would sell that as a drive to these older mini computers so they could drop, you know,

backup their system on a cartridge or a few instead of, know, boxes of these tapes. So we went to the market with that, but we had found a little company in Boulder called Colorado Tech Design, who had taken that exabyte drive and they had built a very simple, well, I wouldn’t say it’s simple. They had a robot for it. And so they could put two exabyte drives.

in this robot and they had a carousel that would rotate around and then mechanisms that were mounted on the drives that would pull one of these eight millimeter X-Bike cartridges into the drive. And we came up with a ⁓ manufacturing license to manufacture it. And we came to realize that

it was mostly unmanufacturable. So for instance, there was a, there was a shell ⁓ that they put over the drives and, we came to realize that they would buy a special ⁓ pot from a local kitchen store when his kitchen specialist stores drill holes in it and screwed on top of the drives. And, and it wasn’t highly repeatable.

And so we got this thing, it was a little bit of a pig in the poke and we had to build a team to go make it viable. And we, made it reasonably viable. We sold several hundred and maybe four or 500 of them. But in the end, was, it was really tough. The drives weren’t very good. They were hard to replace, but we had a team. We’d built an engineering team that had the mechanical.

electronics and firmware skills to do robots. And so, so we re did that design entirely. think I still have one of those old machines, but it didn’t even resemble it and built some of our very first tape libraries, which use X by and Sony AI T and we did a, did four millimeter drives and, we probably sold 10 or 12,000 of those. Uh, and so

You know, that sort of accidental near disaster turned into us having the skillset to do, do these machines. And so we continued on that path and built larger and larger libraries.

HOST OUTRO

It was interesting to me that Nathan and his team realized that trying to be everything to everyone can leave you with limited business prospects. So he made what may have been the counterintuitive choice to others when he sold the software division to Sterling Software and bet everything on tape libraries. That decision

Choosing focus over diversification transforms Spectralogic into the company it is today. I also found value in the 1 to 10x ratio he mentioned for customer revenues. When you sell a large tape library, the purchase price is just the beginning. Over 10 years, if you manage that relationship well, if you support them, upgrade them, maintain that equipment, that relationship, according to Nathan, is worth 10 times more.

That’s not just smart business. That’s fundamental understanding that in enterprise technology, you’re not selling products. You’re building partnerships that can span decades. And competing against the Redwoods, IBM, storage tech, quantum. Nathan didn’t try to outvolume them or undercut them in price. He focused on technology, leadership, and customer intimacy. He built the world’s largest tape libraries and maintained relationships that

outlasted many of those giants. In our next episode, we’ll explore how Spectrologic navigated the industry’s evolution, the challenges of staying independent, and Nathan’s vision for the future of data storage. If you enjoyed this episode, please subscribe to the LTO show wherever you get your podcasts. And if you’re working on your own venture, remember Nathan’s lesson. Sometimes the best engineers are the ones who simply refuse to accept

that something can’t be fixed.

You’ve been watching the LTO show. Thanks for joining us.