Stadium Seating

There's something intimate about watching sports in-person with 20, 30, 50-thousand of your closest friends. The energy. The shared experience. The feeling of being "in it" together. There's nothing like it. It's why Rome built the colosseum 2,000 years ago. It's the reason LA built "The Coliseum TM" a little while later.

There's also something intimate — in a "maybe-I'd-rather-not-share-this-moment" way — about braving stadium restrooms. You face a decision — do you go during the run-of-play when they're empty or do you hold out for halftime and wait in line around the concourse? These are real moments of suspense.

Either way, you count on them working and count on them being equipped. There are snafus, for sure. But the biggest worry is usually the line, not whether there will be toilet paper.

It's a small miracle that this works out. After all, stalls sit empty almost all the time. They're only ever tested during those 15 minutes between halves. Each stall needs toilet paper. No stall can run out. Otherwise, all you have is an uncomfortable porcelain seat with an "obstructed view," of a very long, grumbling line.

But keeping things stocked is important work. If it works poorly — if it doesn't work — you notice. The whole experience can die without this one little thing working. Unseen heroes keep the paper chain moving from the store rooms to the stalls.

Silicon Paper Chain

Rambus is a semiconductor company that designs and sells the interface chips that let processors talk to memory. It's kit allows data to move around quickly and safely. This is an unsexy job as far as semiconductors go — it's not making the gear that trains AI models, or even the memory to feed those engines. Its gear is barely noticeable when you look at a memory chip. It’s interconnects just move data from one part of a computer to another. It delivers the toilet paper or whatever else the system needs.

Days of Yore

When PCs were ascendant, this was more of a blocking and tackling business. Back in the late 1900s, eeking out a little more performance from my PC memory didn't deliver much value while my proto-metaverse pioneers died of dysentery on the Oregon Trail. High-performance memory systems were a relatively small part of the market and more parts were "off-the-shelf."

The Future is Cloudy

Then, The Cloud happened. The proliferation of data centers hosting The Cloud, from AWS, Azure and Google, along with data center specialists and enterprises themselves is driving much higher performance needs for memory systems. With billions at stake and an imperative to have data centers function more like stadium restrooms at halftime, all the time, every element of performance gets looked at. There's a keener focus on squeezing every bit of efficiency out of the whole ecosystem. And more value to find in each tiny corner.

The upshot is that data center operators want higher-value interconnects and systems and want new technologies altogether. Rambus can deliver far greater value in this environment than in the PC-era.

This transition, despite the hype, is still early. These forces will drive demand for interface chips and for Rambus.

The lab’s newest kit improves both power and performance. It has introduced more features to lower customers' operating cost and improve data security. It's introducing a system that will allow pooling of toilet paper, err, memory resources throughout a data center rather than each stall, err, server, having dedicated resources — a bit like going from using port-o-potties to having indoor plumbing. It's a big deal.

This boon comes along-side a successful cycle of long-term customer contract renewals that also brought renewed attention to the company. That attention recently drove Rambus' stock price higher.

Investor expectations for the everyday-hero-of-semiconductors are still restrained, though, compared to the backdrop their end-users’ world-beating ambitions. If shares work out from here, it will be on the deluge of new data centers adopting more and higher value memory systems.

Disclaimer: None of this is investment advice. It's meant to illustrate ways LCM thinks about investing. Things that LCM decides are good investments for LCM clients are based on many criteria, not all of which are covered here. Some or all of LCM's ideas may not be suitable for other investors. LCM does not recommend investing either long or short any position mentioned. LCM may own positions in some of the companies mentioned. Some of its ideas will lose money — investing entails risk.