Elf Poaching
There's a scene in the iconic Will Ferrell Christmas movie Elf, where James Caan lays into his underlings. Caan works at children's book publisher Greenway Press. He's the grumpy father of the film's hero, Buddy the Elf/Will Ferrel. Caan missed his quarterly numbers. His boss—Greenway—made threats. He needs the next home run children's book. His underlings—his "A-team" writers-come back to him with their "big idea:" hire a ringer.
If you're a business, you go hire McKinsey or Bain to get a distilled copy of competitor's strategies. But, like Steve Jobs pointed out, "Good artists copy. Great artists steal." You could poach the talent.
Royal flop
Remember the year 2015? The dress was gold (or was it blue?) The force had awakened. And DRAM memory licensor Rambus had to do some soul searching.
Stick-in-the-mud accountants—FASB—had announced that flat royalties not tied to product sales would no longer be recognized as revenue over the term of a license. Companies would recognize them up front. This mattered to about 0.1% of public companies. Rambus was one of the 0.1%.
Royalties were Rambus' business. This would flatten revenue. Money would still come in the door, but would investors notice? It wouldn't be on the P&L. This would confuse the crap out of investors. They'd have to actually read its financial reports to understand it. What a drag. The company needed to show something on its income statement to investors for the next decade.
Rambus still had plenty of cash coming in the door. It had cash on its balance sheet. It had ideas too, probably: "What about this company…" management suggested to directors, "Integrated Device Technologies—IDTI. They sell memory-related stuff just like what we license. They've doubled that business over the last five years. That seems good. We could buy it."
"Too big,” said the board.
"OK. There's this other company, Inphi.” Management pitched their next idea… “It has a tiny memory interconnect business—things that clip into data centers. It's trying to shed it. We could grab that one?"
"Can we do it cheap?"
"Yep."
"We're in."
Rambus bought Inphi's memory interconnect business in 2016 for $90M. And that's, roughly, maybe, how Rambus pivoted from a royalty company to a product company.
Management had some other ideas too. Too many ideas. Some not great ones. It bought a card emulator for mobile transactions. It bought a serializer/deserializer. It bought some physical interface IP.
These were side quests. In 2018, the company booted its CEO for some undisclosed conduct shenanigans. He'd been helming all the M&A. Rambus promoted its current CEO—Luc Seraphin—from its memory business. Luc shed most of these distractions. The company would focus on the memory business.
It worked. Memory sales growth clipped ahead at 30% for the first few years.
Plot twist
Imagine, then, the board watching with a bit of envy the share price success of IDTI. Its memory business grew by 16% a year through 2018. Slower than Rambus. Its memory business accounted for more than half of that company's growth over the decade to 2018. It attracted a bid from Japanese semiconductor shop, Renesas. Renesas wrapped IDTI into its corporate tent in 2018. It paid nearly 23 times earnings for the outfit.
Meanwhile, Rambus appeared to struggle against the accounting changes. Its overall profit was trash despite the success of the product business. And investors didn't want to pay much for what profit it did make. Cash was coming in the door. Investors didn't care. It wasn't going through the P&L.
Luc went back to the drawing board.
The board pressed, "How can we hit a home run, like IDTI?"
"Here's what we do: We bring in a ringer."
The Elf maneuver.
So, they did. Rambus poached Renesas execs like a bandit. It brought in Sean Fan, the Integrated Device Technologies exec who ran its memory biz. They hired Des Lynch, IDTI's VP of Finance as Rambus' new CFO. The board brought in new Director, Necip Sayiner, President of Renesas Americas and former CEO of Intersil another semiconductor company bought by Renesas. After the move, Luc, Des and the gang pushed Rambus' product growth to a 36% yearly pace over the next six years. The business made frothy margins.
The Sequel
So, success?
The company trades more like a product company now. It's no longer in investors' doghouse. Shares made progress towards the price IDTI won from Renesas. Investors value each dollar of Rambus' profit close to what Renesas paid for IDTI. Rambus is only a bit cheaper now.
Selling DRAM has been tough going the past few years. It's a cyclical market. The cycle has been down. Rambus' customers have been cautious.
Rambus missed a few beats too. Nvidia ghosted the company's and its competitor's devices. It doesn't use any of it. Nvidia has been the story in semiconductors the past few years. Rambus is barely a player in HBM, a higher-tech, higher-margin memory solution for AI than the company's bread-and-butter DRAM.
The company's price isn't screaming cheap anymore. Market headwinds. An AI miss. Does that mean the story's over?
Not quite. Last year the shop finished extending contracts with its big three licensing customers for at least the next five years. The company changed some of the licensing terms. Investors will see those dollars come in through the P&L this time around. The questions facing the company now are product questions.
Will the company continue ambling forward with its existing products? Or
Will Luc and Sean be able to recapture magic with new ones?
On its existing line, Rambus hasn't been immune to the swoon. Product sales fell 1% in 2023. But they picked up again last year as the market pivoted to the newest generation of DRAM. Rambus holds a much stronger position in this generation than the previous one. Sales rose 10% while the overall market slipped. Over the past four years product growth has averaged 20% per year. And, by the end of 2024, product sales tore ahead at a 37% clip again.
Conventional data centers still need Rambus' kit. Those are still popping up all over the place. Its market is growing. Older DCs need replacements and upgrades. Management cut its teeth in every sort of market. It ought to find a place in one that’s expanding. If all Rambus achieves is to continue selling into these, that ought to give investors a good-not-great result.
But, management's also cooked up a couple new ideas. These could land in large new markets. Or they could fall flat. If the sequel is a blockbuster, it'll turn on the product chops management's built over the past decade. Since adopting a product business from nearly nothing, management's grown it at a 34% pace. It's outpaced the market by miles. It's outpaced Rambus' share price too. Given that, investors might consider giving management a chance to cook.
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 and its 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. See full disclaimer here.