The same AI boom Rivian hopes to use in its cars is making basic components harder and more expensive to buy, CEO RJ Scaringe and hardware chief Vidya Rajagopalan said in an interview with Kleiner Perkins, InsideEVs reported. Scaringe said the company now spends an "enormous amount of time" planning around memory, using it carefully, and thinking strategically about how to source it.
A shortage of memory chips, especially DRAM, is rippling across the tech industry. DRAM makers have been directing capacity toward higher-margin sales to AI firms, leaving tighter supplies for automakers. Apple blamed a data-center-driven rise in memory costs when it raised prices on laptops and other devices earlier this year, InsideEVs noted. China's Li Auto, in a filing earlier this year, pointed to "significant supply tightness and price volatility in memory components, such as DRAM and NAND flash," which it said are essential for its smart cockpit and driver-assistance systems. GM and Ford both flagged higher memory costs on recent earnings calls and announced long-term supply deals with Micron this summer.
On Rivian's latest earnings call, Scaringe said the company is looking for ways to offset rising costs of memory and other commodity parts so it can still hit the R2's advertised $45,000 base price.
Not just memory
"We deal with this all the time, supply chain shortages, it's a bane of our existence," Rajagopalan said. With the data-center boom, she said, "a lot of the e-glass for [printed circuit boards] is really hard to find," and a shortage in one area can throw a program off. Multilayer ceramic capacitors, or MLCCs, are another part Rivian did not used to worry about. "MLCC capacitors, you never thought about them. You're thinking about them now, right?" she said.
The squeeze comes as Rivian is also building up its own AI hardware. An in-house processor, the RAP1 chip, is set to reach customer R2s next year, and together with a new lidar sensor is meant to support higher levels of automated driving, InsideEVs said. Rivian is trying to cash in on AI with an in-car assistant, automated driving, and a robotics project, while the same data-center demand is making the ordinary parts inside those systems harder to buy.
InsideEVs framed the squeeze as a problem for ordinary parts in increasingly software-heavy vehicles, not a shortage of one exotic chip. DRAM makers are sending capacity to higher-margin AI customers, and that shift is now showing up in earnings calls and in the cost of building a $45,000 R2. Rajagopalan's point about e-glass and MLCCs is that a bottleneck can appear in materials automakers barely used to track. Scaringe's "enormous amount of time" line is about planning and sourcing, not a claim that Rivian cannot get parts at all.