Rubin’s LPDDR – A forecast from Citrini Research says Nvidia’s Rubin AI platform could require more than 6 billion GB of LPDDR memory in 2027—potentially exceeding Apple and Samsung combined. With consumer electronics already facing higher memory, storage, and graphics costs,
For shoppers hoping the next phone upgrade will come with better deals, a new AI roadmap is sharpening the squeeze—this time at the memory level.
Citrini Research projects that Nvidia’s next-generation Rubin AI platform will require more than 6 billion GB of Low-Power Double Data Rate memory (LPDDR) in 2027. LPDDR is the low-power memory used in devices such as smartphones, tablets, and ultra-thin laptops.
If the forecast proves right, Nvidia’s Rubin could end up consuming more LPDDR than Apple and Samsung combined.. That prospect lands at a particularly painful moment for consumers: rising memory costs have already been filtering into prices across consumer electronics. and the next demand surge from AI chipmakers risks making those downstream effects last longer.
Rubin—named after astronomer Vera Rubin—is central to Nvidia’s push into generative AI. Nvidia says the chips are designed to meet generative AI’s growing demand for real-time reasoning and will be twice as fast as Blackwell, Nvidia’s current flagship AI platform.
In March, Nvidia announced that between Blackwell and Rubin it had locked in $1 trillion in orders through the end of 2027.. The order book looks like good news for Nvidia and its investors.. For consumers. though. the timing is harder to ignore: the AI ramp is arriving as many households are preparing to replace the devices they bought during the pandemic.
That refresh cycle is already underway.. One widely used rule of thumb is that televisions are typically replaced every 6.6 years, according to Circana.. That would put more than 20% of the sets in use globally in an upgrade zone. and the move toward smart services increases the need for onboard memory.
Computers are also in the middle of an upgrade push, but the cost barrier has climbed sharply. RAM prices have risen anywhere from 150% to more than 200% over the past year. Storage prices—what consumers pay for hard drives and SSDs—have followed a similar trajectory.
Even Nvidia’s own PC ecosystem is feeling constrained pricing, with the company continuing to prioritize AI demand over the PC market.
Game consoles are part of the same story.. For the first time in nine generations of gaming hardware. console prices are going up instead of down on systems that have been out for a while.. Nintendo raised the price of the Switch 2 from $450 to $500 in the U.S.. earlier this month.. In March. Sony increased the price of the PlayStation 5 by as much as $150. with the high-end PS5 Pro now selling for $900.. Last October. Microsoft increased the prices of the Xbox Series X and Xbox Series S for the second time in six months; the systems now cost $650 and $600.
If smartphones and tablets see similar price hikes, the impact on consumer spending could be significant—especially during the holiday shopping season, when manufacturers roll out new devices and demand for upgrades typically spikes.
The memory squeeze looks poised to deepen beyond this year.. Nvidia’s use of LPDDR is expected to surpass that of either Apple or Samsung individually this year. though not both companies combined.. By next year, Nvidia is projected to consume 6.041 billion GB of LPDDR memory.. Apple’s projected capacity is 2.966 billion GB, while Samsung’s is expected to reach 2.724 billion GB.
Nvidia is not the only one leaning harder into LPDDR. Google and AMD are also expected to ramp up their memory usage, though at lower projected levels than Nvidia.
As multiple AI players step up at once. the broader effect is straightforward: AI’s growing appetite for LPDDR is likely to tighten supply and push prices higher.. Even though Samsung and Apple remain major customers for memory manufacturers. they are unlikely to commit to purchase volumes on the same scale as AI chipmakers.
The demand is also unlikely to cool soon because the stakes inside the tech sector are rising.. Memory makers are being pulled into a wider competition for AI leadership, driven in part by U.S.. companies racing against China for AI supremacy.. In January, Nvidia founder and CEO Jensen Huang acknowledged how competitive the industry has become.. “The number of startups that have emerged in China … speaks to the vibrancy and capability of the Chinese technology industry. ” he said.
For consumers, that matters because sustained AI competition can keep memory demand elevated—and elevated demand tends to show up somewhere. In this case, the forecasted memory pull from Rubin and the broader AI sector could continue to trickle down as higher prices across consumer electronics.
Nvidia Rubin LPDDR memory DRAM Apple Samsung Jensen Huang Blackwell AI hardware demand memory prices consumer electronics smartphone upgrade costs
So basically Nvidia is gonna make phones cost more? Great.
I don’t even understand LPDDR but if it’s “more than 6 billion GB” then yeah that sounds like it’ll hit regular people. Everything already costs more now, so AI is just the excuse to keep raising prices.
Wait I thought Apple and Samsung were the ones hogging memory… but the article says Nvidia Rubin might use more than them combined? That seems backwards to me lol. Also “twice as fast” doesn’t mean my phone will be twice as good so I’m not buying it.
This is gonna sound dumb but if AI chips need more “low power memory” then why do my current devices feel slow? Maybe it’s the storage, not the memory. And if TVs get replaced every 6.6 years like they say, we’re overdue anyway, so companies are just gonna charge us more because “AI roadmap” or whatever. I swear it’s always something new with Nvidia.