AI is not just fighting for GPUs anymore. By 2027, the next big pressure point could be the same type of low-power memory used in smartphones — and that matters if you are planning to buy a future iPhone, Samsung Galaxy, or gaming phone in Malaysia.
According to figures cited by Citrini Research, NVIDIA’s upcoming Rubin AI platform alone is expected to require more LPDDR memory in 2027 than Apple and Samsung combined. That is a wild shift, because LPDDR has traditionally been closely linked with mobile devices, where compact size and lower power draw are critical.
The estimate puts NVIDIA Rubin demand at over 6,000 million GB of LPDDR memory in 2027. For comparison, Apple is projected to need around 2,966 million GB for iPhones, while Samsung is expected to require about 2,724 million GB. Together, Apple and Samsung would sit at roughly 5,720 million GB, meaning Rubin racks alone could be about 6% higher than both smartphone giants combined.
Why is AI suddenly eating phone-style memory? AI server platforms are moving toward huge memory capacities in tight, power-conscious layouts. LPDDR5 and LPDDR5X fit that requirement nicely because they offer high density while keeping power usage lower than some traditional memory approaches. For massive AI racks running next-gen workloads, that combo is very attractive.
And Rubin is not the only platform in the queue. The report also points to AMD’s MI400 family and other agentic AI-focused systems adding even more pressure. Micron has already introduced 256GB LPDDR5X SOCAMM2 memory, while SK Hynix is mass-producing 192GB LPDDR5X SOCAMM2 modules for NVIDIA’s Vera Rubin platform. AMD’s Verano CPUs and MI455X chips for Helios racks are also expected to support LPDDR5X.
For Malaysian buyers, the impact is straightforward: if AI servers keep absorbing more LPDDR supply, phone brands may face tighter component availability. That can translate into delayed launches, fewer units, or higher prices — especially on flagship phones where memory configurations keep climbing. Think iPhone Pro models, Samsung Ultra devices, foldables, and gaming phones that already sit in the premium RM bracket.
This does not mean your next phone will instantly become impossible to buy, bro. Apple, Samsung, Micron, SK Hynix, and other major players are not sitting still. The memory industry is expanding production, and newer facilities are being built to narrow the supply gap. But building fabs and ramping memory output is not something that happens overnight.
The bigger story is that AI infrastructure is now competing directly with consumer tech for key components. We already saw how GPU demand changed PC gaming prices over the years. Now, memory could become another battleground, except this time it touches nearly every smartphone buyer too.
If these estimates hold, 2027 could be a messy year for phone makers. AI companies want massive memory pools for bigger models and more advanced workloads. Smartphone brands still need LPDDR for millions of devices. Somewhere in between, consumers in Malaysia and SEA may feel the pressure through pricing, stock levels, or slower availability of certain models.
So yes, NVIDIA Rubin sounds like server-room stuff. But the ripple effect could land right in your pocket when you are comparing phone prices on Shopee, Lazada, or at the mall.
Source: Wccftech Gaming