Mobile World Congress is where vendors from across the mobile ecosystem get together to show off the latest devices, components and services. One of the most essential players in the mobile industry is Micron, which offers memory products including LPDDR5X memory and Universal Flash Storage NAND. As games, AI models, video files and photos get bigger and use richer data, storage technologies are pushed to the limit of speed, capacity and size. So it’s notable when a company like Micron has an important new product to introduce to the market at MWC.
UFS 4.0 Update
Micron’s UFS 4.0 product was introduced in June 2023. This product combines Micron’s 232-layer NAND flash with its proprietary controller technology to build a minuscule SSD for smartphones, tablets and other mobile devices. PCs generally use M.2 NVMe SSDs, which are traditionally much larger and consume more power, so UFS is mostly for mobile uses. In fact, the OnePlus 12 smartphone I just reviewed has UFS 4.0, although there are no teardowns yet to verify whose UFS is inside that phone. UFS 4.0 also offers better than 25% lower power consumption compared to the previous-generation UFS 3.1. It also has impressive sequential speeds of 4.3 GB/s read and 4 GB/s write, which is very good but still lags behind NVMe PCIe 5.0 SSDs, which are currently in 14 GB/s territory—but which also require a lot more space and power.
Many flagship phones are currently using UFS 4.0, including Samsung’s flagship line, which is starting at 256GB capacity. Micron’s UFS 4.0 modules top out at 1TB capacities but also offer 512GB and 256GB. These days, 128GB capacities don’t seem like a viable option, and I welcome that for many reasons. Truthfully, when I review phones now and have been using them regularly for a year, I find that 256GB is barely enough storage anymore, and 512GB is a much better fit for me. Micron also claims that UFS 4.0 solutions with the 232-layer NAND are 40% faster at loading applications than the previous generation’s 176-layer UFS 3.1.
Next-Gen UFS 4.0 Module
The UFS 4.0 module from Micron reduces the size of the complete package from 11mm x 13mm (143 square mm) to 9mm x 13mm (117 square mm), representing a nearly 20% reduction in area while maintaining the same speeds and capacity. Micron claims that this is the world’s most compact package size. Space on wearable devices or smartphone motherboards is always at a premium, so smartphone OEMs welcome any reductions in size. I believe that with these reductions, we could also potentially see larger capacities without needing new and even higher-density NAND.
Three New Key Features
In addition to the smaller size, Micron is also shipping its next-generation UFS 4.0 solution with three new features specifically aimed at enhancing the flagship smartphone experience. One-button refresh is a feature that helps to automatically cleanse and defragment data so that the smartphone’s files, including its operating system, continue to run in a like-new state. OBR also results in faster read/write performance, resulting in 10% faster app launches, better camera responsiveness and smoother multitasking.
Meanwhile, high-performance mode helps optimize device performance during intensive use by prioritizing critical actions over background tasks that are idling. This also results in an improvement of more than 25% in speed when launching applications, which Micron attributes to “twice-as-fast” storage access during heavy use.
Last but not least is zoned UFS or ZUFS, which allows an OEM to set up specific zones for read/write traffic, allowing faster access to critical data and improving drive wear over time. Micron says that implementing ZUFS under full array fragmented conditions shows more than 140% improvement in random reads and 180% in sequential writes. Micron says that it developed these features in partnership with its customers in joint labs in the U.S., China and South Korea.
What Does This Mean For AI Applications On Phones?
Micron’s new UFS 4.0 product should give phones enough storage to fit the enormous foundational models of localized AI, as well as accommodate the speeds at which data must go into and come out of those models. While I do believe that bigger and faster storage will absolutely enhance the AI experience, storage alone will not solve the challenges of on-device AI, because there are still considerations related to the NPU and memory capacity and speed. That said, if your on-device storage is not fast enough, it will bog down AI applications. This is why Micron and other companies are so excited about the potential impact of improved storage for on-device AI. Beyond that, I’ve seen how quickly one can run out of space when using AI models on my PC; I believe that AI will drive OEMs to increase storage capacities even on mobile devices to the point that 256GB becomes the default.
Wrapping Up
Micron continues to innovate in the mobile space, whether it’s delivering some of the world’s fastest memory with LPDDR5X or with innovative, AI-friendly storage in the next-generation UFS 4.0 package. There was a lot of talk about AI at MWC, which will require so many players in the mobile ecosystem to work together to solve the hardware and software challenges that AI presents.
Judging from this product announcement at MWC, it appears that Micron and its OEM partners are up to the challenge. They have built what looks like an excellent storage solution that should stand the test of time and will enable developers to reliably deploy next-generation AI applications that otherwise would’ve been impossible.
