SANTA CLARA, June 27 — Marvell Technology has unveiled a new approach to making every gigabyte of memory count, introducing its Structera CXL controllers that can compress data by up to 3.64 times. The move comes as AI workloads devour memory at an unprecedented rate, and the broader tech industry faces widespread shortages of memory and storage solutions. Here’s what it means in plain terms.
Think of computer memory like counter space in a busy kitchen. When you run out of room, you can’t cook efficiently.
Marvell’s new hardware essentially compresses the ingredients as they’re being put away, then expands them again when the chef needs them — all without the chef even knowing it happened. The two new controllers, Structera X and Structera A, are part of Marvell’s CXL family. What sets them apart from most CXL controllers on the market is that they offer compression for data residing in memory.
Most competing controllers don’t provide any kind of compression for the data set. The secret sauce is a purpose-built hardware block called the Compression-Decompression Block, or CDB.
This dedicated silicon solution operates at full bandwidth. Unlike software-based compression, which can bog down a system, the CDB compresses data while it’s being written to DRAM and decompresses it when it’s being read — all in real time. The process happens independent of the host CPU, freeing up more memory resources for the platform.
Marvell’s compression block also adheres to specifications submitted to the Open Compute Project. This makes it the only CXL memory controller that offers inline compression as an in-production feature, according to the company. The practical upshot is significant.
As AI workloads continue to consume massive amounts of memory, the need for high capacities and bandwidth has never been more pressing. Memory shortages across the tech sector have prompted companies to come up with innovations to mitigate the capacity and bandwidth requirements for AI.
Marvell’s solution aims to squeeze more usable memory out of existing hardware. Structera X is designed as a CXL memory expansion controller with support for both DDR5 and DDR4 standards. The controller features four ARM Cortex M7 cores, the CDB block, Multi-Channel DMA, 56 MB of last-level cache, and four DDR channels.
It also includes AES-XTS 256-bit memory encryption and decryption. Structera A CXL is the other new controller in the lineup, though Marvell has provided fewer details on its specific capabilities in this announcement.
For everyday users, the benefits may not be immediately visible. But for companies running massive AI models in data centers, the ability to effectively expand memory capacity without physically adding more DRAM modules could mean significant cost savings and performance improvements. It also means less pressure on the supply chain for memory chips, which have been in short supply.
The question now is how quickly data center operators will adopt the new controllers. With memory shortages showing no signs of easing, the pressure is on for solutions that can stretch existing resources further. Marvell’s Structera lineup appears positioned to do exactly that.
{“@context”: “https://schema.org”, “@type”: “NewsArticle”, “headline”: “Marvell’s Structera CXL Accelerators Compress Data Up To 3.64x”, “description”: “SANTA CLARA, June 27 — Marvell Technology has unveiled a new approach to making every gigabyte of memory count, introducing its Structera CXL controllers that…”, “datePublished”: “2026-06-27 04:05:21T09:30:00+00:00”, “author”: {“@type”: “Person”, “name”: “Erika Johnson”}, “publisher”: {“@type”: “Organization”, “name”: “InfoPulse Today”, “url”: “https://newsfeed.infopulsetoday.com/”}}
