Samsung just kicked off mass production of its fastest SSD ever — and it’s not for your laptop. The PM1763 is the company’s first PCIe 6.0 solid-state drive, built specifically for AI servers and high-performance computing (HPC) systems that power tools like ChatGPT and Gemini. Here’s what’s inside, who it’s really for, and why it signals a bigger shift in how the tech industry thinks about storage.
What Is the Samsung PM1763?
The PM1763 is an enterprise SSD — short for solid-state drive, the fast storage chip inside computers. Samsung announced it on July 8, 2026, and it’s already rolling off production lines.
In plain terms, it’s a storage drive designed for massive server farms, not personal devices. It uses Samsung’s ninth-generation V-NAND flash memory paired with a brand-new 4-nanometer controller chip. The result: up to 28,400 MB/s sequential read speeds and 21,900 MB/s sequential writes in the 16TB configuration. That’s roughly twice the performance of its predecessor, the PM1753. It comes in three sizes — 4TB, 8TB, and 16TB.
For perspective, a typical consumer NVMe SSD you’d put in a gaming PC reads data at around 7,000 MB/s. The PM1763 is four times faster. But it uses a completely different form factor (E3.S) and runs in liquid-cooled server racks, so you couldn’t physically install one in a regular computer even if Samsung would sell you one.
Why Does AI Need Faster Storage?
Here’s the part that often gets overlooked: training and running AI models isn’t just a GPU problem. Large language models (LLMs) — the engines behind AI chatbots, coding assistants, and image generators — need to load enormous amounts of data before they can do anything useful. Samsung says the PM1763 can transfer a 40GB LLM in approximately 1.4 seconds.
That sounds like a niche flex, but it matters. When your data storage can’t feed the GPU fast enough, those expensive AI accelerators sit idle. Wasted GPU cycles at data-center scale translate directly into wasted millions of dollars. As AI models grow toward trillions of parameters, the gap between compute power and storage speed is turning into one of the industry’s defining bottlenecks.
Samsung’s recent 18x profit surge forecast, driven by AI memory, makes the strategic motivation clear. The company isn’t just selling storage — it’s building the plumbing for the AI economy.
Samsung Isn’t First — Micron Got There in February
There’s an important asterisk. Samsung’s PM1763 is Samsung’s first PCIe 6.0 SSD — but it’s not the industry’s first. Micron started mass-producing the 9650 in February 2026, claiming the title of the world’s first shipping PCIe Gen6 data center SSD.
Micron’s drive hits similar top-line numbers: up to 28 GB/s sequential reads, up to 5.5 million random read IOPS, and liquid-cooling support. And SK Hynix has confirmed it will enter the PCIe 6.0 supply chain before the end of 2026.
What makes Samsung’s entry interesting isn’t the raw specs — it’s the signal. When the world’s largest memory manufacturer moves a product into mass production, it typically means the wider ecosystem is ready. Server platforms from AMD (EPYC Venice), Intel, and NVIDIA (Vera architecture) with native PCIe 6.0 support are expected to ship in the coming quarters. The hardware is finally catching up to the interface.
What Makes the PM1763 Different?
Beyond speed, Samsung is emphasizing two things: cooling and security.
Cooling. The PM1763 is designed specifically for liquid-cooled server environments using direct-to-chip (D2C) technology. AI data centers are increasingly ditching fans entirely in favor of liquid cooling to manage the heat generated by racks packed with GPUs and high-speed drives. Samsung says the PM1763 delivers 1.8x better power efficiency than its predecessor, which is critical when data centers are fighting for every watt they can save.
Security. The drive supports post-quantum cryptography (PQC) — encryption algorithms designed to resist attacks from future quantum computers. It also includes support for TDISP (TEE Device Interface Security Protocol), which secures data pathways in virtualized server environments. This is forward-looking stuff. Nobody’s cracking current encryption with a quantum computer today, but the data stored on these drives might still be sensitive in five or ten years.
When Will Consumers Get PCIe 6.0?
Not for a while. PCIe 6.0 enterprise SSDs like the PM1763 and Micron 9650 use specialized form factors (E1.S, E3.S) that don’t fit into standard desktop or laptop slots. They run at power levels (18–25 watts) that would drain a laptop battery quickly and generate heat that consumer chassis aren’t designed to handle.
More importantly, most consumer workloads don’t even push PCIe 4.0 SSDs to their limits yet — let alone 5.0. The jump to 6.0 for everyday users is probably three to four years away, at minimum. What you will notice sooner is faster cloud services, snappier AI responses, and shorter wait times for generative AI outputs — powered by exactly these types of enterprise drives sitting in a server room somewhere.
Storage as the New Bottleneck
The race to faster SSDs tells a broader story about where the tech industry is headed. For years, GPUs grabbed all the headlines — and all the investment. But now that companies have spent billions stacking NVIDIA H100s and Blackwell chips into every available data center rack, the bottleneck is shifting. Storage and memory are increasingly the thing holding AI back.
Samsung, Micron, SK Hynix, and Kioxia all know this, which is why the enterprise SSD market is expected to roughly double in revenue through 2026, according to TrendForce. Kioxia is working on drives that hit 10 million IOPS. Solidigm plans to ship 245TB SSDs before the year ends. Samsung itself has a 512TB PCIe Gen6 SSD on its roadmap for 2027.
None of that is coming to your laptop. But it is shaping how quickly AI can train new models, how fast your cloud apps respond, and how much that next GPU-heavy service costs to run. In the AI era, storage has quietly become the hardware that matters most — even if nobody’s buying it at Best Buy.
FAQs
What is PCIe 6.0 and how does it differ from PCIe 5.0?
PCIe 6.0 doubles the data transfer bandwidth of PCIe 5.0, reaching roughly 128 GB/s across 16 lanes. It uses a new signaling method called PAM4, which packs more data per signal cycle. For practical purposes, it means enterprise SSDs can move data twice as fast — but consumer devices won’t adopt it for several years.
Which companies are making PCIe 6.0 SSDs in 2026?
As of mid-2026, Micron shipped the first Gen6 drive (the 9650) in February, Samsung began mass production of the PM1763 in July, and SK Hynix confirmed it will enter the PCIe 6.0 supply chain by year’s end. Kioxia and SanDisk are also developing next-gen drives for AI data centers.
Will PCIe 6.0 SSDs make consumer PCs faster?
Not anytime soon. Current consumer SSDs mostly use PCIe 4.0 or 5.0, and everyday tasks like gaming, browsing, and video editing rarely saturate those speeds. The Gen6 interface matters most for large-scale AI workloads where storage bottlenecks cost real money.
What does liquid cooling mean for SSDs?
Liquid-cooled SSDs use direct-to-chip coolant channels instead of fans to manage heat. As AI servers pack more high-power components into dense racks, traditional air cooling can’t keep up. Samsung’s PM1763 and Micron’s 9650 are both designed for these liquid-cooled environments.
How does post-quantum cryptography protect data on SSDs?
Post-quantum cryptography (PQC) uses encryption algorithms that can’t be broken by quantum computers — which could one day crack today’s standard encryption. Samsung’s PM1763 includes PQC support to future-proof sensitive data stored in enterprise environments.
The post Samsung’s First PCIe 6.0 SSD Could Load AI Models in Just 1.4s appeared first on Memeburn.