AMD brings Zen 5 and 3D V-Cache to Ryzen Pro 9000 series workstation chips
3D V-Cache was a gaming party trick. AMD just crashed into the workstation segment with six new Ryzen PRO 9000 chips, two with X3D stacking, the flagship hitting 128MB of L3 cache at 170W.
AbJimroe
AMD spent years telling gamers that more cache means faster performance. Now it's making the same argument to architects, video editors, and simulation engineers.
AMD
AMD just did something that should excite anyone running a professional workstation, and probably annoy anyone who believed 3D V-Cache was exclusively for gamers.
The chip maker has announced six new Ryzen Pro 9000 Series desktop processors built on Zen 5 architecture. And, for the very first time, these processors bring 3D V-Cache technology to the commercial desktop market.
AMD Unsplash
What makes the X3D models special?
Among the six, the headline chip is the Ryzen 9 Pro 9965X3D, which comprises 16 cores, 32 threads, up to 5.5 GHz peak clock speed (with boost), and 128 MB of total L3 cache. While everything else might sound normal, the 128MB of cache is significantly larger than the standard 64MB.
The extra 64MB comes from AMD’s 3D V-Cache technology, which physically stacks additional last-level cache directly on the processor die and drastically reduces data latency under heavy workloads.
Below the Ryzen 9 Pro sits the Ryzen 7 Pro 9755X3D. It features eight cores, up to 5.2 GHz clock speed (with boost), and 104 MB of L3 cache. Both the chips do something that no other Ryzen Pro CPU has done before: break the 65W power limit.
While the 9965X3D runs at 170W TDP, the 9755X3D peaks at 120W. All the other CPUs, including the Ryzen 9 PRO 9965, Ryzen 9 PRO 9955, Ryzen 7 PRO 9755, and Ryzen 5 PRO 9655, are non-X3D models ranging from six to 16 cores and 65W to 170W, allowing OEMs to develop a full range of products on them.
AMD Unsplash
When can you get the new AMD chips?
All six chips support up to 256GB of ECC DDR5 RAM, PCIe 5.0 connectivity, and AMD’s Pro Technologies suite, including enterprise security, remote manageability, and a long-term platform stability stack.
The company is positioning its new Ryzen chips for media and entertainment. Think 4K/8K video editing machines, architecture, or engineering workflows that often include 3D modelling/rendering, and local AI inference workloads.
Systems powered by the new Ryzen PRO 9000 Series processors are expected to roll out in the second half of 2026. The first confirmed OEM system, based on the new chips, will come from Lenovo in the third quarter of this year (it’s called the ThinkStation P4).
Unfortunately, these chips won’t appear in the retail market, as they’re OEM-only. The pricing will also follow the standard Pro series practice: it won’t be disclosed publicly.
For more than five years, Shikhar has consistently simplified developments in the field of consumer tech and presented them…
Fluffy robot seals are being used for mental health care at a UK hospital
A fluffy robotic seal could soon be your unlikely companion at mental health appointments.
A baby seal robot for mental health care might sound like something out of a sci-fi movie, but Kent and Medway Mental Health NHS Trust is actually piloting one in its outpatient clinics, and the idea makes a lot of sense when you dig into it.
It’s a PARO robot (a personal robot designed for therapeutic purposes) which looks and behaves like a baby seal. It responds to touch, sound, light, and temperature with gentle movements and soft sounds. Think of it like a therapy animal, but without the unpredictability of a real animal. You can hold it, interact with it, and it reacts back.
Android apps can now track your habits and whereabouts to offer task suggestions
Contextual Suggestions is Android acting on what it already knows, and the fact that most users will never find the off switch is exactly why you should pay attention.
Google just flipped a switch on Android phones, and you probably didn’t notice. A feature called Contextual Suggestions is currently rolling out to Android devices, enabled by default.
It does what the name implies: watching what you do, learning your habits, and using the information to provide you with contextual suggestions or actions that you might perform next (via compatible apps).
Google could be prepping a powerful new Gemini AI model to outsmart ChatGPT
Google is expected to unveil a stronger Gemini model as AI rivals pull ahead in coding and agents
Google could announce a new Gemini AI model at I/O on May 19, and the timing is aggressive. A report says the release is expected to land near OpenAI’s GPT-5.5 class, while still trailing Anthropic’s Mythos, the model now shaping the industry’s frontier-model conversation.
For Google, the awkward part isn’t raw talent. A strong model can grab headlines, but developers don’t rebuild workflows for a leaderboard. They switch when a tool saves time, reduces cleanup, and survives real projects without becoming another tab to manage.
/cdn.vox-cdn.com/uploads/chorus_asset/file/25262608/DSC06592.jpg)
