52
Technology2h 32m ago

A local AI inference mini PC, featuring the AMD Ryzen AI Max+ 395 chip, can now run 235-billion-parameter models on x86 architecture, potentially allowing developers to save on cloud inference subscription costs.

Shenzhen, China

Who
AMD, GMKtec
What
A local AI inference mini PC, featuring the AMD Ryzen AI Max+ 395 chip, can now run 235-billion-parameter models on x86 architecture, potentially allowing developers to save on cloud inference subscription costs.
When
Sun, 14 Jun 2026 09:54:52 GMT · 2h 32m ago
Where
Shenzhen, China ·
Why
The AMD Ryzen AI Max+ 395 chip integrates CPU, GPU, and NPU sharing a single 128 gigabytes unified memory pool, allowing it to load large AI models that discrete graphics cards struggle with due to VRAM limitations.
The Frontline Impact

How this affects you

This development could significantly reduce operational costs for developers currently paying for cloud AI inference, as the hardware pays for itself in under a year compared to $440 monthly cloud subscriptions. It also offers enhanced privacy and control over data by enabling local execution of large language models.

Story chain

3 events in this thread

Verified Sources & Citations

Credibility ratings reflect the AI ingestion pipeline's assessment of source provenance.