After an XDA writer replaced Synology's own apps with Immich, Navidrome, Plex and Jellyfin, the open-source server maker explained how it builds NAS systems with no vendor in the middle.
The open-source server maker explained why plain Ubuntu plus KVM is the host it recommends, why Proxmox and IncusOS run on the same kernel module, and offered its in-house eVirt framework on a consulting basis.
The open-source server maker released six months of its own daily memory price tracking and explained why the industry's three memory futures, sealed HBM, flat CAMM2 modules and CXL fabric, make upgradeable slots worth more than ever.
Anthropic's machine-checked proof was eleven days of sustained, verifiable AI work. The open-source server maker explains why that changes how businesses should size private AI hardware, and publishes the numbers.
The open-source server maker released the per-terabyte street prices it tracks daily to quote storage systems, showing enterprise and NAS drives 40 to 90 percent more expensive than in March, and what that changes for anyone sizing storage now.
Measured on the company's own bench: 54 tokens per second on Qwen3-14B, 26 on Qwen3.6-27B from a single card - and the provisioning behind those numbers is now a standard priced offering: $1,495, or $2,495 with a private RAG stack, included on flagship orders.
The open-source server maker published the numbers behind this year's component inflation - an 87 percent GPU list-price increase in under 18 months, server memory roughly four times higher per gigabyte, flash premiums of 30 to 50 percent - and the rent-versus-own math those numbers change.
Every eRacks server already ships with the open-source AI stack pre-installed free. The new service closes the last gap: models selected for the customer's field, tuned to their hardware, air-gapped, and handed off live to their team.
New Good, Better, Best ladder pairs current-generation AMD Threadripper and EPYC processors with a pre-installed open-source AI stack, and cuts mid-tier prices substantially from the prior line.
Single-card 32GB VRAM at $1,000 cost cuts price-per-GB-VRAM by 40-80% versus NVIDIA pro silicon. 4U flagship scales to 256GB unified VRAM for 70B-class language model inference.
WD Purple Pro and Seagate SkyHawk AI drives, 25/100GbE networking, and ZFS sequential-write tuning - on-premise storage built for IP camera systems and healthcare archives. From $4,995.
Every AILSA, AIDAN, AINSLEY, and AISHA ships with vanilla Ubuntu 26.04 LTS plus Ollama, Open WebUI, vLLM, PyTorch, CUDA, and Docker - ready to run Llama 4, Qwen 3, and DeepSeek V3.2 in 10 minutes.
Open-source NAS servers with full Linux, ZFS, TrueNAS, and no vendor licensing fees - on-premise storage that pays for itself in months versus cloud subscription costs.