I asked XDB to build a local AI workstation without spending RTX 5090 money
XDB can plan your local AI rig on a budget
In the pursuit of creating an AI workstation using consumer-grade components, it becomes increasingly costly and complex. XDB, XDA's PC building planner, simplifies this process by handling all the technical details and allowing users to create a custom build from the ground up. While gaming builds are the most common application, XDB can also assist in setting up workstations tailored to specific workflows, and it can accommodate very specific needs as well.
To build an AI workstation without spending the equivalent of an RTX 5090, an individual can input their desired workloads and expected performance into XDB. By setting a budget, users can ensure their chosen components stay within their financial limits. In this case, a budget of just over $2,500 was set to provide sufficient memory for running AI models while remaining reasonable for a workstation.
The recommended AI workstation features an XFX Swift Radeon RX 9070 XT 16GB graphics card, priced at $900. This card offers 640GB/s of memory bandwidth, which is significantly higher than other options within the same price range. The card is compatible with various local AI tools, including Ollama, llama.cpp, LM Studio, vLLM, and ComfyUI, making it a versatile choice for AI workstations.
The build also includes an Intel Core Ultra 7 270K Plus processor, an ASRock B860M-H2 motherboard, Cooler Master MasterLiquid 360 Core II cooling, 32GB of Patriot Viper Elite 5 memory, a SanDisk Optimus 5100 SSD, a be quiet! Pure Power 12 power supply, and an NZXT H5 Flow case. The total cost of these components amounts to $2,304, leaving $196 of the budget unused.
Although the choice between AMD's ROCm and Nvidia's CUDA platforms for local AI workflows is a common consideration, XDB provides useful insights into the compatibility of these tools with different hardware. In this case, the build was assembled with an AMD-based GPU and compatible software, resulting in a reasonably priced AI workstation that stayed under budget.
Written by urgent.news from XDA Developers's reporting — not their text. Machine-written — may contain errors; check the original before relying on it.
