Could an Xbox Series X cluster with 160GB unified memory host local LLMs? A Reddit thought experiment
RecursiveCTE · reddit · 2026-09-16
A Reddit user explores using Xbox/PlayStation consoles as local AI hosts, based on a discussion with Gemini, and asks the community for technical feedback.
Key points:
- Microsoft offers an official Developer Mode for Xbox One/Series X|S, theoretically opening the hardware
- Consoles pack 16GB unified memory; with free access you could theoretically run a 4-bit quantized 3B-7B open-weight LLM, leveraging GPU memory bandwidth like an entry-level workstation
- Scaling to a 4-console cluster — Xbox Series X dev kits carry 40GB GDDR6 each — would yield 160GB of fast unified memory, enough for large open-weight models or high-concurrency multi-agent setups
The post is largely speculative; real feasibility hinges on dev-mode sandboxing and the lack of CUDA ecosystems, which commenters are likely to dissect.
More from Infra
- Only 17% of IT leaders have unified AI observability strategy, Omdia survey of 500 finds — anacondainc · 2026-09-16
- Amazon SVP Peter DeSantis: model and silicon roadmaps must anticipate each other years ahead — dawnsongtweets · 2026-09-16
- Perplexity's CobbleDB cuts p99 latency from 123ms to 24.2ms, replacing DynamoDB — perplexity_ai · 2026-09-16
- CobbleDB vs DynamoDB: median batch-read latency 31.4ms→5.6ms, 20%+ cost savings — perplexity_ai · 2026-09-16
- Inside CobbleDB: partitioned RocksDB MultiGet reads with same-zone-first routing — perplexity_ai · 2026-09-16
- Perplexity: two engineers plus hundreds of AI agents built CobbleDB in two months — perplexity_ai · 2026-09-16