Home › AI news › Digest, 20 Oct 2026 › DeepSeek open-sources DeepInfer, the engine behind its servers
DeepSeekOpen source
Trending on X
DeepSeek open-sources DeepInfer, the engine behind its servers
What happened
DeepSeek released its in-house inference framework that powers DeepSeek-V4.1-Flash, free and open source.
Why it matters to you
Running open models at scale just got cheaper for everyone.
Key number: runs on NVIDIA SM100 GPUs and Huawei Ascend 950 chips
Sources
More from DeepSeek
Newest first.
More AI news from 20 Oct 2026
The other stories of that day.
- Otter.ai · Otter.ai raises $320M from General Catalyst
- OpenAI · OpenAI's GPT-6.1 Terra spotted in Microsoft code
- Nvidia · ByteDance, Alibaba, Tencent order $35B of Nvidia chips
- Courts · Judges citing fake AI-made law face few consequences
- OVHcloud · OVHcloud plans its own AI model in early 2027
- GMO · AI hacking tool used on Korean banks also hit Japan
- SoftBank · SoftBank credit risk hits 4-year high on AI funding
- OpenAI · OpenAI rushes to raise funds before Anthropic IPO
- US government · US sanctions 5 Chinese entities over Google AI trade secrets
- Microsoft · Microsoft adds guard agents to watch rogue AI agents
- Amazon · Amazon's TorchNeuron lets AI apps switch off Nvidia chips
- OpenCode · Hackers breach OpenCode, run up AI bills
- Google · Google reportedly building Gemini 4 Carbon
- NVIDIA · Nvidia releases free Nemotron 3.5 models for agents and PCs
- OpenAI · Ex-OpenAI engineer in bugging case held as flight risk
- US government · US bars DeepSeek from civilian government agency systems
- MiniMax · MiniMax launches M3.1 Flash agent model
- Anthropic · Anthropic's Claude Fable 5.5 shows signs of launch
- OpenAI · OpenAI tests 'Aurora' image models on Arena
- Tenstorrent · Chip startup Tenstorrent raises $1.5B at $6.2B value
- Reuters · Reuters: hacker used AI agents against 100+ company sites