DeepSeek open-sources inference engine; Anthropic reports $80.7B loss

DeepSeek has open-sourced DeepInfer, the inference engine powering its production models, under an Apache 2.0 license, allowing for commercial use. This release, built on vLLM and inspired by other frameworks, targets NVIDIA and Huawei hardware and includes optimized kernels and features for structured output generation. Separately, Anthropic's draft IPO filing revealed a net loss exceeding $80 billion in the first half of 2026, largely due to accounting charges, despite significant revenue growth and a projected $750 billion expenditure on compute capacity. AI

IMPACT DeepSeek's open-source inference engine could accelerate deployment of LLMs, while Anthropic's financial disclosures highlight the immense compute costs associated with frontier model development.

RANK_REASON The cluster covers a significant financial disclosure from Anthropic related to its IPO and an open-source release of a production inference engine from DeepSeek. [lever_c_demoted from significant: ic=1 ai=0.7]

Read on dev.to — LLM tag →

• Anthropic
• DeepInfer: Open-Source Deep Learning Deployment Toolkit for Image-Guided Therapy
• DeepSeek
• GitHub
• Hacker News
• Huawei
• Hugging Face
• Meta*
• Mistral AI
• Nvidia
• Prometheus
• Reuters
• SGLang
• TensorRT-LLM
• The Wall Street Journal
• TorchTPU
• vLLM
• Voxtral 3

• infra
• funding

AI-generated summary · Google Gemini · from 1 sources. How we write summaries →