Skip to content

Su

Business coverage of technology, AI and cryptocurrency

Last update

Search results for «Inference Costs»

12 posts

Today 1

  1. Cloudflare AI Inference Adds Zero-Trust Security for Enterprise Use

    Cloudflare has unified zero-trust controls with its Workers AI and AI Gateway services to let enterprises run inference across multiple providers with policy enforcement and observability.

    3 min

29 October 1

  1. Google Cloud AI platform Unveils Ironwood TPU for Enterprise Inference

    Google Cloud launched Ironwood, its seventh-generation TPU, delivering substantial inference cost reductions and performance gains for enterprise AI deployments.

    3 min

22 October 1

  1. Cloudflare AI inference Unifies Workers AI and Gateway Under Zero-Trust Controls

    Cloudflare integrated its Workers AI and AI Gateway platforms on August 7, 2026, adding zero-trust identity controls and extended audit logging for enterprise users.

    3 min

21 October 3

  1. llm-d 1.0 Released by IBM, Red Hat and Google for Production Distributed Inference

    IBM, Red Hat and Google released llm-d 1.0 on October 20, 2026, adding production features for distributed LLM inference across clouds and on-prem environments.

    3 min

  2. DeepSeek DeepInfer Open Source Release Cuts LLM Inference Costs

    DeepSeek released its DeepInfer inference framework under Apache 2.0 on October 20, 2026, targeting lower costs for large-scale LLM serving.

    2 min

  3. MiniMax M3.1 Flash Launches with 70% Lower Agentic AI Costs

    MiniMax released its M3.1 Flash model on October 19, 2026, claiming major gains in speed and efficiency for agentic AI applications at flat API rates of $0.15 input and $0.60 output per million tokens.

    3 min

16 October 1

  1. Microsoft Maia 300 Inference Accelerator Targets 40% More Tokens per Watt by 2027

    Microsoft announced the Azure Maia 300 at the OCP Global Summit on October 12, 2026. The new inference accelerator enters production in 2027 with claims of substantial efficiency improvements over prior Maia generations and competing GPU systems.

    3 min

5 November 1

  1. NVIDIA AI Agents Context Window Study Finds 16K Tokens Maintain Accuracy

    A new NVIDIA preprint demonstrates that frontier models retain performance on agent tasks with frequent compaction to 16K-token windows instead of full long contexts.

    3 min

3 November 1

  1. OpenRouter Token Volume Hits 194 Trillion in One Week With 10 Quadrillion Annualized Run Rate

    OpenRouter processed 194 trillion tokens in the week of October 26, 2026, setting an annualized run rate above 10 quadrillion tokens for the first time.

    3 min

26 October 2

  1. Moonshot Kimi K3.1 Pro Launches With 1M Context Window and Open Weights Promise

    Moonshot AI released Kimi K3.1 Pro on October 23 with a 1M-token context window, MoE design, and planned open weights release by November 4.

    3 min

  2. Nvidia Flora open source delivers 125x cheaper robot video labeling than GPT-6 Astra

    Nvidia released Flora on October 20, 2026, providing three Qwen3.5-based models that label manipulation videos for robotics training at a fraction of prior costs.

    3 min

25 October 1

  1. Nscale Anthropic Deal Advances on Second $45 Billion Compute Agreement in West Virginia

    Nscale and Anthropic are in advanced talks for a second $45 billion AI compute agreement covering another 460 MW at the Monarch Compute Campus in West Virginia.

    3 min