推进私有AI计算：引入安全服务器端内存技术
该动态宣布为面向个人AI的私有AI计算（Private AI Compute）引入安全的服务器端内存技术。由于目前提供的发布信息较为简短，文章主要聚焦于在云端服务器环境中强化内存隐私与数据安全，旨在兼顾大规模算力扩展与个人AI数据的机密性。具体技术细节及完整实现方案尚需进一步披露。
DC AI 热点
全部 AI 动态
9月24日2026-09-24
Watermarking in vLLM
How vLLM implements distribution-preserving Gumbel-max text watermarking with efficient GPU kernels, statistical detection, speculative decoding, and repeated-c
9月23日2026-09-23
v0.30.1rc0: [ROCm][CI] Add MI355 dense NVFP4 and MoRI kernel mirrors (#58281)
Signed-off-by: Andreas Karatzas akaratza@amd.com Co-authored-by: OpenAI Codex noreply@openai.com
黑客声称入侵了 FBI 窃取雇员信息
勒索组织 ShinyHunters 声称入侵了 FBI 窃取了逾 2TB 雇员数据。该组织的一名发言人称，这次行动不是出于经济动机，而是要求 FBI 更正或撤回此前发表的声明，其中包含大量不实的指控。ShinyHunters 称它利用了 FBI 招聘网页的一个 Oracle PeopleSoft 的 0day 漏洞，该漏洞允许在服务器上远程执行代码。该组织随后篡改了页面，替换为已被其控制的横幅和图片(This site has been seized by ShinyHunters)。ShinyHunters 从 FBI 管理的 AWS GovCloud 服务器上下载了约 2TB 至 3TB 的数据，这些数据涉及 FBI 的现有和前雇员，以及求职者。 FBI 在今年五月就 ShinyHunters 发出安全警告，称该组织采用“骚扰策略，向受害者及其家属发送威胁性短信和拨打骚扰电话，在某些情况下还包括恶意报假警（swatting）”。ShinyHunters 声称这些指控不实。
v0.34.4-rc0: mlx: speed up Qwen 3.8 prompt processing (#18550)
mlx: speed up Qwen 3.8 prompt processing Use MLX's gated-delta kernel for long scans and fold dense MLP global scales into SwiGLU. address comments
Topology-Aware Workload Scheduling with NVIDIA Topograph
AI factories are power-limited systems that deliver maximum value when fully optimized. GPU workload placement is a key optimization. Poor workload placement...
9月22日2026-09-22
Right-size generative AI endpoints with concurrency sweeps on Amazon SageMaker AI
Concurrency sweeps help you right-size a generative AI endpoint on Amazon SageMaker AI by systematically benchmarking it at increasing load levels. This post walks through deploying a model, running automated concurrency sweeps with the CreateAIBenchmarkJob API, and using the results to make data-driven capacity decisions about fleet size.
Cloudflare Python Workers are now generally available
Cloudflare Python Workers are now generally available After a two year preview, Cloudflare's support for running Python code in their server-side Workers platform is now stable: "Python is now a first-class, fully supported language on the Cloudflare Developer Platform". A neat thing about this is how it works. Cloudflare are running Python compiled to WebAssembly via Pyodide in their V8-based workerd runtime. This comes with some limitations, documented here - most notably both multiprocessing and threading are non-functional in the WebAssembly VM. One particularly interesting detail of this is the local development environment story - their
Security auditing in the age of (good enough) AI
Comments
Simplifying Model Serving Across Multiple GPUs with NVIDIA TensorRT Multi-Device Integration in NVIDIA Dynamo-Triton
The compute and memory demands of generative AI increasingly exceed what a single GPU can provide. NVIDIA TensorRT multi-device inference is a new capability...
NVIDIA Launches DSX Ready to Qualify Power and Cooling Products for AI Factories
Every AI factory needs power and cooling that fit its computing architecture. As AI infrastructure expands, power, cooling, water, site and grid constraints are shaping what builders can deploy. Choosing products that fit the complete factory design helps builders turn computing capacity into useful AI output. To help builders make those decisions, NVIDIA is introducing […]
How BMW Group detects cost anomalies across 14,000 cloud accounts
BMW Group operates CLEA, a FinOps platform monitoring more than 14,000 cloud accounts. This post shows how BMW added automated daily cost anomaly detection, moving from reactive dashboards to proactive alerts using Prophet forecasting, AWS Step Functions, and a serverless pipeline that processes every account for about $50 per month.
9月21日2026-09-21
Turn Your Latest Observations Into Timely Weather Decisions With NVIDIA Earth-2
Weather-sensitive industries increasingly have access to observations that offer an earlier, more local view of changing conditions. Energy companies collect...
从“算力焦虑”到“算力理性”：Akamai解答AI推理时代的ROI之问
过去两年，全球AI产业的关键词是“抢GPU”。从Meta、微软到xAI，动辄十万卡级别的集群建设不断刷新着人们对算力规模的想象。在这一背景下，笔者有幸采访到了Akamai云计算首席技术官Jay Jenkins，围绕AI算力的分配逻辑、推理场景的隐性成本、分布式架构的合规价值以及多智能体时代的基础设施挑战，进行了深入探讨。
9月19日2026-09-19
v0.34.3-rc1
server: allow registry cross-host redirects among allowlisted hosts (…
v0.30.0rc2
[Bugfix][NIXL] Avoid receive reports for notification-only requests (…
Amazon SageMaker Inference: 2026 year-to-date launches in review
Amazon SageMaker AI shipped 13 inference launches in year-to-date across two deployment paths: fully managed endpoints and Amazon SageMaker HyperPod Inference. This post reviews each launch, from inference recommendations and capacity-aware instance pools to tiered KV caching and disaggregated prefill and decode.
Introducing Kimi K3 on Amazon Bedrock
Kimi K3 from Moonshot AI is now available on Amazon Bedrock, giving you a powerful new open-weight option for coding and knowledge work. It offers native vision, a 1-million-token context window, and explicit prompt caching to reduce latency and input costs.
9月18日2026-09-18
How CSIRO built scalable, cost-optimized genomic variant querying on AWS
Learn how researchers at CSIRO, Australia's national science agency, built Serverless Beacon (sBeacon), a scalable serverless solution for securely querying genomic variant data on AWS. sBeacon implements the GA4GH Beacon standard using Amazon S3, AWS Lambda, Amazon DynamoDB, and Amazon Athena to support production-scale clinical and research applications.
How Equinix cut operational overhead with a shared services architecture on Amazon EKS
Equinix, the world's digital infrastructure company, built a shared services architecture on Amazon EKS to eliminate the operational sprawl of its self-managed Kubernetes environment. Learn how a multi-account North Star architecture centralized governance and shared services, delivering 4x faster deployments and 40% less operational overhead.
v0.34.2-rc2: mlxrunner: Release freed KV buffers during speculative decode
The decode loop releases MLX's pool of freed buffers every 256 generated tokens, which is also how often the KV cache grows and drops its previous, smaller buffers. The check fires only when the token count lands exactly on a multiple of 256. Speculative decoding emits several tokens per round, so most rounds step over the boundary and the pool is never released. Each growth at a long context leaves several GB of buffers that no later allocation can reuse, so the runner's footprint keeps climbing over a long generation until the system runs out of memory. We now release the pool whenever a round crosses a multiple of 256 tokens, which is what
Scaling Multi-GPU Video Captioning with PyNvVideoCodec and vLLM
How to leverage NVIDIA Hardware Video Decoders to Achieve Multi-GPU Scaling in Video Captioning and Description tasks.
9月17日2026-09-17
Building cloud-native PACS on AWS
A hybrid cloud architecture pattern for modernizing medical imaging on AWS. Learn how multi-hospital networks can centralize PACS archives, enable cross-facility interoperability, and use Amazon S3 storage tiers to manage cost and retention at scale.
How DHI Group accelerates generative AI workloads from idea to production using hackathons
Learn how DHI Group partnered with AWS to move generative AI workloads from idea to production using a structured hackathon. This post covers the Hackathon Acceleration Package, the winning ClearanceJobs and AgileATS agentic architecture on Amazon Bedrock AgentCore, and the principles that make hackathons a repeatable path to production.
v0.30.0rc1: [Bugfix] Isolate supplemental FlashInfer BF16 autotuning (#57285)
Signed-off-by: jiahanc 173873397+jiahanc@users.noreply.github.com Co-authored-by: OpenAI Codex codex@openai.com
Zombie Workloads Haunt Data Center Efficiency Efforts
Abandoned jobs, instances, and volumes can run indefinitely. FinOps tools help, but GPU-era AI demands new zombie-hunting methods.
Helping bring the world’s first large-scale, near-zero emissions steel plant online
Video opens with archival footage of smoky, coal-fired steel factories, then transitions to clean, modern exterior shots of Stegra's new facility and a clear digital animation showing how green hydrogen replaces coal to produce near-zero emission steel.
proto-v0.3.0
Release vllm-proto 0.3.0
v0.34.2-rc1: mlxrunner: lay out model by contract, checkpoint and construction
model is one package with three jobs: the contract between the runner and the architectures, the opened checkpoint, and building nn layers from checkpoint tensors. Its files did not say which was which. base.go carried the folded package's name over the interfaces and the registry, root.go held the safetensors header scan next to Root, and quant.go mixed the nvfp4 global-scale helpers with quant parameter resolution. base.go becomes model.go, named for what it holds. root.go keeps Root and Open; TensorQuantInfo and the header scan join quant.go, so everything the checkpoint says about quantization is read and resolved in one file. The global-
Translating CUDA Tile Operations from Python to Rust Using Agentic AI
cuTile Rust (cutile-rs) is a tile-based system for safe, idiomatic GPU kernel authoring in the Rust programming language. Extending the Rust ownership model to...