Search “AWS”

Amazon stockpiles NVIDIA GB300 servers but has no power to run them, exposing the power bottleneck of AI infrastructure

According to internal emails seen by Business Insider, Amazon receives NVIDIA server racks worth billions of dollars each month; in August alone AWS expects to receive about $5 billion worth of GB300 racks, and thereafter about $3 billion more each month, but some US AWS sites lack sufficient power, so this hardware can only be temporarily stored in transit warehouses. This phenomenon, known as dark GPUs, is putting financial pressure on data centers; Amazon's free cash flow turned negative in the second quarter for the first time since 2023, and capital expenditure this year is expected to be about $220 billion.

Amazon CEO Jassy on $220 Billion in Capital Expenditure: Compute Is Still in Short Supply, Inference Costs Need to Fall by Another Order of Magnitude

Amazon CEO Andy Jassy said in an interview on the a16z podcast that the company's capital expenditure is expected to reach $220 billion in 2026, but compute capacity is already sold out for 2026, essentially fully booked for 2027, and a large share of 2028 quotas has been locked in. He described enterprise AI demand as dumbbell-shaped: a large number of production systems in the middle have yet to be deeply integrated because inference costs are too high, and costs need to fall by another order of magnitude before an explosion in adoption occurs. Jassy also noted that 99% of AWS's top 1,000 customers use Graviton, which offers about 40% better price-performance than x86, while Trainium has a 30% to 40% advantage over comparable hardware.

Orange Cat Universe completes tens of millions Pre-A round financing, Gao Bingqiang invests in three consecutive rounds

Pet smart hardware company Orange Cat Universe recently completed a Pre-A round financing of tens of millions of yuan, with Professor Gao Bingqiang and his Gaofeng Patience Fund, Qianhai Innovation Investment, and Changlei Venture Capital continuing to add investment; the funds will be used for technological innovation, product iteration, patent system development, and global offline channel expansion. The first product, PawSwing Neo, is a powerless pure-mechanical cat bed grooming collector, priced at about 200 US dollars, with a return rate of less than 3% and grooming efficiency of 3 to 4 milligrams per joule. In 2026, the company has entered Europe's largest pet specialty retail chain channel, with more than 3,000 offline stores stocking its products.

AI accelerates cyberattacks; Anthropic, OpenAI, Microsoft and others disclose defense and agent control developments

gihyo.jp reviews multiple releases and reports from late October, organizing AI-era defense challenges along two lines: vulnerability remediation implementation and permission management. Anthropic Deputy CISO Jason D. Clinton says that within about 18 months, high cyber capability will spread to small, cheap models, but fixes still need to be applied to running systems. The article cites OpenAI's discovery of Brotli's CVE-2026-111497 during testing in an isolated environment, and introduces remediation and credential control solutions such as AWS Continuum, Google Agent Compute Engine, and 1Password for Agents, as well as Microsoft's report on Storm-0940 using AI to automate attacks.

Amazon SageMaker AI Multi-Turn Reinforcement Learning Practical Guide: Training gpt-oss-20b with a Wordle Example

AWS has published a practical guide for multi-turn reinforcement learning (MTRL) on Amazon SageMaker AI. Using Wordle as an example, it deploys the agent and reward function to Amazon Bedrock AgentCore, and the training job updates LoRA adapters via GRPO. In the example, gpt-oss-20b at default reasoning effort improved its solve rate on held-out words from 61% to 84%, at a training token cost of about $339 and an AgentCore runtime cost of about $79. The article also details failures such as two inverted reward design errors, token invalidation under concurrency, and tool calls degrading into plain text, and recommends evaluating under the decoding settings used in final serving.

Databricks vs. Oracle 2026 Comparison: Why Open Lakehouse Architecture Replaces Traditional Data Warehouses

In its 2026 comparison article, Databricks argues that its open lakehouse architecture stores 100% of customer data with Delta Lake and Apache Iceberg, unifying BI, ETL, stream processing, ML and AI, whereas Oracle uses proprietary formats and relies on PL/SQL and Exadata. Databricks separates storage and compute; SQL Serverless warehouses cost nothing when idle and are billed per second; Oracle bills by ECPU and autoscales up to 3x base capacity, making costs harder to predict. Databricks also offers Unity Catalog unified governance, Genie natural language analytics and OpenSharing cross-platform sharing, supporting AWS, Azure and Google Cloud.

PyTorch Conference 2026 held in San Jose, announces progress on TorchTPU and TorchNeuron

PyTorch Conference North America 2026 was held on October 20–21 in San Jose, attracting more than 3,500 attendees and focusing on accelerating open intelligence and projects including PyTorch, vLLM, DeepSpeed, Ray, Helion, and Safetensors. The event announced TorchTPU, a native PyTorch TPU backend developed in collaboration between Google and Meta, as well as TorchNeuron, a native backend for AWS Trainium, which entered public testing under Apache 2.0 on October 20. Shopify shared a case study of building a continuous learning loop with PyTorch and vLLM, saying it compresses production failures into model weights every day and has reduced serving costs by 96%.

Streaming low-latency speech for AI agents with Kokoro TTS on Amazon SageMaker AI

AWS demonstrated a solution for hosting the Kokoro-82M text-to-speech model on Amazon SageMaker AI that lets AI agents stream-synthesize speech immediately as each sentence is generated, so playback does not have to wait for the full response to finish generating. Kokoro-82M has only 82 million parameters and can run on a single ml.g5.xlarge GPU instance; the container provides two routes, POST /invocations and WS /invocations-bidirectional-stream. The article also explains how binary WebSocket frames are handled in bidirectional streaming, and precautions for keeping the session from closing.

How Amazon Quick Deploys AI in HIPAA-Compliant Environments

Amazon Quick is a HIPAA-eligible service covering web, desktop, and mobile clients, and can be used to process PHI after signing an AWS BAA. Its chat agent can retrieve across unstructured data such as clinical notes and structured data such as claims and member records, and execute multi-step tasks through connectors including Slack, Microsoft Outlook, Salesforce, Jira, and SharePoint; it also supports connecting custom tools via MCP. HIPAA-eligible does not equal HIPAA compliant—users still need to complete data classification, configuration, governance, and administrative safeguards themselves.

How IskayPet unified enterprise knowledge with Amazon Bedrock Managed Knowledge Bases

IskayPet built an enterprise knowledge assistant on Amazon Bedrock Managed Knowledge Bases, covering retail operations, veterinary care, HR onboarding, and procurement, supporting four languages across three countries without self-managing RAG infrastructure. The assistant ingests existing documents directly through the native SharePoint connector, combines hybrid search of semantic and BM25 keyword search, managed reranking, and document-level permission controls, and returns answers with source citations. The entire solution, from the store frontend to the AWS architecture deployment, was completed by Kiro with Claude Haiku 4.5.

Amazon Polly generative engine major upgrade: adds 14 new voices and supports Hindi and Finnish for the first time

The Amazon Polly generative engine has completed an upgrade, adding 14 new voices across 9 locales and, for the first time, officially offering Hindi and Finnish in the generative engine. After the upgrade, it offers a total of 57 generative voices, covering 11 languages and 25 locales, and has expanded to AWS GovCloud (US-West), for a total of 11 AWS Regions. Existing voice names and API calls remain unchanged; this is an in-place upgrade with no price change and no code changes required.

How Rescale Built an AI Triage Agent with Strands Agents and AI-DLC

Rescale built an AI triage agent based on the AWS AI-DLC framework and Strands Agents, going from problem definition to pre-release deployment in just five days and reaching production within weeks. The agent runs in an Argo Workflows CI/CD pipeline on Amazon EKS, calls Claude Opus 4.8 on Amazon Bedrock to classify nightly Playwright test failures as infrastructure issues, code defects, flaky tests, or environment problems, and queries pod logs via MCP, combining them with a historical failure knowledge base to generate structured remediation reports.

No More Guesswork: A Measured Approach to TTS Autoscaling Targets on Amazon SageMaker AI

AWS deploys the 82M-parameter TTS model Kokoro-82M in a custom container on an Amazon SageMaker AI real-time endpoint, using measured Q_max (the highest concurrency a single instance can sustain within the SLO) and T_total (the end-to-end time of one scale-out) to derive the scale-out threshold, scale-in threshold, and cooldown period, replacing target values set by guesswork. The approach scales on the natively published SageMaker AI ConcurrentRequestsPerModel metric, avoiding the hidden failure of a custom metric that no one reports.

AAIF: An Adaptive Agent Framework Based on Strands Agents

AWS built the Adaptive Agent Intelligence Framework (AAIF) with Strands Agents, using llama3.2:1b as the orchestration layer to classify requests: routine tasks are handled by a local SLM via Ollama, while complex analysis is routed through boto3 to frontier models on Amazon Bedrock. This architecture can run on edge devices such as a Raspberry Pi with 8 GB of memory, or be deployed fully in the cloud or in a hybrid setup, thereby reducing API costs while maintaining low local latency.

From personal coding tools to development platforms: trust and cost in AI programming

Massimo Re Ferre, Senior Principal PM at AWS, believes that the key for teams moving from personal coding assistants to agent development platforms is not model capability, but whether agent output can be verified at low cost—trust rises with inspection capability, not with benchmarks. Agent costs are billed by task activity; about 80% of developers use only 20% of the budget, while 20% of heavy users consume the rest, so the focus should be on the cost per unit of effective work rather than seat price.

OpenAI and Google Reach AI Distribution Agreement; GPT-6 Astra and Codex to Launch on Google…

OpenAI and Google have reached an AI distribution agreement. Within weeks, OpenAI will list GPT-6 Astra and Codex on Google Cloud's Model Garden, extending its model coverage to all three major U.S. clouds. Customers can purchase these models through Google Cloud Marketplace and count the spending toward their committed cloud consumption quotas, eliminating a separate procurement process; regulated customers can also run agent workflows built with OpenAI models within Google Cloud's access control and data governance tools. The agreement only became possible after April 27, 2026, when Microsoft's license to OpenAI technology became non-exclusive and Microsoft stopped paying OpenAI a revenue share on Azure resales; AWS had previously been first to launch Codex on Bedrock.

WSJ sees Anthropic IPO draft: half-year revenue $16.3 billion, compute commitments about $450 billion

The Anthropic IPO draft seen by WSJ shows half-year revenue of $16.3 billion, a book loss of $80.7 billion, and compute commitments of about $450 billion. The draft discloses fixed payments of $252 billion to three major cloud providers over the next 7 to 10 years, including $111 billion to Google, $110 billion to AWS, and $31 billion to Microsoft, payable whether or not used. There is also over $160 billion in Google chip leasing (via Broadcom) and up to $86 billion in Nvidia GPU rental, plus $150 billion in deals under negotiation.

English translation is being prepared.

English translation is being prepared.

English translation is being prepared.