Smartsheet has launched a generally available Model Context Protocol (MCP) Server that lets AI assistants read, write, and manage Smartsheet data through a standardized interface, according to the AWS Machine Learning Blog. Smartsheet says the server provides more than 40 tools for sheets, rows, columns, and automations, including natural-language task updates and row creation. The report also describes integrations with Amazon Quick and Kiro for business workflows and developer access to project context.
An AWS Machine Learning Blog post describes how Octave (formerly Hexagon SIG) built a secure multi-tenant AI foundation for its Reliance quality management platform on Amazon Bedrock, using Cohere Embed v4 for multilingual and multimodal search and Amazon Bedrock AgentCore to run each user session in a dedicated, sanitized microVM. Octave says the foundation embedded 2 billion documents and that its first feature, a Complaint Triage Assistant, has run in production for three months. By Octave's own accounting, the shared layers cut deployment time per use case by 50 percent, while a Cohere social post repeats claims of premium accuracy and lower vector storage costs without independent measurement.
AWS Machine Learning Blog presents a reference architecture that combines Keyfactor EJBCA with AWS services to authenticate AI agents, verify user delegation and enforce revocation across MCP tool calls. Its example uses Amazon Bedrock, the AWS Strands Agents SDK and FastMCP, with signed request envelopes, RFC 8693 audience-scoped tokens, OCSP checks and identity cross-validation. The post says it also provides complete example source code.
Amazon said it will invest more than €16 billion (about $18 billion) in Germany in 2027, a plan announced at the Invest in Germany Summit in Berlin, according to The Next Web, Bloomberg and cnBeta. The largest share, €4.8 billion, goes to AWS to expand its Frankfurt data-center region and related AI infrastructure. Amazon is also opening a second robotics and AI research lab in Berlin, doubling its research presence there, without disclosing the facility's cost.
AWS Security Agent successfully exploited 33 of 37 CVEs and retrieved verification flags in an isolated, CTF-style test, an 89.2% success rate reported by iThome. The test used Claude Opus 5 with a 10-hour limit per vulnerability; researchers excluded two initial successes caused by test-environment flaws and noted that two other attempts disrupted applications or services before obtaining a flag.
AWS released TorchNeuron on October 19, an open-source tool built on a native PyTorch backend that lets developers move PyTorch AI code from Nvidia GPUs to its own Trainium2 and Trainium3 chips by changing a single configuration line, according to 钛媒体. The tool is released under Apache 2.0 and entered public beta on October 20 at the PyTorch conference in San Jose; the report says the first version does not yet support features such as dynamic input shapes and skips roughly 22,000 PyTorch tests, while Nvidia still holds an 80% to 85% revenue share of the data center AI accelerator market. The report also says Amazon disclosed that Trainium and Graviton combined have surpassed a $25 billion annualized revenue run rate, and that Anthropic has trained and run Claude on more than one million Trainium2 chips.
AWS announced that new Amazon SageMaker AI customers can now sign up with an existing Google, GitHub, Apple or Amazon account, or an email address, in two to three steps, and then land in a fully configured SageMaker Studio environment created automatically in about 20 seconds. AWS says most customers no longer need the advanced AWS account sign-up, phone verification or a credit card, and that permissions for model customization, training, evaluation and deployment are pre-configured, with eligible accounts getting default access to G5, G6 and G4dn GPU instances. New customers get $100 in AWS Free Tier credits usable across SageMaker AI, including Studio notebooks, training jobs, model customization and endpoint deployments, with a chance to earn another $100; the Free Tier ends after six months or when credits run out, at which point usage stops automatically. The new sign-in experience applies only to new AWS accounts; existing accounts see no change.
SandboxAQ on Monday launched Switch Trust and Switch ROI, two tools that are part of its Switch AI platform, which debuted in August. According to SandboxAQ, Switch Trust inventories the AI agents, models and tools running in a company's systems — including ones its security team may not know about — tests them against attacks such as prompt injection, and monitors live input and output, blocking or hiding leaked passwords and personal data while logging each agent's actions; Switch ROI links each agent's spending to outcomes such as support tickets closed or tasks automated and suggests savings. The company says both products work with agents from any vendor and are free for up to five managed agents, with no paid prices listed; Mount Sinai West's MSW Lab is using Switch Trust while its doctors build medical software with AI agents.
Amazon announced TorchNeuron, an open-source native PyTorch backend that lets PyTorch code run on its Trainium AI chips by changing a single device setting rather than rewriting the application, according to Amazon. The tool entered public beta on October 20 at the PyTorch Conference in San Jose under the Apache 2.0 licence, supporting Trainium2 and Trainium3. Amazon says the first release still has gaps — some PyTorch features such as dynamic input shapes are unsupported and around 22,000 PyTorch tests are skipped, per The Next Web's account of the PyTorch blog post — and AWS says it is working with the PyTorch community to remove code that ties the framework to specific chips.
Research firm SemiAnalysis published a report forecasting more than 50GW of global data center capacity additions in 2027, up from 15.6GW in 2025 and roughly 25GW in 2026. It says its Datacenter Industry Model shows over 24GW across 40 campuses in advanced construction, most more than halfway done with exterior cooling and electrical equipment installed, all slated for 2027 delivery. The report acknowledges constraints in power, labor, supply chains and capital, but argues behind-the-meter generation, prefabricated designs and pre-leasing are easing them; TechNews, citing the report, said the next bottleneck could shift back to chips and memory. Techmeme's summary cites about $1.3 trillion of lease contracts signed year-to-date, while TechNews gives roughly $1 trillion.
In a blog post, venture investor Tomasz Tunguz argues that Anthropic should be assessed as a cloud business rather than only an AI model company, comparing its Q2 2026 revenue of $12 billion with AWS at $42 billion, Azure at $39 billion and Google Cloud at $25 billion. He cites 1,374% year-over-year growth, a $46 billion annual run rate and $518 billion in data-center commitments, while noting that two customers accounted for nearly 25% of 2025 revenue, per Reuters reporting from an unfiled prospectus. He also cites Ramp vendor data showing 87% of businesses buying a coding agent use Claude Code, and Artificial Analysis figures putting Claude Opus 5.5 at $5.98 per task versus $3.26 for GPT-6 Astra and $1.99 for Gemini 4 Argon.
AWS chief financial officer John Felton said spending on AI infrastructure and services is sustainable because it is driving meaningful innovation, according to an Axios report from October 16 cited by IT之家 on October 18. Felton said AWS had nearly $500 billion in backlog as of the second quarter and that its AI server investments have a payback period of under three years; AWS reported quarterly revenue of $42.2 billion, up 36.7% year over year, while its AI business's annualized revenue run rate exceeded $25 billion with triple-digit year-over-year growth. DoNews also published an automatically generated summary repeating the same remarks. The reports noted that some investors and analysts worry about an "AI bubble" amid rising debt, high valuations and circular financing deals.
An AWS Machine Learning Blog post describes a continuous quality-evaluation loop built with Amazon Bedrock AgentCore for an internal sales assistant that answers business questions in plain English through Amazon Quick, with Performance Insights, Customer Insights and Data Insights sub-agents working behind it. Before launch, evaluation found the Customer Insights Agent reported $410.08 of revenue as "$410.08 Million," offered an annual recurring revenue estimate that appeared nowhere in the data, and presented its own calculations as sourced facts, while monitoring stayed green. The framework runs from a golden dataset and a sentinel of 15 cases to a quality gate that replays sandboxed, frozen tool responses and scores 200 held-out questions end-to-end with an LLM-as-judge rubric; the post says one mismatched coordinated release dropped the pass rate from 78 percent to 38 percent.
The UK Competition and Markets Authority on October 16 proposed giving Microsoft a Strategic Market Status designation in business software and opened a consultation, with feedback due November 13 and a final decision expected by February 2027. The CMA said Microsoft holds a strong position in all five areas it examined — PC operating systems, productivity software, security software, server operating systems and relational database management — and that switching barriers and bundling practices may limit rivals. It will further examine whether customers can freely choose AI tools and whether competitors can take on Copilot, along with security-software bundling and how licensing affects cloud choice; Microsoft said all its products, including Copilot, are open and interoperable.
Amazon CTO Werner Vogels told a media roundtable in Shenzhen on 14 October that companies should educate themselves about AI before adopting it, saying "fear is a very bad way to make decisions" and warning against FOMO-driven rollouts, according to 钛媒体 and InfoQ 中文. He said Amazon still hires the same number of junior engineers as before, and described an internal language called Dogwood that specifies what an agent may and may not do, enforced by a Dogwood engine for Strands Agents users; he also said the Strands Decider decision model is fully open source, including weights and input datasets. Vogels questioned benchmark rankings — "sometimes a model that is good enough is enough" — and praised Kimi's underlying Mooncake architecture as impressive, while 钛媒体 reported Amazon plans about $200 billion in 2026 capital spending, mostly on AI infrastructure, with AWS second-quarter revenue up 37% year over year.
AWS, Anthropic, OpenAI and Google have all recently announced expansions of their forward-deployed engineer (FDE) programs, sending specialists to help customers realize AI benefits, according to CNBC and TechRadar. TechRadar reported that OpenAI is backing its FDE business with a $4 billion investment and AWS is footing a $1 billion bill, while Anthropic plans to train 10,000 specialist engineers by the end of 2027. Citing Lightcast research, it said US FDE job ads rose from about 200 in each of 2023 and 2024 to roughly 1,200 in 2025 and 7,300 in the first nine months of 2026 — a 6x year-over-year rise — with three of the top four advertisers being consultancies Deloitte, Accenture and KPMG, and Google the only tech firm.
North.cloud, a startup that helps companies track their cloud, AI and data spend, closed a $60 million Series B led by PeakSpan Capital, taking total funding to $67 million, CEO Matt Biringer told Axios Pro. Founded in 2023, the company says it manages close to $2 billion in cloud spend and has helped customers save more than $400 million cumulatively — figures it disclosed itself and that a 36Kr commentary notes were not independently audited. That commentary also reports that North is extending from cloud cost optimization into AI and data platform spend management, but that its TokenFlow page for tracking cross-model AI costs still showed a waitlist and testing-stage status when checked.
AWS published the Agent Toolkit for AWS on GitHub, an officially supported set of MCP servers, skills and plugins meant to help AI coding agents build, deploy and manage applications on AWS. The toolkit can be configured from the terminal through the AWS CLI, and has separate install paths for Claude Code, Codex, Cursor, Kiro and fx; the Claude Code plugins are listed on Anthropic's official plugin marketplace. Its AWS MCP Server is described as covering 300-plus AWS services through one authenticated endpoint, with sandboxed script execution, real-time documentation search, and enterprise controls including CloudWatch metrics, IAM condition keys and CloudTrail audit logging. Those capability descriptions come from the project README and are AWS's own account.
The AWS Machine Learning Blog published a walkthrough for deploying a supervisor agent on the Amazon Bedrock AgentCore runtime that uses the Strands Agents framework to route requests between presentation-generation and summarization tools. The setup uses Anthropic Claude Sonnet 4.6 for content creation, Amazon Titan Text Embeddings v2 for vector search and Stability AI Stable Image Inpaint for slide images, backed by an Amazon OpenSearch Serverless retrieval-augmented generation pipeline; a public code repository is included. AWS says the workflow completed a presentation in under 90 seconds in its internal testing, a figure that has not been independently verified.
36Kr reports that startup Reactor is building a real-time inference platform for video and world models, saying developers can use a unified SDK and API to build interactive applications with more than 60 FPS and under 40 milliseconds of latency. Reactor had raised a $59 million seed-plus-Series-A round and recently added follow-on Series A funding from NVIDIA and Sapphire Ventures, bringing its total raised to $74 million. The report also says Reactor worked with an AWS team to run real-time video generation on a single Trainium2 chip, though the performance and deployment claims were not independently measured.
Posts on the vLLM Blog and the AWS Machine Learning Blog describe a push-based NIXL connector for prefill-decode (P-D) disaggregated inference in vLLM. In the default pull mode, the decode instance allocates destination blocks and issues an NIXL READ only after prefill finishes, so coordination and setup sit on the time-to-first-token critical path; in push mode the decode side registers blocks in advance and the prefill worker hands the transfer to a background thread that issues an NIXL WRITE. The authors say the feature was built and validated on AWS Trainium clusters with EFA and then contributed to vLLM as an accelerator-agnostic feature.
Bloomberg Intelligence forecasts that games streamed over the internet will grow into a $38 billion market by 2035 and that the dedicated video-game console may eventually stop being a mass-market product. The research, led by BI analyst Nathan Naidu and reported by Bloomberg on Tuesday, argues AI both cuts streaming costs and supplies the data centers that run it. But the case depends on surplus GPU capacity pushing compute prices down late this decade, which has not happened: The Next Web notes AWS raised GPU prices about 20% in July for reserved capacity, its second rise in six months. BI estimates Microsoft's Xbox Cloud Gaming holds roughly 40% of streaming with about 220 million hours last quarter, sees Nvidia's GeForce Now as technically advantaged and Tencent leading in China, puts Nintendo in the worst position for treating streaming as peripheral, and says Sony is better placed via PlayStation Plus and PlayStation Portal.
An AWS Machine Learning Blog post describes how Anchor Browser works with Amazon Bedrock AgentCore payments: when an agent's managed browser session hits an x402-protected page or API returning HTTP 402 Payment Required, AgentCore payments checks the session budget, signs the transaction through the configured wallet provider and lets the browser retry the request once, so the task continues without human intervention. The post says Anchor Browser's Agentic Payments was developed with the AWS Generative AI Innovation Center Partner Agent Factory and is now available on AWS Marketplace. It also says wallet secrets, spending limits and CloudWatch audit trails sit in the AgentCore service layer rather than in the agent runtime, with wallet connectivity via Coinbase CDP or Stripe Privy.
An AWS Machine Learning Blog post describes Discovery AI, a natural-language query agent that Orca Security built with AWS on Amazon Bedrock to turn plain-language questions into filters for its Sonar query system. Sonar spans more than 900 cloud infrastructure and application models, 18 operators and nested parent-child relationships, which previously required users to hand-write JSON queries. Orca reports that compared with its initial prototype, which loaded the full schema into one prompt, input tokens per workload fell 93 percent from roughly 2 million to 140,000, cost fell 90 percent and end-to-end latency fell 75 percent from about 60 seconds to 15 seconds. The pipeline pairs retrieval-augmented example retrieval with schema-aware generation and splits models by stage: Anthropic Claude Haiku 4.5 for model selection and query generation, Amazon Nova 2 Lite for guardrails and validation, Amazon Titan for embeddings and Cohere Rerank 3.5 for reranking. Orca says Discovery AI has handled thousands of queries per day in production since March 2026.
SpaceXAI's Grok 4.7 is now available on Amazon Bedrock with configurable reasoning effort, tool calling and structured output, according to the AWS Machine Learning Blog. SpaceXAI says the model uses a larger base model than Grok 4.6 and a longer reinforcement-learning run on a task mix weighted toward multi-hour problems, citing gains such as Terminal-Bench 4.0 rising from 20.3% to 37.6%, plus a 500K-token context window. AWS built a tool-calling agent on three search-requiring tasks with one verifiable answer each and ran it six times per effort level for 72 runs: 67 were correct, with all 18 xhigh runs correct and 15 of 18 at low effort, while mean output tokens were about 1,184 at low versus 9,147 at xhigh.
An AWS Machine Learning Blog post describes several kernel optimizations AWS implemented for fused Mixture-of-Experts models running on GPUs. It says a swap-AB matrix transposition aligned with NVIDIA Hopper's WGMMA instruction layout produced a 2–4 percent end-to-end throughput improvement for large MoE models on H200, and was upstreamed to vLLM 0.24.0 or later. A second change added do_not_specialize annotations to eliminate repeated Triton JIT recompilation, which AWS says improved time to first token by 9 percent at 8-way tensor parallelism and is available in vLLM 0.22.0 or later. AWS says the optimizations are open sourced, applied in the LMI deep learning container on Amazon SageMaker AI, and together delivered up to 10 percent higher throughput and up to 9 percent faster TTFT; these figures are vendor-reported and not independently verified.
At its late-September Pharma Day in London, Roche’s Genentech gRED head Aviv Regev presented a “Lab-in-the-Loop” system that trains models on experimental and clinical data and feeds predictions back into the next round of experiments. Roche says 40% of pipeline decisions from Q4 2025 through Q2 2026 had traceable AI or computational contributions, rising to 44% in Q2 2026, with 41% of those contributions classed as “critical”; its TargetNexus target-assessment agent is meant to take part in 80% of research portfolio decisions by Q4. The report notes Roche has not disclosed how it judges AI’s contribution, that FourWeekMBA’s rough calculation puts critically AI-influenced decisions at about 16%, and that the rise in Phase III success from 65% to above 80% is not split between AI and Roche’s stricter project-review framework.
AWS Machine Learning Blog published a tutorial on calling OpenAI's GPT-6 Sol and GPT-6 Luna through Amazon Bedrock global cross-Region inference from five Asia Pacific source Regions, plus GPT-6 Astra from Singapore. It says all three accept text and image input, client-side tool use and prompt caching, while differing in available reasoning-effort levels. The post also covers the Converse API, OpenAI-compatible APIs and coding clients, along with IAM and SCP permissions, quota management and CloudWatch/CloudTrail monitoring.
An AWS Machine Learning Blog post walks through deploying one large language model across two GPU nodes on Amazon EKS using the multi-node Ray LLM Deep Learning Container, with separate prefill and decode nodes behind a single OpenAI-compatible endpoint. The container ships Ray Serve LLM, NVIDIA's NIXL library and the Elastic Fabric Adapter stack, and the example uses Qwen3.5-9B to show prefill/decode disaggregation configured entirely through YAML with no application code. The post also warns that NIXL silently falls back to a TCP-speed transport when it cannot use libfabric, so it recommends checking worker logs to confirm the backend.
Citing The Information (relayed by Stocktwits and Chosun Ilbo), reports say the mystery AI cloud customer IREN Limited teased on an earnings call is Anthropic, which has signed a multi-year cloud contract that has not been announced; neither company confirmed it, and the value and term were not disclosed. IREN shares jumped in Asian trading on Monday, rising as much as 8.5% to $38.17, after closing at $35.19 on Oct 9, a low not seen since late July. The report also says Anthropic has signed data-center leases with more than 10 cloud providers including AWS and Google, totaling over 15GW and about $530 billion, with capacity projected to grow from 1.4GW in 2025 to 5GW and 9.5GW and to exceed 15GW in 2028; IREN CEO Daniel Roberts said in August that it had signed a top AI lab without naming the customer. The report notes that research firm SemiAnalysis last week questioned the quality of IREN's older British Columbia data centers, and that heavy lease costs could pressure Anthropic's margins.
What problem it solves
Developers using AI coding CLIs need a way to delegate work among multiple agents and coordinate their execution. CAO addresses this by running provider CLIs in isolated terminal sessions and giving a supervisor tools for coordinating workers.
Key features
Multi-agent delegation — A supervisor delegates work to specialist agents, with parallel and sequential execution suppor…
What problem it solves
AI-generated code can violate the rules its developers intended, especially when those rules are described in natural language and the resulting code is not reviewed closely. Bend addresses this by representing important rules as laws and requiring mathematical proofs when the code is edited.
Key features
Proof-enforced application laws — Applications can declare rules in `L…
What problem it solves
Agents can understand diagram content but often produce generic layouts that do not match a site's visual identity. Diagram Design addresses this by applying a defined editorial style to agent-generated diagrams and producing publishable, self-contained files without a manual design workflow.
Key features
Self-contained HTML and SVG output — Each diagram is written as one HT…
What problem it solves
Organizations need a distributed storage layer that can serve objects, files, tables, and vectors while keeping administration and horizontal expansion straightforward. SeaweedFS addresses this by providing these capabilities on shared storage and allowing capacity to grow by adding volume servers.
Key features
Unified distributed storage — A single cluster exposes S3 object…
What problem it solves
Modern systems need controlled access to database credentials, API keys, and other secrets across platform-specific infrastructure. Teams also need secure storage, key rotation, encryption, revocation, and detailed audit logs without building a custom solution.
Key features
Secure secret storage — Vault stores arbitrary key/value pairs and encrypts data before writing it to …
What problem it solves
VPN traffic can expose protocol characteristics that are examined during network inspection. The project addresses this by encrypting UDP tunnels, applying configurable traffic masks, and rotating session keys while providing bounded packet handling and live traffic-quality checks.
Key features
Encrypted and forward-secret sessions — The InitialV2 exchange confirms a fresh s…
What problem it solves
Leaked credentials can be difficult to identify and assess because they may be stored across repositories, chats, wikis, logs, filesystems, and other systems. Finding a credential pattern alone does not establish whether it is active, so the project combines discovery, classification, validation, and analysis.
Key features
Broad secret discovery — Scan sources including Git …
What problem it solves
Creating intelligent AI agents that can collaborate on complex tasks can require code and substantial integration work. Docker Agent addresses this by letting users define agents in YAML, assign tools, and run them through the Docker CLI.
Key features
Multi-agent orchestration — Create teams of specialized agents that delegate tasks automatically.
Source: README
Extensible t…
What problem it solves
AI coding agents need AWS-specific tools, knowledge, and guardrails to work effectively with AWS services. Teams using different coding agents also need a consistent way to configure AWS API access, guidance, and reusable skills.
Key features
AWS MCP Server — The managed MCP Server gives agents access to AWS APIs, isolated script execution, and current documentation through …
What problem it solves
Organizations need to build and maintain custom applications such as dashboards, admin panels, customer 360 tools, IT automations, and service management systems. This platform streamlines those application-development workflows while supporting databases and APIs.
Key features
Low-code custom application development — Build internal applications through a low-code platform …
What problem it solves
Ad-hoc AI coding can lose context as projects grow. AI-DLC keeps requirements, decisions, implementation, tests, and operational work connected through an audited lifecycle.
Key features
Harness-neutral workflow core — A single core runs across multiple coding-assistant harnesses, with thin harness-specific integrations.
Source: README
Structured delivery lifecycle — The wor…
What problem it solves
Teams need to capture, search, and analyze multiple observability signals without managing fragmented tools or complex infrastructure. OpenObserve addresses this with a unified platform, a single-binary deployment option, and familiar SQL and PromQL query languages.
Key features
Unified observability — Logs, metrics, traces, RUM, dashboards, alerts, incidents, pipelines, and…
What problem it solves
Developing against AWS services often requires a cloud account, authentication token, or paid feature gates. Floci provides AWS-shaped services locally so existing AWS SDKs, CLI commands, infrastructure-as-code workflows, and test suites can run without those dependencies.
Key features
Local AWS services without an account — Run AWS-compatible services locally without a clou…