OpenAI told investors in late September that its annualized revenue was close to $50 billion, below the roughly $70 billion that had circulated in the market, according to the Financial Times. The gap does not mean OpenAI actually earned $20 billion less; it reflects investors having tried to estimate both companies using Anthropic's method, which pushed OpenAI's revenue base higher. The core difference is how cloud-partner revenue is recognized: Anthropic counts sales through partners such as AWS and Google Cloud at the full amount, while OpenAI does not include some partner sales the same way. Both companies comply with US generally accepted accounting principles, but judge differently who owns the customer relationship and who delivers the product; Anthropic reported an annualized revenue run rate of about $65 billion at the end of July.
SailPoint CEO and founder Mark McClain, speaking to Hong Kong and Taiwan media during a late-September visit to Hong Kong, said enterprises are moving toward hybrid workforces of humans and AI agents and need a unified control plane to manage identity risk across both. He said non-human identities in some companies already outnumber human identities by 100 times, and argued AI agents should get only the minimum access needed for their work. SailPoint released agentic-AI security offerings in May and October and acquired Israeli startup Entro in June; McClain also said AI development should not stop, with safety controls acting like brakes that let the technology go faster.
vLLM and the AWS Machine Learning Blog describe bi-directional KV cache transfer in vLLM: in prefill-decode disaggregated serving, a decode node can return previously computed KV blocks to the prefill node on later turns, avoiding recomputation of the conversation history. On a 1P1D deployment using Amazon EC2 P5en instances with AWS EFA, the authors report TTFT reductions of up to 1.6x for Qwen3-32B and up to 3.3x for zai-org/GLM-5-FP8, noting that results vary with workload and configuration. Using it requires vLLM v0.21.0 or later with NixlConnector, enabling bidirectional_kv_xfer on both prefill and decode servers, and a KV-cache-aware router, with decode-side block retention governed by decoder_kv_blocks_ttl (default 480 seconds).
The Wall Street Journal reported that Amazon is negotiating to acquire San Francisco AI startup Decart in a deal that could be valued at around $7 billion, a report relayed by Bloomberg, IT之家 and others. Founded in 2023, Decart sells a cross-chip optimization layer called DOS that lets AI workloads move across Nvidia GPUs, Amazon Trainium and Google TPU, and also builds the real-time video model Lucy and the interactive-environment model Oasis. Reports say Anthropic discussed a roughly $6 billion acquisition in August and walked away after due diligence, while Nvidia was also named as a possible suitor; Amazon and Decart did not comment, and TechRepublic noted CTech floated a possible $5.5 billion valuation.
An analysis published via 36Kr's Muhe Business Finance says Zhipu's GLM-5.3 went live on Amazon Bedrock during China's October National Day holiday under a usage-based revenue-share deal; Zhipu's Hong Kong shares reportedly rose more than 7% that day, and Goldman Sachs upgraded the stock to buy with a HK$1,560 target, raising its end-2026 ARR estimate for Zhipu from $2.7bn to $3.2bn. The piece contrasts this with DeepSeek, which has been available on Bedrock since March 11, 2025 via Bedrock and Custom Model Import but is described as prioritising distribution and open weights over direct monetisation, a difference it ties to Zhipu's post-IPO revenue pressure versus DeepSeek's High-Flyer backing and its reported RMB 51bn first external round in June 2026.
A long technical analysis walks through Cerebras's Wafer Scale Engine: instead of dicing a wafer, it turns a full 300mm wafer into a 21.5cm x 21.5cm chip with 84 dies, roughly 900,000 active cores, 4 trillion transistors and about 44GB of on-chip SRAM, with claimed on-chip bandwidth of 21PB/s. The piece argues that keeping weights in SRAM next to the compute units sidesteps the HBM memory wall that leaves GPUs idle while shuffling weights for every token. It also details reticle stitching across scribe lines, whole-wafer testing and fault-tolerant routing, and a four-layer “engine block” assembly of cold plate, wafer, flexible connector and PCB (23kW peak power, vertical power delivery, microchannel liquid cooling).
Synopsys said it has entered a multi-year strategic agreement with Amazon worth more than $1 billion, using Synopsys silicon IP and AI-driven engineering software to speed development of Amazon's custom chips and AWS infrastructure. The two will also optimize Synopsys software for Trainium and Graviton, while Synopsys engineering teams adopt Amazon EC2 and Amazon Bedrock to accelerate their own product development.
Citing two people familiar with the matter, Cailianshe reports that Mississippi utility Entergy warned Amazon earlier this year it might have to cut grid power to the company's Mississippi data center during peak demand to keep other customers supplied. Such warnings to large customers are rare for local utilities, and the previously unreported notice reflects how surging AI cloud power demand is pushing utilities toward unprecedented measures. The warning prompted AWS to rework its automatic power-switching system, which moves data centers from the public grid to backup sources such as gas generators.
An AWS Machine Learning Blog recap says Amazon Bedrock added several capabilities for AI builders in September: Amazon Bedrock Managed Agents powered by OpenAI is now in public preview, and the AgentCore runtime has lower cold-start latency and scales sessions to zero when idle. Strands introduced an open source agent harness that AWS says matches popular harnesses on accuracy while using 28 percent fewer tokens, plus Strands Decider 2B, an open source 2B-parameter decision model that returns answers locally in about 115ms. Bedrock also lists OpenAI's GPT-6 Astra, Sol and Luna, along with Claude Fable 5.1, Claude Opus 5.5, Claude Sonnet 5.5, Moonshot AI's Kimi K3 and xAI's Grok 4.6 and 4.7, and added native ServiceNow, Salesforce and Zendesk connectors to its managed knowledge base.
In an Amazon Machine Learning Blog post, Postman and AWS describe the production architecture behind Agent Mode, Postman's AI-native way to work across API testing, documentation, discovery and implementation. The post says the team expected model quality and prompt design to be the hardest problems, but deeper challenges came from integrating an agent into an 11-year-old product built around interface-driven assumptions. The described patterns include dynamically scoping tools per task, favoring schema-based reads over tool proliferation, building purpose-made context handlers per entity, and treating the context window as a scarce budget. Inference runs on Amazon Bedrock with cross-Region inference, multi-tier prompt caching and model-dependent data-retention settings. Agent Mode requires user approval before actions that change application state and uses Amazon Bedrock Guardrails to redact personally identifiable information. Postman says its production implementation is proprietary and not available as a public sample repository.
The piece frames an agent as "model + harness" and splits the harness into a development half (memory, tools and MCP, retrieval, prompt orchestration) and an operations half (observability, evals, guardrails, routing, cost control, deployment and scaling) — capabilities the model itself does not provide. It then builds the same finance agent, FinBot, two ways: a managed Amazon Bedrock AgentCore Harness configured via CreateHarness/InvokeHarness calls, with per-session Firecracker microVMs, managed memory and automatic tracing; and a self-managed LangChain agent on the reader's own Kubernetes cluster behind the open-source Agent Router (formerly Envoy AI Gateway). The trade-offs come down to where provider credentials live, who carries the operational burden, and whether you pay a vendor bill or infrastructure plus engineering time.
A report republished with authorization by TMTPost says Amazon eliminated more than 30,000 corporate roles between October 2025 and October 2026 — roughly 10% of its 350,000 white-collar staff and the largest layoff in its 30-year history — with over 70% of the cuts at L5–L7 middle-management levels. In the same period, the article says, Amazon announced 2026 capital expenditure of about $200 billion, mostly for AI infrastructure, roughly 1.5x the prior year's $131 billion and nearly 37% above Wall Street expectations. It also cites an estimated $3 billion in annual savings from the first 14,000-person round (per Cybernews) and says Amazon raised its total Anthropic investment from $8 billion to $33 billion in April 2026, with future plans up to $65 billion, in exchange for Anthropic's commitment to spend more than $100 billion on AWS over the next decade.
AWS Machine Learning Blog detailed how Incarna’s agents use Amazon Bedrock AgentCore payments to pay BlockRun per request for model inference. BlockRun is described as a pay-as-you-go inference router serving more than 90 models from over 15 providers over x402, while AgentCore payments handles wallet connections, transaction signing and infrastructure-enforced spending limits, with settlement in USDC on Base. The post says Incarna cut x402 integration work from months to days and put an end-to-end pay-per-inference flow into production; those results and capabilities are vendor claims without independent verification.
The open-source trycua/cua project has shipped Cua Spaces 0.1.0, a macOS app that gives agents full macOS or Linux desktops running on your Mac, other machines you own, or your own cloud account (AWS, Google Cloud or Modal), with shared human-plus-agent desktops, teleporting of signed-in apps, and sessions kept in a local encrypted Cua Keyvault. The repository also ships Cua Driver for driving native desktop apps and browsers across macOS, Windows and Linux via CLI, MCP or typed SDKs, Lume for local Apple Silicon VMs, the small specialized CUA-S1 'System 1' decision models whose weights are hosted on Hugging Face, and Cua Bench for building tasks, evaluating agents and exporting trajectories. Cua Spaces is source-available under FSL-1.1-MIT and becomes MIT two years after each release, while the rest of the repo is MIT; Spaces is free for individuals, with Pro and Teams plans listed as coming soon.
An analysis in TMTPost argues that the binding constraint on AI infrastructure is shifting from chips to time, land and electricity: citing a Reuters report from July, it says Meta may spend about $145 billion on AI infrastructure in 2026 and deploy roughly 14GW of capacity, while buying more CPUs to handle persistent agent workloads. It notes Blackstone walked away from the Digital Gateway data-center campus in Virginia after 22 months, with developer QTS withdrawing its state Supreme Court appeal; Reuters reported Oracle issued a force-majeure notice over power-related delays on the Jupiter project it runs with Blue Owl for OpenAI; and JLL data shows 57% of data-center projects slipped by more than three months in 2025. The piece also cites a newly announced US energy package of up to about $200 billion — including eight large nuclear plants, an Alaska LNG project and Texas gas generation — as evidence that new AI projects increasingly build their own power supply.
Amazon carried out a fresh round of job cuts starting October 6 across customer service, Marketplace support and retail engineering, with teams in India and the UK also affected; a person familiar said fewer than 1,000 employees were impacted. An Amazon spokesperson confirmed to Business Insider that a "small number" of roles were eliminated, mainly in the Stores business. The move extends a series of reductions since October 2025: about 14,000 corporate jobs cut that month, another 16,000 in January 2026 touching AWS units including Bedrock, AI cloud services and Redshift, and roughly 17,000 in January–September 2026 per Layoffs.fyi — more than 30,000 in under a year. At the same time, capital spending is hitting record levels, with reports putting 2026 capex at about $200 billion (one report says $220 billion), mostly for AI infrastructure, up sharply from $131 billion in 2025.
An AWS technical post describes how Amazon Quick and Amazon Bedrock Knowledge Bases enforce access control in retrieval-augmented generation. The two-stage design first filters candidates using ACLs stored in the vector index, then verifies each candidate at query time by calling the authoritative source—such as Google Drive APIs via service-account impersonation—so only authorised passages reach the LLM. AWS says this avoids stale or mis-mapped permissions from replicated ACL syncs, and cites Mondelēz International, which it says has deployed Amazon Quick for more than 35,000 employees across four regions.
In an AWS blog post, Qlik detailed the architecture behind Qlik Answers, its enterprise question-answering product: employees ask in natural language and get sourced answers drawn from knowledge bases, live analytics apps, glossary definitions or documents, across more than 40,000 customers and 11 Regions. The system is split into entry, routing, answer, specialist-agent, conversational-analytics, retrieval (Amazon OpenSearch Service) and model-access layers; model calls go through Qlik's own LLM gateway to Amazon Bedrock, with Bedrock Guardrails applying content filtering and contextual grounding validation to every request and response, and Amazon SageMaker AI hosting models not yet available in-Region. Qlik says that since February 2026 general availability its Discovery Agent has surfaced over 100,000 discoveries, that most Qlik Cloud accounts with agentic tools enabled keep using them, and that customer Lintech International saw 75% faster responses and up to 7 hours per week returned to managers — all figures reported by Qlik itself.
An AWS Machine Learning Blog post walks through automating fixes after an AWS DevOps Agent investigation. Amazon EventBridge picks up the investigation-completed event, triggers an AWS Lambda Durable Function, and Amazon Bedrock proposes remediation actions drawn from a curated allowlist of purpose-built Lambda tools; read-only tools run autonomously while infrastructure-changing actions suspend for human approval. The demo scenario increases an undersized Lambda function timeout.
AWS published a post on how organizations can use the ISO/IEC 42005:2025 standard to run AI system impact assessments and fold them into existing enterprise risk, privacy and security reviews. It covers the standard's guidance on when to assess, lightweight triage, required documentation and reassessment triggers, plus Annex D for integrating existing assessments without duplication and Annex E for a standalone template. AWS also points to its Well-Architected Responsible AI Lens and its ISO/IEC 42001 implementation guide, and says Amazon Bedrock, Amazon Q Business, Amazon Transcribe and Amazon Textract hold ISO/IEC 42001 certification.
An AWS Machine Learning Blog tutorial walks through deploying a voice travel concierge on Amazon Bedrock AgentCore: Amazon Nova 2.5 Sonic handles real-time speech-to-speech, Strands Agents defines the agent, and AgentCore Gateway exposes airline backend endpoints as MCP tools, with Amazon Bedrock Knowledge Bases answering policy questions via RAG. The post says Nova 2.5 Sonic improves tool calling and multi-step agentic task completion and supports bidirectional streaming, asynchronous tool calls, latency masking and barge-in. The solution deploys with AWS CDK and a React front end on Amplify.
Z.ai's GLM 5.3 is now available on Amazon Bedrock, according to an AWS blog post. The 753B-parameter mixture-of-experts open-weight model is optimized for coding and long-horizon agentic tasks, and Z.ai reports a leading CyberGym score of 84.5 for defensive security work. Bedrock offers managed APIs, cross-Region inference, prompt caching and service tiers, with access currently limited to eligible enterprise customers.
NVIDIA researchers posted NeMo-DCR (Delta-Compressed Refit), a method that synchronizes policy updates between training and rollout clusters by sending only weight changes while remaining bit-exact against a dense refit. The paper reports that about 1% of BF16 training weights change stored values per step, and that at 3% and 5% change rates refits of 30B–1T models run 12–40x faster than a transport-only full-checkpoint reference; a 1T relay-tree refit at 3% takes 150 seconds versus 87.5 minutes to move a full checkpoint between two AWS regions. The code is open-sourced in NVIDIA NeMo RL PR #2444.
An AWS Machine Learning Blog post demonstrates a RAG application that uses LangChain with Amazon Bedrock Managed Knowledge Bases, comparing two retrieval paths. The standard Retrieve API runs a single hybrid search, while the AgenticRetrieveStream API runs a planning loop that splits a multi-part question into sub-queries, judges whether the evidence is sufficient, and searches again if it is not; with generate_response enabled it returns a grounded answer with citations. The post also covers the IAM permissions, relative costs and appropriate use cases for each path.
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…