All news in category “Vendor and Hyperscaler Watch”

⚡ AWS announced that high-performance, storage-optimized Amazon EC2 I7i instances are now available in the Asia Pacific (Hyderabad) and Canada (Central) regions. Powered by 5th-gen Intel Xeon Scalable CPUs and 3rd-gen AWS Nitro SSDs, I7i delivers up to 23% better compute and substantial NVMe storage improvements over I4i. Instances support torn-write prevention, real-time NVMe performance statistics, and sizes up to 48xlarge plus bare metal options.

🔒 Amazon Keyspaces (for Apache Cassandra) now supports Logged Batches, enabling multiple INSERT, UPDATE, and DELETE operations to be executed as a single atomic transaction. This ensures that all writes in a batch succeed or none are applied, improving consistency across rows and tables for use cases such as finance, inventory, and multi-entity profile updates. The feature preserves Cassandra's atomicity guarantees, integrates with CQL, scales serverlessly with your workload, and is available today in all AWS Commercial and AWS GovCloud (US) Regions. Customers pay only for the standard write operations processed within each batch.

🚀 Amazon EC2 M8a instances are now available in US East (N. Virginia) and Asia Pacific (Tokyo). Powered by 5th Gen AMD EPYC processors (code-named Turin) with up to 4.5 GHz, M8a delivers up to 30% higher performance, up to 19% better price-performance versus M7a, and 45% more memory bandwidth. They show workload gains up to 60% for GroovyJVM and 39% for Cassandra, are SAP-certified, come in 12 sizes including two bare-metal options, and run on sixth-generation AWS Nitro Cards. Customers can purchase M8a via Savings Plans, On‑Demand, or Spot.

🤝 Google Cloud is increasing local AI compute in India with its AI Hypercomputer powered by Trillium TPUs, enabling training and serving of advanced Gemini models with data residency and sovereignty controls. New local offerings include batch support for Gemini 2.5 Flash, a preview of Document AI, and real‑time grounding using Google Maps for location‑aware responses. Google is also supporting Indic Arena at IIT Madras with cloud credits to benchmark Indian multilingual models and to help grow the local AI ecosystem.

🔒 Mozilla has rolled out enhanced anti-fingerprinting protections in Firefox 145, initially active in Private Browsing and Enhanced Tracking Protection (ETP) Strict mode. Phase 2 measures add targeted noise to background image reads, restrict reported fonts to standard OS sets with select language exceptions, coarsen touch reporting, report screen height minus 48 pixels, and always report two processor cores. After testing these changes will be enabled by default; users can disable them per-site for compatibility. The release also removes the 32-bit Linux build.

🔒 AWS Backup now supports Amazon EKS, providing a fully managed, centralized solution for backing up cluster state and persistent application data. The agent-free integration replaces custom scripts and third-party tools with a native, policy-driven service that offers automated scheduling, retention management, immutable vaults, and cross-Region and cross-account copies. You can restore entire clusters, specific namespaces, or individual persistent volumes to support disaster recovery, compliance, or pre-upgrade protection.

🔒 AWS announced the 2025 H1 IRAP report is now available on AWS Artifact for Australian customers. An ASD-certified IRAP assessor completed the evaluation in September 2025, and four services were newly assessed at the PROTECTED level: Amazon Application Recovery Controller, AWS Global Accelerator, Amazon Q Business, and AWS Resource Explorer. AWS also published an IRAP documentation pack aligned to ACSC guidance and the ISM (March 2025) to help customers assess and architect PROTECTED workloads. Customers can request inclusion of additional services via their AWS representatives.

⚡ Effective today, all new Amazon MSK Provisioned clusters with Express brokers support Intelligent Rebalancing at no additional cost. The feature automates partition balancing when clusters scale up or down, maximizing capacity utilization and removing the need for manual or third-party partition management. AWS reports Intelligent Rebalancing runs up to 180× faster than Standard brokers and scales brokers without impacting client availability.

🚀 Zeotap migrated its Customer Data Platform from ScyllaDB to Bigtable to address scaling challenges, operational overhead, and highly spiky workloads. The cloud-native stack—using Dataflow, a home-grown streaming engine, Memorystore as a cache, Bigtable as the hot store, and BigQuery for analytics—delivers predictable low-latency reads and writes at scale. The transition yielded a 46% reduction in TCO and a ~20% drop in operational tasks while enabling sub-second SLAs and faster ML deployment.

🔐 Microsoft’s November 2025 progress report on the Secure Future Initiative outlines governance expansion, engineering milestones, and product hardening across Azure, Microsoft 365, Windows, Surface, and Microsoft Security. The update highlights measurable gains — a nine-point rise in security sentiment, 95% employee completion of AI-attack training, 99.6% phishing-resistant MFA enforcement, and 99.5% live-secrets detection and remediation. It also introduces AI-first security capabilities, new detections, and 10 actionable SFI patterns to help customers improve posture.

🧠 Gemini Code Assist on GitHub now supports persistent memory that learns from merged pull request interactions to capture a team's coding standards, style, and best practices. The memory is stored securely in a Google-managed project specific to each installation and is applied selectively to relevant reviews. It infers reusable rules from review threads and uses them both to shape initial analysis and to filter draft suggestions so the agent adapts over time and reduces repetitive feedback.

🚀 Google Cloud describes a full-stack solution for running high-scale Reinforcement Learning (RL) with LLMs, combining custom TPU hardware, NVIDIA GPUs, and optimized software libraries. The approach addresses RL's hybrid demands—reducing sampler latency, easing memory contention across actor/critic/reward models, and accelerating weight copying—by co-designing hardware, storage (Managed Lustre, Cloud Storage), and orchestration on GKE. The blog emphasizes open-source contributions (vLLM, llm-d, MaxText, Tunix) and integrations with Ray and NeMo RL recipes to improve portability and developer productivity. It also highlights mega-scale orchestration and multi-cluster strategies to run production RL jobs at tens of thousands of nodes.

🚀 Google Cloud announces general availability of the N4D machine series built on 5th Gen AMD EPYC 'Turin' processors and Google's Titanium infrastructure. N4D targets cost-optimized, general-purpose workloads — web and app servers, data analytics, and containerized microservices — with up to 96 vCPUs, 768 GB DDR5, 50 Gbps networking, and Hyperdisk storage. Google cites up to 3.5x web-serving throughput versus N2D and material price-performance gains for general compute and Java workloads.

🔒 Small and medium-size businesses face rising, measurable cyber risk as ransomware incidents increase and attacks spike during the holiday season. Resource constraints and end-of-life Windows 10 devices magnify exposure, while firmware-level and endpoint gaps can defeat traditional defenses. A layered, defense-in-depth approach across silicon, the operating system, and endpoints reduces attack surfaces. Business-grade devices such as the ASUS Expert Series integrate these protections to turn necessary upgrades into strategic security investments.

🌐 Amazon now supports Internet Protocol version 6 (IPv6) addresses for S3 Express One Zone gateway VPC endpoints, enabling access over IPv6 or DualStack without additional translation infrastructure. This applies in all Regions where the storage class exists at no extra cost. You can enable IPv6 for new or existing endpoints via Console, CLI, SDK, or CloudFormation. See the S3 User Guide to get started.

🔬 Amazon Braket notebook instances now include native support for CUDA-Q, enabled by upgrading the underlying OS to Amazon Linux 2023 to deliver improved performance, security, and compatibility for quantum development and production-ready workflows. Developers can run GPU-accelerated quantum circuit simulation alongside access to QPUs from IonQ, Rigetti, and IQM within the managed notebook environment. This eliminates the need for local deployment or separate Hybrid Jobs, streamlining hybrid quantum-classical experimentation. CUDA-Q support is available in all Regions where Braket operates.

🚀 Anthropic's Claude Sonnet 4.5 is now available in Amazon Bedrock within AWS GovCloud (US‑West and US‑East) via US‑GOV Cross‑Region Inference. The model emphasizes advanced instruction following, superior code generation and refactoring judgment, and is optimized for long‑horizon agents and high‑volume workloads. Bedrock adds an automatic context editor and a new external memory tool so Claude can clear stale tool-call context and store information outside the context window, improving accuracy and performance for security, financial services, and enterprise automation use cases.

🔁 AWS Control Tower customers can now automatically enroll member accounts simply by moving them into an Organizational Unit (OU). When moved, Control Tower applies the destination OU's baseline configurations and controls and removes the originals from the prior OU, removing the need for manual re-registration. This streamlines provisioning by allowing accounts to be created and then placed in the correct OU using the AWS Organizations console or the CreateAccount and MoveAccount APIs. Customers on landing zone version 3.1+ can opt in by toggling the automatically enroll accounts flag or by setting RemediationTypes to Inheritance_Drift in the CreateLandingZone or UpdateLandingZone APIs.

🐍 Cloudflare has announced Python Workflows in beta, enabling developers to orchestrate multi-step, durable applications on Workers using Python. The feature aims for feature parity with the existing JavaScript SDK while adapting APIs to Pythonic idioms—using decorators for step callbacks and snake_case naming for method calls. Under the hood it leverages Pyodide and CPython in the runtime, exposes WorkflowStep as an RPC-backed JsProxy for at-most-once durable execution, and supports DAG-style concurrency via asyncio.gather. Targeted use cases include data pipelines, ML/LLM training loops, and autonomous agents where step-level retries, state persistence, and explicit wait points simplify orchestration.

🔒 Google Public Sector announced it has achieved CMMC Level 2 certification, validated by a certified third-party assessment organization (C3PAO). The certification confirms that its internal systems used to process and store Controlled Unclassified Information (CUI) meet DoD cybersecurity expectations. While the certification covers Google’s internal systems and does not extend to customer environments, Google highlights support for the Defense Industrial Base through FedRAMP-authorized cloud services and published compliance resources, including a Google Workspace CMMC Implementation Guide, to help partners accelerate their own CMMC journeys.