All news in category “Vendor and Hyperscaler Watch”

🚀 Imgix serves over 8 billion images and videos daily and has migrated its real-time processing stack to G4 VMs on Google Cloud, powered by NVIDIA RTX PRO 6000 Blackwell GPUs. The move delivered a 50% reduction in median latency and a 5–6× increase in throughput per node without rewriting core application code. Imgix combines nvJPEG, NVENC/NVDEC, custom Vulkan compute shaders and CUDA libraries to accelerate decoding, transformation and encoding, while autoscaling, self-healing GPU management and a 2.5PB GCS cache enable fast, reliable global delivery.

🔒 AWS outlines its approach to AI sovereignty, emphasizing customer control over data, deployment location, and access across the AI stack. It highlights infrastructure choices—AWS AI Factories, Outposts, Local Zones, Dedicated Local Zones, and the AWS European Sovereign Cloud—to meet regulatory and operational needs. AWS emphasizes technical protections like the AWS Nitro System, identity controls (IAM and Amazon Bedrock AgentCore Identity), and certifications such as ISO/IEC 42001 to reinforce transparency and trust.

🔐 Idira is Palo Alto Networks' next-generation identity security platform, unveiled at IMPACT following the company's integration with CyberArk. It discovers every human, machine and AI agent, inventories entitlements across network, cloud, endpoints and browsers, and evaluates whether access is necessary. Idira replaces standing accounts with dynamic, just-in-time privileges and automates continuous governance, shrinking the fragmentation that delays incident response. The platform embeds AI to surface risky entitlements and drive rapid remediation, while integrating with Strata, Cortex and Prisma to enforce controls where users and agents work.

🔒 OpenAI has unveiled Daybreak, an enterprise-focused cyber-defense platform that combines its large language models with Codex-style agent capabilities and broad integrations across the security ecosystem. The initiative aims to accelerate vulnerability discovery, generate and test fixes within repositories, and deliver audit-ready evidence back into enterprise workflows. Daybreak will be offered in tiers including GPT-5.5, Trusted Access, and GPT-5.5-Cyber, and is being developed with major vendors and government partners.

🔧Cloudflare engineers describe a CUBIC congestion control bug in their open-source QUIC implementation, quiche, where the congestion window (cwnd) becomes permanently pinned at its minimum after an early congestion collapse. Test harnesses that injected 30% loss during the first two seconds revealed per-RTT oscillations and frequent timeouts despite loss stopping. The root cause was an idle-period epoch adjustment ported from the Linux kernel that could advance the recovery epoch into the future; a concise near-one-line change in quiche breaks the death spiral and restores recovery.

🔒 Apple and Google have initiated a beta rollout of end-to-end encrypted RCS messaging between iPhone and Android devices, closing a long-standing interoperability gap. The feature requires iOS 26.5 on supported iPhones and the latest Google Messages on Android, with carrier activation determining availability. Encryption is enabled by default, marked by a lock icon, and the rollout implements the GSMA Universal Profile 3.0 with MLS.

🔐 Apple released iOS 26.5, adding beta support to enable end-to-end encryption for RCS messages across iPhone and Android devices when used with supported carriers and the latest Google Messages. The feature is enabled by default for new and existing conversations and displays a lock icon to indicate encryption. Apple and GSMA say this is part of a cross‑industry effort to modernize SMS. The update also patches over 50 vulnerabilities in iOS and iPadOS.

🚀 Amazon now offers P5.48xl EC2 instances in SageMaker Studio notebooks across US West (San Francisco), Asia Pacific (Tokyo, Mumbai, Sydney, Jakarta), and Europe (London, Stockholm). These instances are powered by NVIDIA H100 Tensor Core GPUs and deliver up to 4x performance improvements and up to 40% lower training cost versus prior GPU generations. They are suited for training and serving complex LLMs, diffusion models, and other generative AI and HPC workloads. See the developer guides for setup with JupyterLab and CodeEditor and consult regional pricing for details.

🚀 Amazon Web Services announced general availability of Amazon EC2 G6 instances on SageMaker notebook instances in additional Asia Pacific (Tokyo, Mumbai, Sydney) and Europe (London, Paris, Frankfurt, Stockholm, Zurich) regions. G6 instances deliver up to 8 NVIDIA L4 Tensor Core GPUs with 24 GB per GPU and 3rd-generation AMD EPYC processors, offering roughly 2x inference performance over G4dn. Customers can use these instances for interactive model testing, generative AI fine-tuning, NLP, translation, computer vision, and recommender systems. Developer guides cover JupyterLab and CodeEditor setup on SageMaker Studio and notebook instances.

🚀 Amazon announces general availability of EC2 P6-B200 instances in AWS US East (N. Virginia) for use with SageMaker Studio notebooks. These instances feature eight NVIDIA Blackwell GPUs, 1440 GB of high-bandwidth GPU memory, and 5th Generation Intel Xeon (Emerald Rapids) processors, offering up to 2x training performance vs P5en. They enable interactive development and fine-tuning of large foundation models directly in JupyterLab or CodeEditor for generative AI workloads.

🚀 Amazon Web Services has announced general availability of Amazon EC2 P5.4xl instances for SageMaker Studio notebooks, powered by NVIDIA H100 Tensor Core GPUs. These instances offer up to 4x faster time-to-solution versus previous-generation GPU instances and claim up to 40% lower training cost for ML models. They are designed to accelerate training and deployment of demanding DL and HPC workloads, including large language models and diffusion models. P5.4xl is available now in select US, Asia Pacific, and South America regions, with developer guides and pricing details provided.

🚀 Amazon Web Services announced general availability of Amazon EC2 G6 instances for SageMaker Studio notebooks in the Middle East (Dubai) and Asia Pacific (Malaysia). G6 instances pair up to eight NVIDIA L4 Tensor Core GPUs (24 GB each) with third-generation AMD EPYC processors, delivering roughly 2× better deep-learning inference performance than G4dn. These instances support interactive model deployment and training for generative AI fine-tuning, NLP, vision, and recommender workloads. Refer to developer guides for JupyterLab and CodeEditor setup and the pricing page for cost details.

🚀 Amazon Web Services announced general availability of EC2 G6e instances on SageMaker Studio notebooks in Dubai, Tokyo, Seoul, Frankfurt, Stockholm and Spain. G6e instances provide up to 8 NVIDIA L40s Tensor Core GPUs with 48 GB per GPU and 3rd‑generation AMD EPYC processors, delivering up to 2.5× performance versus G5. They target interactive model testing, training and generative AI fine‑tuning, and can host LLMs up to 13B parameters as well as diffusion models for image, video and audio generation. Developer guides cover JupyterLab and CodeEditor setup; pricing is available on the AWS pricing page.

🚀 Amazon Web Services has announced the general availability of EC2 P4de instances on SageMaker Studio notebooks in Asia Pacific (Tokyo, Singapore) and Europe (Frankfurt). Each P4de packs eight NVIDIA A100 GPUs with 80GB HBM2e (640GB total), offering 2× the per‑GPU memory versus P4d. AWS reports up to 60% faster ML training and roughly 20% lower training cost compared to P4d, benefiting large high‑resolution datasets and reducing model training time. Developers can follow the SageMaker JupyterLab and CodeEditor guides and consult pricing for cost planning.

⚡ ENA Express now supports high-bandwidth traffic between Amazon EC2 instances in different Availability Zones within a Region, delivering up to 25 Gbps single-flow performance. The feature uses the AWS Scalable Reliable Datagram (SRD) protocol with multi-pathing and advanced congestion control to reduce head-of-line blocking. ENA Express establishes SRD connections automatically when both instances are enabled and supports TCP and UDP transparently. The capability is available at no additional cost across a broad set of Regions and instance types.

🚀 Amazon has expanded Aurora DSQL single-Region clusters to five additional AWS Regions: Hong Kong, Mumbai, Singapore, Stockholm, and Sao Paulo. The service offers serverless, distributed SQL with virtually unlimited scalability, high availability, and minimal infrastructure management. Aurora DSQL provides fast distributed reads and writes to simplify resilience and scaling for always-available applications. It is now available across 18 AWS Regions and is eligible for the AWS Free Tier.

🧬 AWS HealthOmics now supports caching completed task outputs when runs are cancelled, automatically storing those outputs in the customer’s S3 bucket. When caching is enabled, customers can restart runs from the point of cancellation and avoid recomputing tasks that already finished. This capability helps researchers, bioinformaticians, and workflow developers debug and iterate more efficiently. Caching is available for Nextflow, WDL, and CWL runs across all HealthOmics regions.

🔷 At Red Hat Summit 2026, Microsoft and Red Hat highlighted how Azure Red Hat OpenShift supports modernization and production AI by delivering consistent governance, security, and scale. Microsoft was named Platform Modernization Partner of the Year, underscoring joint customer outcomes. Banco Bradesco and Topicus illustrate production AI and regulated lending workloads running on the jointly managed platform. Key advances include OpenShift Virtualization, confidential containers, managed identities, expanded NVIDIA GPU support, and broader regional availability.

📊 Database Center now offers an AI-native manageability interface powered by Gemini, delivering fleet-wide visibility across Google Cloud managed databases. It introduces Gemini-backed recommendation validation (coming soon) to simulate performance impacts before applying changes like new indexes or machine upgrades. Additional enhancements include inventory, end-to-end lineage, and automated health checks to reduce MTTR and operational overhead.

🛡️AWS WAF now supports dynamic label interpolation, letting you forward WAF classification signals to your origin and embed contextual data in responses using a single rule. Using the ${namespace:} syntax in custom request headers, response headers, and response bodies, you can pass entire label namespaces (including AWS Managed Rules, marketplace groups, or custom labels) without separate rules. Interpolation adds synthetic labels like client IP, WAF request ID, and JA3/JA4 fingerprints, adapts headers automatically, and is available in all AWS Regions at no extra cost and with no new API fields or configuration steps.