Cloud · Top stories
Cloudflare launches Streamline developer playground for custom video pipelines
Cloudflare introduced Streamline, a new developer playground demonstrating how to build custom video pipelines using Cloudflare Stream, Workers, and Containers. Streamline allows developers to create dynamic video experiences like live annotations or burned-in subtitles by modifying video and publishing the output as a livestream or hosted video. This provides a framework for developers to implement bespoke video processing workflows on Cloudflare's platform.
Cloudflare is now the fastest provider in 74% of the top 1,000 global networks
Cloudflare announced it is now the fastest provider in 74% of the 1,000 largest global networks, an increase from 60% in April 2026. This improvement means users can access applications, APIs, and websites more quickly, impacting overall internet experience.
Cloudflare expands enterprise features to all plans, including Logpush and Transformers
Cloudflare is making more enterprise-tier features available to all customers, including free and pay-as-you-go plans. This change allows smaller businesses and individual users to access advanced capabilities like Logpush and Logpush Transformers, previously restricted to Enterprise accounts. The move aims to democratize access to Cloudflare's full suite of tools, enabling more users to monitor traffic and analyze data.
Google Cloud Storage SDKs Enable End-to-End Checksums by Default for Data Integrity
Google Cloud Storage SDKs now enable end-to-end checksumming by default for all uploads and downloads. This change improves data integrity by protecting against bit flips during data transfer, even when applications do not explicitly provide checksums.
Cloudflare OS now offers fully managed deployments for agent workspaces
Cloudflare has opened a waitlist for fully managed deployments of Cloudflare OS, an agent workspace designed to connect to company data and systems. This new offering allows organizations to deploy Cloudflare OS with simplified configuration and maintenance, building on the previously released open-source version.
Cloudflare launches Workers KV Instant for faster global data replication
Cloudflare introduced Workers KV Instant, a new mode for its Workers KV service, which utilizes the Quicksilver key-value store. This update provides 100 times faster p99 reads and immediate global updates for data, addressing the need for rapid access to infrequently updated application configurations.
Google Cloud Managed Lustre Introduces Dynamic Tier for Lower Cost and AI/HPC Workloads
Google Cloud Managed Lustre now offers a Dynamic Tier priced at 6 cents/GB*month, allowing users to store all data in a single namespace. This aims to reduce costs and simplify data management for AI and HPC workloads by providing sub-millisecond latency for hot data and linear throughput scaling.
Cloudflare Basin, an open and serverless data platform, is now generally available
Cloudflare announced the general availability of Cloudflare Basin, a serverless data analytics platform built on Apache Iceberg and R2 Object Storage. Basin provides an end-to-end analytics solution, allowing users to collect data from various sources and query it using SQL, addressing the need for portable and cost-efficient data analytics.
Amazon S3 Tables now support all Apache Iceberg V3 data types
Amazon S3 Tables now support all data types specified in Apache Iceberg V3, including variant, nanosecond timestamps, geometry, geography, and unknown. This update allows users to create new V3 tables or upgrade existing V2 tables to utilize features like deletion vectors and row lineage, addressing common performance and data management challenges in large analytics datasets.
Amazon S3 Vectors adds metadata pre-filtering for improved recall in filtered searches
Amazon S3 Vectors now supports metadata pre-filtering, which evaluates metadata filters before similarity searches. This change improves recall for filtered queries in applications like semantic search and RAG, ensuring more relevant results are returned when searching specific subsets of data.
Netlify Edge Functions now run on MicroVMs, achieving 5x speed improvement
Netlify has re-architected its Edge Functions infrastructure to run on MicroVMs within its edge network, replacing a hosted execution service. This change results in a 5x speed improvement at median latency, reducing it from 25-40ms to 5-6ms, and enhances security and reliability for edge computing.
Amazon Aurora PostgreSQL now supports direct querying of Apache Iceberg and Parquet data
Amazon Aurora PostgreSQL now allows direct querying of Apache Iceberg and Parquet data in data lakes, eliminating the need for ETL processes. This integration simplifies combining operational and historical data, reducing operational complexity and supporting AI applications.
Cloudflare Workers introduces 'Issues' for automated error monitoring and agent integration
Cloudflare has launched 'Issues' in open beta for Cloudflare Workers, providing built-in error monitoring that groups exceptions, 5xx responses, and error logs. This feature automatically sends error context to coding agents, streamlining the process of identifying and resolving production issues.
AWS European Sovereign Cloud to Demonstrate Independent Operation on October 24, 2026
AWS will conduct an exercise on October 24, 2026, to demonstrate that its European Sovereign Cloud can operate independently of non-EU infrastructure. During this period, the cloud will function without a connection to the AWS Global Network backbone, with traffic rerouted over the public internet, to verify its operational independence for European customers.
Google Cloud announces general availability of M4N VM family for I/O and memory-bound workloads
Google Cloud has announced the general availability of its M4N virtual machine series, designed for I/O-intensive and high-memory workloads. This new VM family offers higher per-core IOPS and throughput, aiming to reduce total cost of ownership for applications like Oracle databases by optimizing memory-to-vCPU ratios.
Meta Introduces ZGateway for ZippyDB, Reducing Connections by 19x
Meta launched ZGateway, a stateless proxy layer for its ZippyDB distributed key-value store, to manage over one million client hosts and improve reliability. This new architecture reduces total persistent connections by approximately 19 times while handling over one billion operations per second.
AWS Aurora DSQL Adds Foreign Key Constraints with Referential Actions
AWS Aurora DSQL now supports foreign key constraints, including CASCADE and SET NULL, allowing applications to enforce referential integrity directly within the database. This addition addresses a significant adoption blocker for the distributed SQL database, making it more viable for migrations and brownfield environments.
GKE Pod Snapshots Reduce AI Model Load Times by up to 89%
Google has released GKE Pod snapshots to general availability, reducing AI model load times by up to 89% by saving and restoring the running state of workloads. This feature allows large models to bypass lengthy initialization processes, significantly speeding up startup times for AI/ML applications on GKE.
Australian Officials Claim Local Data Center Backlash is Imported US Sentiment
Australian politicians and business leaders state that the growing public opposition to AI data centers in Australia is influenced by US social media narratives, rather than local conditions. They argue Australia's data center buildout is more regulated and smaller in scale compared to the US, making the backlash disproportionate to the actual impact.
Google's Hyperscale AI Datacenter in India Faces Local Opposition Over Environmental Concerns
Google is constructing a $15 billion hyperscale AI datacenter in Tarluvada, India, which is set to be its largest outside the US. The project faces local opposition and accusations of government obfuscation regarding its environmental impact and scale, with villagers fearing economic and environmental costs.
Cloudflare Migrates Main Blog from WordPress to Internal EmDash CMS
Cloudflare migrated its main blog from WordPress to EmDash, an open-source content management system developed internally. This change was made to improve performance and caching, addressing limitations of the previous platform and handling traffic spikes more efficiently.
Netflix Details Workload Attestation for Apache Spark on Amazon EMR
Netflix developed a method to bridge cloud provider identities with internal identity systems for Apache Spark workloads on Amazon EMR. This allows managed compute instances, which initially only have cloud identities, to acquire internal identities necessary for accessing internal services and data. The solution enables consistent access control and auditing for data projects.
Google Cloud Storage launches Storage Intelligence advisor and expands batch operations
Google Cloud Storage has released Storage Intelligence advisor into general availability and expanded capabilities for storage batch operations. These updates provide users with automated insights into storage changes and enable large-scale object management, addressing the complexity of managing growing data volumes.
Google Cloud's Memorystore for Valkey 9.1 achieves 3x QPS and microsecond latency
Google Cloud announced the general availability of Memorystore for Valkey 9.1, offering up to three times higher queries per second (QPS) and microsecond latency compared to Memorystore for Redis Cluster. This performance improvement is due to a redesigned lock-free, multi-queue messaging architecture that optimizes thread communication and dynamic work balancing.
Perplexity Replaces DynamoDB with Custom CobbleDB to Reduce Latency and Storage Costs
Perplexity migrated its core search serving from Amazon DynamoDB to CobbleDB, an internally developed Rust-based distributed key-value store. This change reduced batch-read latencies fivefold and lowered storage expenses by at least 20% by decoupling durable storage from hot-tier retrieval.
AWS Updates Model Context Protocol (MCP) to Remove Session Affinity Requirements
AWS has updated the Model Context Protocol (MCP) specification to eliminate protocol-level sessions, allowing requests to be routed to any available server instance. This change simplifies horizontal scaling for MCP server deployments by removing the need for sticky sessions and shared session stores, shifting state management responsibilities to surrounding infrastructure.
Oracle Japan reports record Q1 sales and profits, driven by cloud growth
Oracle Japan's fiscal first-quarter net sales increased 13% to 74.86 billion yen, with operating profit up 22.7% and net profit up 23.2%, setting new records for the quarter. This growth was primarily driven by a 31.7% rise in cloud revenue, contrasting with a decline in its U.S. parent's shares.
Amazon EventBridge introduces enhanced custom event buses for multi-account event-driven applications
Amazon EventBridge launched enhanced custom event buses to simplify event-driven architectures across multiple AWS accounts. This update allows organizations to deploy a single, centralized event bus, addressing operational complexity and cost issues associated with previous multi-bus configurations.
Google XProf Adds Cycle-Level Kernel Profiling for TPU Workloads
Google has integrated a Kernel Profiling suite into XProf, its open-source profiler for TPU workloads, allowing developers to inspect cycle-level details within custom Pallas kernels. This enhancement addresses inaccuracies in static cost models for custom compilation paths and enables more precise optimization of TPU performance. For example, Google used it to reduce kernel time by 30% in a tiled matmul example by identifying and resolving a memory stall.
Amazon's Project Kuiper Partners with Uztelecom to Bring Satellite Internet to Uzbekistan
Amazon's Project Kuiper, branded as Leo, is partnering with Uzbek ISP Uztelecom to provide satellite internet connectivity to remote regions of Uzbekistan. This collaboration aims to extend internet access in areas where traditional cabled infrastructure is difficult to deploy, complementing Uztelecom's existing network.
Texas Governor Halts Data Center Environmental Permits Amid Grid Audit
Texas Governor Greg Abbott ordered the state's environmental regulator to stop issuing data center permits until an audit of grid connection requests is complete. This expands an existing moratorium on new grid approvals, potentially impacting Texas's growth as a data center hub and risking significant revenue losses.
Google's Managed Service for Apache Spark Introduces Flexible VMs to Mitigate Capacity Stockouts
Google's Managed Service for Apache Spark now offers Flexible VMs, allowing users to define multiple acceptable machine families for Spark clusters. This feature helps prevent cluster creation delays and execution failures caused by compute capacity stockouts in specific zones or regions, particularly relevant with increased AI-driven demand for compute.
New York State Recommends $1 Million Per Megawatt Community Investment from Data Centers
New York State introduced the Community Investment Framework (CIF), advising towns to seek $1 million per megawatt from data center developers for community benefits. This framework aims to help municipalities manage the impact of data centers, particularly AI data centers, which strain resources while offering limited local job creation.
Cloudflare Reduces TLS Handshake Retries from 52% to 3.7% with Per-Origin Measurement
Cloudflare replaced a static assumption in its origin TLS handshakes with a per-origin measurement system, reducing HelloRetryRequests on scanned origins from 52% to 3.7%. This change also decreased p90 handshake latency by over 150 ms by correctly guessing the preferred key agreement algorithm for each origin. The update improves connection efficiency and latency for TLS 1.3 handshakes.
AWS Lambda SnapStart Now Supports Container Images, Addressing Startup Latency
AWS has extended Lambda SnapStart functionality to functions packaged as container images, which previously only supported managed runtimes. This change eliminates the startup latency penalty for Lambda functions that use container images to accommodate large dependencies, allowing them to achieve sub-second cold start times.
Google Cloud Adds Native BM25 Ranking to AlloyDB and Cloud SQL for PostgreSQL 17+
Google Cloud has introduced native BM25 ranking in AlloyDB and Cloud SQL for PostgreSQL 17+, eliminating the need for a separate full-text search backend. This integration provides a unified hybrid search backend for combining semantic vector search with traditional keyword search, which is critical for generative AI and RAG applications.
Google Cloud introduces cross-cloud caching for BigQuery to reduce data transfer costs
Google Cloud has launched a preview of cross-cloud caching for BigQuery, allowing frequently accessed data from other clouds to be cached locally in Google Cloud. This feature aims to reduce data transfer costs and accelerate cross-cloud queries for users of the borderless Lakehouse architecture.
Amazon EC2 T8i Instances with Intel Xeon Scalable Processors Now Generally Available
Amazon Web Services (AWS) has made its new low-cost burstable EC2 T8i instances generally available. These instances, powered by custom sixth-generation Intel Xeon Scalable Processors, offer up to 30% better price performance than previous T3 instances and are designed for low-to-moderate CPU utilization workloads.
Google Classroom experiences widespread outage affecting users
Google Classroom is experiencing a widespread outage, with reports spiking on Downdetector. This incident affects only Classroom, while other Google Workspace applications remain operational.
NVIDIA, Google, and Emerald AI form coalition to accelerate AI data center grid connections
NVIDIA, Google, and Emerald AI launched the AI Energy Management Alliance (AEMA) to standardize flexible AI data centers. The coalition aims to speed up grid connections for data centers that can reduce power usage during peak demand, addressing current grid limitations.