Product Announcement

[Invitation] Galaxy Event August 2026

Consumers can discover a new smartphone option that incorporates high-end AI and camera features from the S26 series. This allows potential buyers to evaluate whether to upgrade their current device.

• Event scheduled for August 27, 2026 at 9 p.m. KST
• Streamed live on Samsung.com and Samsung's YouTube channel
• Introduces the newest addition to the Galaxy S26 family with camera and AI experiences and latest One UI
• Galaxy AI basic features are free under current Samsung Services Terms and Conditions

XBOX @ gamescom 2026 - XBOX Wire

Fans can get hands-on experience with several unreleased games before they launch or watch live gameplay reveals via streaming services.

• Features 25 upcoming games and 140 hands-on gaming stations
• Hands-on playable demos for Call of Duty: Modern Warfare 4, Gears of War: E-Day, Minecraft Dungeons II, and Fable
• Live broadcasts scheduled on YouTube and Twitch

From Atari to EVE Online: Building on 15 Years of AI Research in Games

These advancements could provide gamers with more adaptive NPCs and companions while offering developers better automated testing tools. Historically, game-based AI research has led to real-world scientific breakthroughs like protein structure prediction.

• SIMA 2 is powered by Gemini frontier AI models for real-time reasoning and conversation.
• It achieves human-like play across research environments and commercial games like No Man's Sky, Valheim, and Hydroneer.
• It interacts via standard screen pixels, natural language, and keyboard/mouse controls without requiring game code access.

Amazon EKS now supports certificate authority (CA) rotation with automated lifecycle management

Users with older EKS clusters must perform this rotation to avoid service disruptions caused by expired certificates. It provides a way to maintain security without risking downtime through automated safeguards and rollback capabilities.

• Enables rotation of cluster Certificate Authorities (CAs) with 10-year validity periods approaching expiration for clusters created since 2018
• Operates as a shared responsibility model where AWS automatically updates AWS-managed components, EKS Auto Mode instances, and AWS Fargate nodes
• Includes automated safeguards such as advance notifications, automatic appending of successor CAs, and automatic activation if not scheduled by customers
• Provides a rollback capability to revert to the previous CA to resolve any issues during the update transition
• Available at no additional cost in all commercial AWS Regions via AWS CLI, EKS APIs, CloudFormation, and the AWS console

Offering Zero Data Retention for frontier models

Companies handling sensitive data can use advanced AI safety monitoring without violating security obligations or compromising privacy by allowing data retention.

• Private Safety Processing evaluates patterns across related interactions without giving OpenAI personnel access to customer content
• Supports customer-controlled infrastructure as well as OpenAI storage encrypted with customer-managed keys
• Full rollout and technical white paper planned for September 2026

AWS IAM identity federation to external services is now available in AWS European Sovereign Cloud Region

Organizations using this EU-based cloud can now securely connect to third-party SaaS and cloud providers without relying on risky long-term credentials.

• AWS workloads in the AWS European Sovereign Cloud (Germany) Region can authenticate with external services using short-lived JSON Web Tokens (JWTs).
• Outbound identity federation eliminates the need for long-term credentials or complex workarounds when accessing third-party cloud providers, SaaS providers, and self-hosted apps.
• Administrators can enforce token properties via IAM policies and audit token usage using CloudTrail logs.

How Claude's text watermarking works \ Anthropic

This allows users to verify if text was likely generated by Claude via an upcoming API, increasing transparency regarding AI-authored content.

• Future Claude models will incorporate text watermarking that alters the source of randomness during word selection without degrading quality, content, or speed.
• Anthropic signed the EU Code of Practice on Transparency of AI-Generated Content in July 2026 along with approximately 190 signatories.
• Generated images and supported files will include C2PA content credentials cryptographically signed in the file metadata.
• Anthropic plans to release a watermark detection API and will roll out watermarking to older Claude models launched before August 2, 2026.

Gemini Omni experts answer key questions about the model

It reduces the complexity of video production by allowing creators to edit and generate videos through simple conversation.

• Gemini Omni is a new model enabling creators to generate anything from any input.
• Gemini Omni Flash is the model's first release, simplifying video generation and editing into a conversational workflow.
• Google DeepMind team members Mohammad Babaeizadeh, Anish Nangia, and Sarah Xu shared insights into Omni's vision and creator focus.

Track subtle changes in your body with Health Guardian features on Pixel and Fitbit

Users can identify subtle health shifts early without invasive tests or cuffs to make proactive lifestyle changes. In select regions, emergency detection may save lives by calling services during critical oxygen drops.

• Health Guardian features bring proactive health tracking to Google Health app on Pixel Watch and Fitbit Air.
• Capabilities include Insulin Resistance Trends, Blood Pressure Trends, Sleep Breathing Quality Trends, and Breathing Emergency Detection.
• Features are powered by advanced AI models pretrained on sensor data from millions of opted-in users.
• Google Health Coach assists users in understanding health insights and making lifestyle changes.

3 updates that make Pixel 11 phones even more helpful

These updates provide users with more efficient tools for capturing spontaneous photos and transcribing voice conversations in their daily lives.

• Pixel 11 lineup includes Pixel 11, Pixel 11 Pro, Pixel 11 Pro XL, and Pixel 11 Pro Fold.
• Introduced Magic Capture camera tool, Rambler voice transcription, and HiLight notification features.
• Preorders began August 12, 2026, with store availability on August 20, 2026.

New connected apps are coming to Gemini

Users can perform various real-world tasks, such as booking doctor appointments or streaming music, directly through the AI assistant.

• Third-party tools and services are being integrated directly into Gemini over the coming weeks.
• Integrations cover productivity, local experiences, music, and lifestyle services.
• Partners include Granola, Otter.ai, Wix, Fever, Get Your Guide, Localiza, OpenTable (UK), Ticketmaster, iHeartRadio, Pandora, Angi, Thumbtack, and Zocdoc.

Breathing Emergency Detection, now on Pixel Watch

This can save lives by ensuring help is summoned when a person is alone and unable to call for assistance during a breathing emergency.

• Uses heart rate, motion, and altitude sensors to detect breathing emergencies
• Includes a 30-second countdown alert before automatically contacting emergency services and contacts
• Available on Pixel Watch 4 and 5 in select European countries

Introducing Pixel 11 Pro Fold

A new sign-to-text feature improves communication for the Deaf community through real-time translation. Enhanced durability and faster charging make foldable technology more reliable for daily use.

• Nearly 10% lighter and about 1mm thinner than Pixel 10 Pro Fold
• Powered by custom Tensor G6 chip with 16GB RAM
• Features new HiLight LED notification system on camera bar
• Includes 48MP main camera and 30x Super Zoom telephoto camera
• Supports sign-to-text translation in Live Transcribe and Gboard

ClusterNetworkPolicy in GKE: Balancing control and autonomy for your microservices

Security teams can enforce global compliance and prevent unauthorized data movement without hindering developers' ability to manage their own app communication. This creates a safer multi-tenant environment by balancing central control with developer autonomy.

• ClusterNetworkPolicy (CNP) is an open-source standard developed by the Kubernetes SIG-Policy Working Group.
• Establishes a three-tier hierarchy (Admin, Network Policy, and Baseline) to resolve policy conflicts between security admins and developers.
• Developed in collaboration with the Kubernetes and Cilium open-source communities.
• Available in preview for GKE version 1.36 and later.

Try new AI tools from Google Ads and Analytics

Business owners can spend less time on manual data analysis and more time reacting to market trends or competitor performance. This leads to faster strategic adjustments for business growth.

• Ask Advisor is an in-product AI agent across Google Ads and Google Analytics to analyze performance and execute marketing actions.
• Google Analytics adds AI Overviews at the top of the homepage for instant summaries of performance shifts.
• Google Ads revamped its homepage with AI-powered insights cards and custom prompt capabilities.
• New visual reporting dashboards in Google Ads and Google Analytics generate real-time summaries from text prompts.
• Google Analytics introduced a benchmarking feature to compare campaign performance against anonymized industry averages.

Zero-code, low-cost data ingestion: New BigQuery DTS capabilities

This reduces the engineering time and cost enterprises spend building manual data pipelines, allowing teams to focus on data analysis rather than maintenance.

• Direct ingestion into Apache Iceberg managed tables in Preview from Google Cloud Storage, Amazon S3, and Azure Blob Storage.
• Fully managed remote Model Context Protocol (MCP) Server in Preview for AI application and agent integration.
• New database connectors for Microsoft SQL Server, PostgreSQL, and MySQL.
• New SaaS connectors for Shopify, Klaviyo, HubSpot, and Mailchimp.
• Enhanced migration features for Snowflake and incremental update support for ServiceNow, Salesforce, and Oracle.

Advance Zero Trust for AI: New tools and guidance to secure AI agents and DevSecOps

AI assistants and autonomous workflows introduce new attack surfaces into software development. This allows security teams to identify gaps and apply specific controls to prevent vulnerabilities in AI-integrated environments.

• Introduced a new DevSecOps pillar with 15 control groups and 91 tasks in the Zero Trust Workshop.
• Added a new AI pillar to the Zero Trust Assessment tool with expanded checks for AI, SecOps, and Infrastructure.
• Published the e-book Zero Trust for AI: Rebuilding security controls for autonomous and agentic systems.

How we built a realtime system for responsive voice AI in six months

Users experience sub-second responsiveness in ChatGPT Voice, enabling natural interruptions during conversation. This technology also powers new capabilities like controlling computers and coordinating agents in the desktop app.

• Eliminates turn detector with a full-duplex model that listens and speaks simultaneously
• Delegates complex reasoning and tool execution asynchronously to GPT-5.5
• Rewrote media frontend in Go, bringing p95 latency down to previous p50 levels
• Developed WARP protocol with IETF to cut WebRTC startup handshake from six round trips to one

The borderless Lakehouse: Bring AWS, Databricks and Snowflake data to your AI agents

Companies can reduce the cost and time spent on moving massive datasets while providing AI agents with real-time access to distributed data for more accurate business outcomes.

• Catalog federation in preview for AWS Glue, Databricks Unity Catalog, and Snowflake Horizon
• Zero-copy data integration with SaaS applications like SAP, Salesforce, and Workday
• Spanner Omni and Lakehouse Federation for AlloyDB
• Partner Cross-Cloud Interconnect for zero variable egress cost access to AWS data
• Integration of Knowledge Catalog as an agentic context engine
• Release of Google Cloud Data Agent Kit for building custom data agents

Today, we are announcing a series of updates that give customers frontier-grade security at half the cost. MAI-Cyber-1-Flash is our first cybersecurity model, built ground up to find the most… | Satya Nadella | 361 comments

Organizations can access high-grade security capabilities at half the cost of leading models. This makes it more affordable for companies to find and fix critical vulnerabilities in their software.

• MAI-Cyber-1-Flash is Microsoft's first cybersecurity AI model for code vulnerability detection.
• MDASH combines with MAI-Cyber-1-Flash to achieve world-class performance at 50% of the cost of leading models.
• Project Perception provides an agentic security offering using specialized agents to simulate attacks, detect vulnerabilities, and remediate threats.

NVIDIA Open Sources First GPU-Accelerated Medical Physics Simulation Framework | NVIDIA Blog

By reducing training time and providing transparent data, this technology can accelerate the delivery of consistent robotic surgical care to patients and simplify regulatory approval processes.

• GPU-accelerated open-source framework for medical physics simulation and anatomical modeling
• Integrates classical physics simulation with real-time generative AI simulation via Cosmos-H Dreams
• Demonstrates 8,192 parallel robot-training environments cutting training time from 5 hours to under 2 minutes
• Adopted by medical robotics leaders including CMR Surgical, Johnson & Johnson MedTech, XCath, Inner Logic, and Medtronic Structural Heart

Microsoft announced the launch of Microsoft Defender Experts Threat Intelligence and the expansion of Microsoft Defender Experts MDR with new third-party and multi-cloud coverage powered by Microsoft Sentinel.

• Microsoft announced a new service, Microsoft Defender Experts Threat Intelligence, which provides expert-delivered curated threat insights.
• Microsoft expanded Microsoft Defender Experts MDR to include new third-party and multi-cloud coverage powered by Microsoft Sentinel.
• The expanded coverage is available through Microsoft Defender Experts MDR Plan 2.

Microsoft announced that starting September 1, 2026, it will roll out passkeys as the default phishing-resistant authentication method in Microsoft Entra ID, and will retire native telecom delivery for SMS and voice authentication on February 1, 2027.

• Starting September 1, 2026, Microsoft will begin rolling out passkeys as the default authentication experience in Microsoft Entra ID.
• On February 1, 2027, Microsoft will retire native telecom delivery for SMS and voice authentication in Microsoft Entra ID.
• Admins requiring SMS or voice will be able to configure a supported telecom provider via the Microsoft Security Store starting October 30, 2026.

OpenAI announced GPT-5.6 as the new preferred model in Microsoft 365 Copilot across core apps.

• GPT-5.6 is OpenAI's latest flagship model series.
• It will become the preferred model in Microsoft 365 Copilot apps (Word, Excel, PowerPoint, Chat, and Cowork).
• The model is accessed natively and through the OpenAI API.

Connecting Real People on Facebook

These tools simplify the selling process for frequent users and help people verify identities before meeting or buying items.

• Launched Seller app featuring AI-powered listing creation, unified inbox, and performance insights.
• Introduced Facebook Verified, a free selfie verification badge for user profiles.
• Testing Forum app for Facebook Groups with Admin AI and Group Ask features.
• Reimagining Facebook Creator Studio as a dedicated app with Creator Assistant AI.

Why Scaling AI Compute Performance Requires a New Power Architecture

This enables data center operators to increase compute performance and density without replacing existing building electrical infrastructure, preventing previous facility investments from becoming obsolete.

• Arriving in the second half of 2026
• Creates a hybrid architecture slotting into existing AC infrastructure
• Delivers 800 VDC to compute racks within the row without changes to building electrical systems

Meta introduces new safety protections, including proactive parent alerts and stricter content filters, for teen conversations with Meta AI.

• Meta will proactively alert supervising parents if their teen's Meta AI chat indicates self-harm or suicide risk.
• Flags generated by the dedicated AI detection system will be manually reviewed before alerts are sent.
• Parental alerts are launched for Instagram supervision in the US, UK, Australia, and Canada, with global rollout by the end of the year.
• Meta is developing the capability to contact emergency services for users at imminent risk of self-harm.
• Instagram's stricter Limited Content setting is extended to limit prompts and responses in Meta AI chats.

Apple announced upcoming features for Apple Maps, Find My, Apple Cash, Apple Podcasts, iCloud, and Apple Fitness+.

• Enhanced Flyover and Local Lists in Apple Maps
• Flexible sharing options in Find My
• Split bills using Apple Cash and Visual Intelligence
• Video podcast support for macOS and tvOS

Apple previewed a new suite of child safety features, including Ask to Browse, Time Allowances, and a redesigned Screen Time, coming to iOS, iPadOS, and macOS this fall.

• Apple introduced Ask to Browse, letting parents approve Safari website requests.
• Communication Safety is updated to detect and block gore or violent content in shared media.
• Time Allowances and Schedules provide parents with age-based guidance and flexible app controls.
• Screen Time is redesigned to offer an at-a-glance dashboard of usage and instant controls.
• Apple is collaborating with the American Academy of Pediatrics (AAP) to adapt its Family Media Plan.

The next measure of AI momentum is work transformed

Automating complex multi-step tasks reduces time spent on repetitive administrative work, allowing employees to focus more on strategic decisions and customer needs.

• Stays active in the background with its own identity and permissions.
• Developed from concept to working product in weeks by a small frontier team.

[dev] Claude for Creative Work \ Anthropic

Creative professionals can reduce manual toil and spend less time on repetitive work, allowing them to focus on larger-scale projects and imaginative work.

• Released connectors integrating Claude into top creative and 3D software suites
• Provided a donation to support the open-source Blender project and its Python API
• Partnered with RISD, Ringling College, and Goldsmiths to bring Claude connectors to creative computation curricula

Record, train, and deploy from one place with Strands Agents, LeRobot, and Hugging Face Storage Buckets

This process reduces the time GPUs spend idle during downloads and lowers storage costs by only uploading changed bytes of large robotics datasets.

• Storage Buckets provide a mutable, non-versioned working layer for dataset collection and streaming
• Uses Xet-backed byte-level deduplication with content-defined chunking
• Operates in the same hf:// namespace and uses the standard hf CLI

New currency capabilities for global businesses to cut FX costs

Businesses can avoid paying double conversion fees and hidden markups when operating internationally. This reduces operational overhead by consolidating payments, banking, and FX into one system.

• Multicurrency settlement expanding to 37 markets including Australia, Hong Kong, and Singapore, with up to 18 settlement currencies by end of 2026.
• Instant currency conversion launched for 15 currencies with 24/7 availability and no hidden exchange rate markups or weekend surcharges.
• Introduced API and Dashboard features for automated conversions and platform currency management with Stripe Connect, Treasury, and Global Payouts.