The Big Data & AI Product Monthly Newsletter for June 2026 covers technology updates, product and feature releases, market updates, and customer application practices to help you quickly understand the latest developments in Alibaba Cloud's Big Data & AI offerings.
PAI - Agentic Readiness V1.0: Core CLI & Skills
Users can now operate PAI directly from their preferred Agent environments, without needing to learn or use the PAI console.
PAI - DSW AgentBox: Built-in AI Coding Agents
Leverage AI agents for code writing, debugging, rapid prototyping between training jobs, and full-stack AI-assisted development on GPU instances. Team administrators can use PAI Tokens to configure access in bulk, eliminating the need to issue API keys individually and lowering the barrier for teams to adopt AI coding tools.
MaxCompute - High-Precision DECIMAL256 Data Type
New DECIMAL256 type supports up to 76-digit precision. The auto.promotion setting automatically promotes standard DECIMAL results to DECIMAL256, preventing precision truncation for finance, scientific computing, and other high-precision scenarios.
MaxCompute - Model Service Adds Qwen3.7-Plus, Qwen3-ASR and More
MaxCompute Model Service now supports top-tier commercial models including qwen3.7-plus and qwen3-asr-flash, with a new Singapore region. Both MaxCompute SQL and MaxFrame built-in AI Functions can seamlessly call these models with a single line of code, delivering effortless access to cutting-edge AI.
MaxCompute - MaxFrame ML Algorithm Extensions (maxframe.learn)
The MaxFrame learn module adds Logistic Regression, PCA, TruncatedSVD, OneHotEncoder and other operators, completing the classical ML capability set.
MaxCompute - MaxFrame Multimodal & Rate-Limiting Enhancements
MaxFrame AI Function expands to support audio and video multimodal inputs, with new TPM/RPM rate limiting and LLM operator fusion for improved inference throughput.
MaxCompute - MaxFrame Vector Data Type Support
MaxFrame DataFrame now supports the Vector data type, enabling DataFrames to directly read, write, and process vector columns.
MaxCompute - MaxFrame Local Debug Enhancement
MaxFrame adds local debug support for read_odps_query, making development and debugging more convenient.
MaxCompute - SQL AI Function Multimodal Data Processing
AI_Embedding and AI_Generate functions now support multimodal input parameters, integrating with Bailian multimodal and Embedding models. Binary type input is supported, enabling vectorization of unstructured data such as images and prompt-based LLM processing — unifying structured and unstructured AI analysis within SQL.
Hologres - Long-Term Memory Service
The Hologres Long-Term Memory Service is a managed service for building dedicated memory for AI Agents, 100% compatible with the open-source Mem0 framework. It automatically extracts and persists key information from multi-turn conversations, solving the problems of inconsistent memory and inability to personalize responses across sessions.
DataWorks - Sensitive Data Protection Expanded to More Engines
Sensitive data identification and dynamic masking now covers additional compute engines: MaxCompute, Hologres, DLF, DLF Legacy, EMR Hive, and EMR StarRocks.
DataWorks - Data Agent Major Upgrade
DataWorks Data Agent adds three new panels: a context latency waterfall view for visualizing call chains, output details with right-side preview for code, documents, and images, and a session environment panel showing workspace information — delivering a comprehensive upgrade to interaction experience and readability.
DataWorks - AI-Powered Code Review Summarization
During data development code reviews, reviewers can now use AI to summarize review suggestions, improving review efficiency.
DataWorks - Data Security & Sensitive Data Feature Upgrade
Data security and sensitive data features receive functionality and performance upgrades. New and some existing users will be automatically upgraded to the latest version. A clear sensitive data inventory helps users manage access permissions, apply data masking, and audit data access to strengthen overall data security.
DataWorks - AI-Based Lineage Analysis for Notebook/PyODPS Nodes
Python/MaxFrame code logic involving MaxCompute tables, datasets, or OSS paths as inputs/outputs can now have its lineage intelligently parsed by AI and registered to the lineage chain with one click, significantly improving efficiency.
DataWorks - Data Quality Adds Multiple Rule Templates
DataWorks Data Quality adds rule templates covering validity, consistency, accuracy, and timeliness. New scenarios include JSON field format validation, cross-field value comparison within a single table, aggregated value fluctuation comparison by dimension+metric, and time-series data latency detection — enabling users to quickly implement quality controls for metric-oriented data.
DataWorks - Data Integration Adds New Sources & Sync Capabilities
New capabilities include: full-database offline sync from OceanBase to MaxCompute; single-table offline sync supporting DLF (Paimon, blob) format reads; full-database real-time sync from PolarDB sharded tables to StarRocks; single-table offline sync supporting Flink Fluss source reads; full-database real-time sync from Oracle to DLF (Paimon); and full-database offline sync from SQL Server, ADB, and Oracle to ADB.
DataWorks - New Data Integration Agent
A new Data Integration Agent enables one-stop task creation, modification, operations, and diagnostics. The Agent comes pre-configured with scenario-based data integration capabilities to accelerate user onboarding.
EMR - Serverless Spark Supports Spark Clusters
Spark Cluster is a persistent cluster mode provided by EMR Serverless Spark. Once started, the cluster runs continuously and tasks can be submitted directly without waiting for resource provisioning — ideal for scenarios requiring fast task startup.
EMR - AI Assistant Officially Released
New task center with health inspection configuration for scheduled comprehensive instance checks; slow SQL alerting with configurable thresholds per business scenario; IM channel push configuration to send inspection reports and alerts directly to DingTalk/Feishu/WeCom groups with mobile access; expert skills covering diagnostics, optimization, consultation, and development; and API/SDK access to embed the AI Assistant into enterprise platforms.
EMR - Serverless Spark Supports Ray Job
Ray Job support allows users to submit tasks to Ray clusters for execution via the Ray Job interface.
EMR - Serverless Spark Supports Cache Clusters
Create and manage cache clusters to accelerate data access through data preloading and local caching, improving I/O performance for Spark jobs.
EMR - Serverless Spark Supports IP Whitelist
IP whitelist configuration, combined with RAM Policy, restricts Spark workspace access to specified IP ranges, enhancing network security.
EMR - Serverless Spark Supports Hologres Data Catalog
New Hologres integration allows users to add Hologres in the data catalog for accessing and querying related data.
EMR - Serverless Spark Best Practice: Fluss Integration
New best practice for Fluss integration, helping users implement Spark Streaming read/write for real-time Fluss data processing.
EMR - Serverless Spark Best Practice: Paimon Hybrid Search
New best practice for Paimon vector + scalar hybrid search, helping users leverage Paimon's vector indexing for efficient hybrid retrieval.
Elasticsearch - Event Center Intelligent Diagnostics
Provides comprehensive intelligent diagnostics for the ES Event Center.
Elasticsearch - AMD Turbo2 New Instance Specs
New turbo2 series AMD instances deliver up to 20%+ performance improvement in write and query scenarios, offering enterprise users a higher-performance option.
Elasticsearch - Proprietary v9.4 Released
The new version supports advanced features including Agent Builder, Workflow, and DiskBBQ, combined with Alibaba Cloud proprietary features like OpenStore and FalconSeek for significant multi-scenario performance gains.
Elasticsearch - ES Agent AI Assistant (Public Beta)
Elasticsearch launches an AI assistant featuring an intelligent diagnostics assistant and an inspection assistant, helping users self-diagnose cluster issues and reduce operational complexity.
OpenSearch - Vector Search Edition Index Data Recovery
Supports recovering data from indexes and rebuilding indexes from recovered records in the Vector Search Edition, resolving the previous limitation where index rebuilding could only start from empty data.
OpenSearch - Vector Search / Recall Engine Monitoring & Alerting Upgrade
Monitoring and alerting system upgraded to Cloud Monitor, with performance and UX improvements.
MaxCompute Agentic Ecosystem Debut: Let AI Directly "Read" Your Data Warehouse
25 posts | 0 followers
FollowAlibaba Cloud Big Data and AI - May 15, 2026
Alibaba Cloud Big Data and AI - June 11, 2026
Alibaba Cloud Big Data and AI - April 13, 2026
Alibaba Cloud Big Data and AI - January 21, 2026
CloudSecurity - June 15, 2026
CloudSecurity - May 18, 2026
25 posts | 0 followers
Follow
ECS(Elastic Compute Service)
Elastic and secure virtual cloud servers to cater all your cloud hosting needs.
Learn More
Big Data Consulting for Data Technology Solution
Alibaba Cloud provides big data consulting services to help enterprises leverage advanced data technology.
Learn More
MaxCompute
Conduct large-scale data warehousing with MaxCompute
Learn More
Big Data Consulting Services for Retail Solution
Alibaba Cloud experts provide retailers with a lightweight and customized big data consulting service to help you assess your big data maturity and plan your big data journey.
Learn MoreMore Posts by Alibaba Cloud Big Data and AI
