HiddenMerit Daily · Issue 16

📊 HiddenMerit Daily · Issue 16

Focus on Database Frontiers, Practical Insights for DBAs
May 9–11, 2026 ｜ 5 Selected Global Breaking News

01｜OceanBase CEO Yang Bing: Real‑Time Online Processing of Unstructured Data Is the Biggest Database Need in the AI Era

On the afternoon of May 9, at the 2026 China Economic Annual Observation and Xinhua Finance Global Ecosystem Partners Conference, Yang Bing, Vice President of Ant Group and CEO of OceanBase, stated that whether unstructured data can truly be processed online in real time is the biggest need in the AI era. How to build an open, real‑time, multi‑modal, and hybrid data lake is the core problem that the technology foundation of the AI era must solve.

Yang Bing believes that the software industry is undergoing a fundamental transformation from "serving people" to "serving agents". Databases in the AI era must simultaneously handle two major demands: core business critical workloads and AI innovation workloads – intelligently searching and retrieving multi‑modal data at the retrieval level, breaking through the integration of structured, unstructured, and semi‑structured data storage, and providing enterprises with unified data services. Quoting the ancient saying "long divided, must unite", he interpreted that the data foundation has evolved from a single database to data warehouses and data lakes, and now a new system of lakehouse integration has emerged.

· DBA Perspective: Yang Bing pushes the core of future data architecture from "storage" to the deep intelligent integration of "retrieval and analysis". For DBAs, facing a flood of massive heterogeneous data, pure structured optimisation is no longer sufficient – you need to master multi‑modal data modelling, layered storage design, and data models optimised for hybrid query and federated analysis for AI agents. This is not just adding skills; it's changing tracks.
· CTO Perspective: Yang Bing's remarks provide a clear strategic direction for technology leaders who are confused about AI applications – do not focus narrowly on single‑product replacement, but build a lakehouse‑integrated foundation from the perspective of the entire data architecture. This means prioritising databases that can handle both structured and unstructured data and have multi‑modal storage and retrieval capabilities when making technology selections.
· Investor Perspective: This is another core strategic statement from OceanBase following its global results report. As the industry moves from discussing "whether it needs to be done" to exploring "whether this is the right way", the valuation logic for databases in the capital market must also upgrade – extending from traditional TP capabilities and compatibility to a deep assessment of multi‑modal data foundation and agent architecture support capabilities.

02｜SAP Acquires Dremio and Prior Labs in One Week, Invests €1 Billion in a Structured Data AI Lab Over Four Years

On May 4, SAP announced the acquisition of open‑source high‑performance data lakehouse platform Dremio, expected to close in Q3 2026. The goal is to integrate Dremio's technology into SAP Business Data Cloud, enhancing real‑time analytics on internal and external data and improving vector AI workload efficiency. On the same day, SAP also announced the acquisition of German AI startup Prior Labs and plans to invest €1 billion over the next four years to build a "frontier AI lab" for structured data in Germany. The lab will operate as a relatively independent R&D entity, continuously maintaining open‑source versions while leveraging SAP AI Core, Business Data Cloud, and the Joule agent layer to rapidly embed research outcomes into SAP products. Previously, SAP also spent approximately $11 billion to acquire data streaming platform Confluent.

· DBA Perspective: Dremio's core value is its federated analytics capability for AI lakehouses – providing federated views across heterogeneous data sources without physically moving data. Following SAP's acquisition, there will be increasingly complex cross‑cloud, cross‑engine metadata governance scenarios. DBAs need to transform into "data federated architects", extending core skills from SQL tuning to cross‑platform metadata management, data virtualisation, and global access control.
· CTO Perspective: The acquisitions of Dremio and Prior Labs represent a battle for dominance over the enterprise data stack. SAP is rapidly building full‑chain AI capabilities from data integration to structured reasoning. The strategic significance for CIOs and CTOs is: can your ERP and data lake remain independent? The speed of enterprise software system integration over the next three years will exceed the governance capacity of most companies.
· Investor Perspective: SAP's acquisition spree once again demonstrates the determination and strength of enterprise software giants to "crowd into the AI infrastructure track". High‑premium acquisitions and heavy investment in data AI R&D provide strong reference points for exit strategies for data platform startups worldwide.

03｜Dameng DM9 Appears at 2026 Mobile Cloud Conference, Discloses Large‑Scale Deployment Results in Telecom Operators

On May 7, the 2026 Mobile Cloud Conference was held in Suzhou. As a key partner of Mobile Cloud, Dameng presented multiple strategic new products and delivered a keynote speech titled "DM9: New Starting Point, New Generation", showcasing cutting‑edge technologies of domestic databases in the AI era and industry empowerment practices, attracting visits from a large number of industry users including China Mobile Group, Guizhou Mobile, Shenzhen Mobile, and Singapore Mobile.

DM9 innovatively achieves a "centralised + distributed + TP + AP + AI" five‑in‑one architecture upgrade, deeply integrating five major pain points: the dilemma of choosing between centralised and distributed, the separation of transactions and analytics, the disconnect between on‑premises and cloud, the gap between databases and AI, and the split between software and hardware. In the telecommunications sector, Dameng has achieved large‑scale deployments, with its largest distributed cluster at Fujian Mobile carrying nearly 100 business systems and a cumulative total of 256 nodes. Its Qiyun Database Cloud Service System has become the first domestic relational database product on China Mobile Cloud.

· DBA Perspective: Under the "sprint"‑like high concurrency and strong transaction pressure of telecom operators, the large‑scale delivery of Dameng's 256‑node cluster is effectively a "hard" admission certificate for domestic databases to replace overseas core business systems. Mastering DM9's "five‑in‑one" mixed workload tuning is the key to unlocking critical industry opportunities for DBAs.
· CTO Perspective: The five‑in‑one architecture successfully encapsulates the dilemmas of different technical routes into a complete logic, allowing the application layer to smoothly schedule computing resources across different business workloads. The launch of the first domestic RDS product on Mobile Cloud is the most powerful validation – it marks Dameng's achievement of a "critical game point" in domestic replacement in core telecom operator scenarios.
· Investor Perspective: The 256‑node distributed cluster across nearly 100 business systems in a telecom operator is one of the most complex and difficult technical validations in the domestic database industry. Dameng has demonstrated its commercialisation capability in large‑scale core deployments. As the "de‑Oracle" trend moves into deep water, Dameng will gain stronger bargaining power in the government and enterprise market. Capital markets should focus on the independent valuation of its core technology assets.

04｜Dameng's Two Database Versions Pass MIIT's Authoritative Shared‑Storage Architecture Test

Recently, in the first half of 2026 batch of "Trustworthy Database" tests organised by the China Academy of Information and Communications Technology (CAICT), both version V8 and V9 of the Dameng Database Management System successfully completed all requirements of the "Basic Capabilities of Shared‑Storage Based Database Architecture" product test. The test systematically validated five dimensions: shared‑storage collaboration, basic functionality, high availability and disaster recovery, compatibility and adaptability, and security capabilities. The test standard, based on the industry standard "Technical Requirements for Shared‑Storage Based Databases", covered a complete capability framework of 28 mandatory items and 11 optional items.

Dameng's shared‑storage database is a truly shared‑data‑file cluster database, where multiple instances read and write the same data file simultaneously. It fully supports load balancing, failover, fused caching, and other features. Its compatibility and security have obtained authoritative certification, benchmarking against internationally leading shared‑storage cluster database products. The successful completion of this dual‑version test marks Dameng's ability to provide high availability and high‑performance stable operation in core business system replacements in critical industries such as finance, telecommunications, and energy.

· DBA Perspective: Financial‑grade data demands harsh requirements for strong consistency and stability when concurrently writing the same data file from multiple nodes – a technical test that many domestic databases dare not attempt. Dameng's V8 and V9 delivered a perfect score from CAICT. DBAs can incorporate shared‑storage cluster capabilities into baseline indicators for "core system replacement", using the high‑availability report to gain stronger negotiating power during technology selection.
· CTO Perspective: The one‑time, dual‑version pass of the highly credible CAICT cluster standard test provides quantitative technical validation of Xinchuang deployment readiness for finance and telecom‑grade scenarios. This certification will accelerate Dameng's coverage in core industry replacements and provide CTOs with stronger technical support for their selection decisions.
· Investor Perspective: The ability to replace "critical core systems" is the core metric for determining whether a domestic database can hold its own in high‑end markets such as finance. Dameng's dual‑version success widens the technology gap between Dameng and many second‑tier domestic competitors. The trend of financial and telecom industry customers increasing their orders with Dameng is likely to become more pronounced in the secondary market.

05｜MongoDB 8.3 Built for AI Speed: Vector Search Adoption Doubles, Revenue Growth Shows Phased Bottlenecks

In Q4 FY2026, MongoDB reported total revenue of $695.1 million, up 27% year‑on‑year, with Atlas revenue accounting for 72% and growing 29%. Atlas vector search adoption nearly doubled year‑on‑year. AI‑powered data services showed positive usage feedback, and although the revenue contribution from AI workloads is still early, platform growth momentum is clear. MongoDB has become a leader in AI‑ready data platforms, with vector search adoption growing rapidly, Atlas revenue share continuously increasing, and its open platform developer advocacy and AI capability integration forming a new competitive moat. The global NoSQL market is projected to grow at a CAGR of 29.85% through 2030, with generative AI, including vector search, directly contributing about 4.5% to that growth. MongoDB now serves over 65,200 enterprises, adding 2,700 new customers last quarter, with average annual revenue per large customer (ARPC) still increasing.

· DBA Perspective: Even in the era of large models, MongoDB remains at the forefront of non‑relational databases in terms of creativity and API evolution. However, the release of 8.3 is not the endpoint, and the financial signal of slowing growth suggests that enterprise willingness to pay for AI is more cautious than expected when measuring ROI. DBAs should carefully evaluate the side effects of AI on existing document model designs and avoid fully vectorising business data prematurely.
· CTO Perspective: MongoDB Atlas's strong cloud growth (+29%) offers enterprises an effective option to replace on‑premises inertia and reduce operational abstraction by moving to the cloud. However, management faces the long‑term contradiction of marginal revenue growth decline and profitability pressure. When selecting technology, decision‑makers should temporarily reduce the proportion of investment in pure vector components and increase the weight of future cost‑reduction assessments for AI products.
· Investor Perspective: Atlas's 29% growth and the doubling of vector search keep MongoDB differentiated in the highly competitive cloud data market, but the leading indicator of total incremental revenue convergence has already flashed a warning. The next inflection point to watch is whether AI workloads can reach double‑digit revenue contribution – once that threshold is crossed, MongoDB's valuation centre will be stably reset.

📌 Issue Summary

News Core Keywords DBA Actions CTO/Decision‑Maker Focus Investor Perspective
OceanBase CEO strategic statement Lakehouse integration, AI agent foundation, multi‑modal storage & retrieval Shift from structured to multi‑modal & federated analysis; design data models for agents Build lakehouse‑integrated foundation; prioritise databases supporting multi‑modal Upgrade valuation logic to multi‑modal foundation & agent support capability
SAP's two acquisitions in one week Dremio, Prior Labs, Iceberg, €1B AI lab Transform from tuner to data federated architect; expand cross‑engine metadata governance Monitor future data stack convergence trends; prepare data governance systems early M&A spree opens exit channels for startups; ecosystem reshaping provides investment windows
Dameng DM9 at Mobile Cloud Conference Five‑in‑one architecture, 256‑node operator deployment Master five‑in‑one mixed workload tuning – the entry ticket to critical industry Xinchuang replacement Five‑in‑one simplifies selection; operator partnership validates technology Full‑scale delivery across 100+ systems proves commercial maturity; government/enterprise market leverage increases
Dameng dual versions pass CAICT shared‑storage test Shared‑storage cluster, five dimensions, 28 mandatory items Include shared‑storage cluster capability as baseline for core system replacement CAICT certification accelerates Xinchuang adoption in finance/telecom Dual‑version success widens tech gap with second‑tier rivals; customer orders likely to tilt
MongoDB FY26Q4 earnings Atlas +29%, vector adoption doubled, revenue growth converging Carefully evaluate AI impact on document model; pause pure vector component investment Cloud DB can offload on‑premises burden; cautiously assess AI ROI Watch if AI workloads cross double‑digit revenue share to trigger valuation re‑rating

HiddenMerit Team Production
Slogan: 绩优隐于内，金石启新程 | Hidden deep. Merit bold. Forge ahead.