On December 3rd, at the "Jazzyalpha Gravity Annual Summit", Zhu Jianyong, Co-founder and Vice President of KeenData, was invited to attend the roundtable forum themed "Where are the Next Focus Points for AI Infrastructure Construction". From three dimensions of business logic, technological evolution and industrial implementation, he systematically elaborated on the strategic value and future direction of data infrastructure in the AI era.

The Core Bottleneck of AI Implementation in China: "Data Ready"

The development of artificial intelligence is driven by the synergy of three core elements: computing power, algorithms and data. None of them can be dispensed with, and only when they form a joint force can artificial intelligence be truly put into application.

In the field of computing power, China has built a competitive computing power system through forward-looking layout and vigorous promotion of large-scale computing center construction, ranging from supercomputing centers to intelligent computing centers. At the algorithm level, the domestic scientific research and industrial circles have also achieved many breakthroughs.

Zhu Jianyong pointed out that as competition in the fields of algorithms and computing power intensifies, data has become the biggest shortcoming for the large-scale implementation of AI—that is, how to make massive, diverse and heterogeneous data truly usable, easy to use and efficiently serve AI. Currently, the industry is mainly facing three core challenges:

Difficulty in activating historical data: The PB-level data accumulated by the Internet and mobile Internet over the past 20 years urgently needs to be reconstructed in terms of storage, computing, circulation and utilization efficiency.

Difficulty in integrating multi-source heterogeneous data: From structured databases to semi-structured logs and then to multi-modal data such as audio and video, traditional data platforms are difficult to support the unified processing required by AI.

Insufficient support for large-scale implementation: Existing data warehouse and platform architectures cannot meet the large-scale, high-concurrency and real-time engineering needs of AI scenarios.

The key to solving these problems lies in building a Data & AI integrated data infrastructure.

The Core Engine of Data Infrastructure: Data & AI Integrated Platform

Data infrastructure is a new type of infrastructure aimed at releasing the value of data elements, covering hardware, software, standards, specifications, systems and mechanisms. It is designed to provide full-life-cycle service capabilities from data collection, convergence, processing, circulation to application, operation and security.

Among them, the Data & AI integrated platform serves as its technical foundation and core engine.

For enterprises, the essence of this platform is to connect the entire workflow of "data engineering" and "AI engineering", realizing the two-way empowerment of "Data for AI" and "AI for Data". It is not only an upgraded form of traditional big data platforms, but also a reconstructed data processing paradigm for the AI-Native era, emerging as the "core production tool" for intelligent transformation.

In the artificial intelligence era, data platforms are subject to new requirements: they need to connect upward with foundation models to provide strong support for scenario-specific model optimization and innovative application implementation; they also need to undertake computing power resources downward, fully release the advantages of computing power, and realize the optimized scheduling and efficient utilization of computing resources.

Based on this, KeenData proposes that data infrastructure needs to be systematically constructed from four technical levels:

Unified storage and computing layer: Supports efficient access of multi-modal data and dynamic scheduling of GPU computing power.

Engineering integration layer: Achieves full-life-cycle integration of data engineering and AI engineering (collection → cleaning → modeling → optimization → deployment).

Data governance layer: Combines AI capabilities to automatically build high-quality and iterable training datasets.

Resource management layer: Establishes a business scenario-oriented data asset operation system to improve the efficiency of data resource utilization.

It is worth emphasizing that as the core supporting platform, KeenData Lakehouse is not merely a software platform, but a core competitiveness that enterprises must continuously iterate. Its essence is a comprehensive system deeply integrating "advanced technology + mature software + AI engineering". It not only solves the problem of technology implementation, but also shapes a new enterprise management method through the core model of "centralized management and decentralized empowerment", serving as the best practice carrier for the deep integration of software into enterprise management.

It connects technical engineering, data management, AI operation and business collaboration, helping enterprises establish a new collaboration mechanism based on data and AI needs. Ultimately, it drives the all-round digital and intelligent transformation of organizations from management models and business processes to value creation, turning transformation from a slogan into sustainable growth results.

Rooted in China, Going Global: Building a World-Class Data & AI Data Infrastructure

KeenData has been deeply engaged in the field of Data & AI integration for more than six years, building the AI-Native Data & AI integrated platform KeenData Lakehouse. Integrating the "AI-Native" design concept, the platform independently developed the AI-in-Lakehouse intelligent-driven architecture, connecting the entire workflow from data engineering → model training/inference → Agent factory → intelligent applications. With its platform capabilities of "trustworthiness + intelligence + system", it promotes the new "Data & AI" infrastructure, supporting large organizations to move from data-driven to intelligence-driven. The platform breaks through the traditional architecture of separated data and AI, unifying the lake-house integrated engine, OLAP data governance and AI technologies to form a streamlined and efficient All-in-One technology solution. The independently developed multi-modal computing engine completes data cleaning to result analysis in a single pipeline, increasing GPU inference throughput several times. Combined with KMI inference acceleration, model quantization and Unity Catalog, it realizes cross-modal intelligent governance.

Supported by methodology + technology + products + practices, KeenData has successfully served nearly 200 large organizations in more than 20 industries including manufacturing, industry, energy, finance and retail, tailoring data infrastructure and data bases adapted to their business needs with remarkable implementation results. Meanwhile, KeenData actively responds to national policies related to Digital China and data elements, deeply participates in the planning and construction of government-side data infrastructure and trusted data spaces, undertakes projects of trusted data spaces and pilot demonstration zones in multiple key cities in China, fully implementing its core capabilities in both government and enterprise scenarios, and continuously expanding the path for releasing data value.

Relying on mature experience in Data & AI data infrastructure construction and core technologies accumulated in China, KeenData has proactively expanded into overseas markets, exporting advanced domestic technologies, products and methodologies to countries and regions around the world. It helps them build core capabilities for development in the AI era and promotes the development of local artificial intelligence industry and digital economy. KeenData has established profound cooperative relations with customers in many countries around the world including Saudi Arabia, Singapore, South Africa, Japan, Malaysia and the Philippines, and works hand in hand with global partners to build new industrial advantages, contributing Chinese wisdom and strength to the development of the global digital economy.