What GraphAI does

GraphAI Co., Ltd. is a South Korea-based deep tech startup founded in 2022 by KAIST Professor Min-soo Kim that provides a full-stack enterprise AI data infrastructure platform called the Akasic Platform. The platform is composed of four integrated modules: AkasicIN for data ingestion and preparation (OCR, parsing, normalization); AkasicDB, a unified graph-vector-relational DBMS distributed as a PostgreSQL extension featuring a proprietary Traversal-Join operator and Triple Storage architecture; AkasicON, an ontology-based governance layer with What-if simulation; and AkasicAI, an agentic decision-making engine that combines vector, graph, and SQL retrieval to produce validated, evidence-grounded recommendations. The underlying technology stems from more than a decade of graph data processing research at KAIST's InfoLab and has been validated at ACM SIGMOD 2026, VLDB 2025, and ACM KDD 2025, with benchmarked performance claims of up to 142x faster than leading vector DBs and 99.5% search accuracy. GraphAI targets large enterprises in defense, financial services, telecommunications, and manufacturing, with on-premises and air-gapped deployment support for security-sensitive buyers.

The company generates revenue primarily through enterprise software licenses, delivered via direct enterprise field sales with PoC-to-commercial-contract conversion, plus cloud-hosted deployments on AWS EC2, Azure VM, and GCP Compute Engine; AWS Marketplace listing is planned for early 2026. Pricing is not publicly disclosed and is handled through a consultative 'Request a Demo' motion, consistent with long enterprise sales cycles. Named customers and design partners include Hanwha Ocean, Hanwha Systems, LG U+, KT Cloud, KB Securities, the Korea Institute of Defense Science and Technology (ADD), and Hankook Tire, alongside consortium partnerships under KISED's InnoWave League and LG U+'s Shift open innovation programs. The company has raised approximately KRW 3.6 billion (~USD 2.7M) in cumulative funding through a 2022 seed round and an October 2024 pre-Series A round, and maintains dual R&D centers in Daejeon (headquarters, adjacent to KAIST) and Seoul (Gangnam).

Name

GraphAI

Legal name

GraphAI Co., Ltd.

Website

https://graphai.io

Company type

Private

Founded year

2022

Operating status

Operating

Headcount range

1–10 employees

Short description

GraphAI (founded 2022, Daejeon) builds the Akasic Platform — a full-stack enterprise AI data infrastructure unifying graph, vector, and relational databases with an ontology and agentic decision engine — selling on-premises and cloud-deployed software licenses to large enterprises in defense, finance, telecom, and manufacturing.

Ownership category

akta.pro rank

Product category

Enterprise AI Data Infrastructure

NAICS

Software Publishers (5132)

SIC

Services-Prepackaged Software (7372)

akta.pro primary industry

Enterprise AI Data & Knowledge Platforms (Vector Databases, Knowledge Graphs) (HDAEANAH)

akta.pro secondary industries

Graph Databases (HDAEAAAE), Data Engineering, Integration & ETL/ELT Services (BPAEAHAB), End-to-End Enterprise AI Platforms (MLOps & Model Lifecycle Management) (HDAEANAA)

Keywords

• Unified database engine
• AI data infrastructure
• Graph database platform
• Vector database software
• Enterprise AI platform

Headquarters

HQ city

Daejeon

HQ country

South Korea

HQ region

Asia

Offices2 records

Markets served

GTM type

B2B

Offering type

Software

Cost components

Technology or R&D, Personnel, Marketing or Sales, Operations, Infrastructure

Revenue model

1. Enterprise Software Licenses (Akasic Platform): GraphAI sells its Akasic platform (AkasicIN, AkasicDB, AkasicON, AkasicAI) as enterprise software to large organizations. Given the enterprise sales motion, long sales cycles, and PoC-to-commercial-contract conversion model, revenue is generated primarily through software licenses, likely annual or multi-year contracts. The platform is delivered as an on-premises or private cloud solution, with support for air-gapped environments — typical of enterprise software license structures.
2. AkasicDB Cloud/Hosted Deployment: AkasicDB is delivered as a PostgreSQL extension (commercial software). The company offers a cloud-based playground (db.akasic.cloud) for evaluation and targets cloud deployment on AWS EC2, Azure VM, and GCP Compute Engine, suggesting a potential usage-based or infrastructure-backed revenue stream for cloud deployments.

Pricing tiers

Go-to-market motion3 records

Distribution channels4 records

Marketing channels6 records

Core offering

GraphAI develops and sells the Akasic Platform, a full-stack enterprise AI data infrastructure solution that integrates data ingestion, a unified graph-vector-relational database engine, ontology-based governance, and an agentic AI decision-making engine. The platform is delivered primarily as on-premises or air-gapped enterprise software, with AkasicDB offered as a commercial PostgreSQL extension for hybrid and cloud deployments.

Product overview

GraphAI offers the Akasic Platform (아카식 플랫폼), a full-stack enterprise AI data infrastructure solution comprising four integrated modules: AkasicIN for data ingestion and transformation into AI-ready standardized formats; AkasicDB as the unified Graph-Vector-Relational DBMS engine; AkasicON for ontology-based governance and policy enforcement; and AkasicAI as the agentic AI decision-making engine. The platform implements a co-optimized architecture where data preparation, unified storage, governance, and AI decision-making are designed to work together within a single pipeline. AkasicDB v1.0.0 was released in June 2026 as a PostgreSQL extension.

Differentiator

Problem solved

Functional benefit

Brands

• Akasic Platform: Full-stack enterprise AI data solution comprising AkasicIN, AkasicDB, AkasicON, and AkasicAI modules for enterprise AI pipeline from data preparation to agentic decision execution
• AkasicIN
• AkasicDB
• AkasicON
• AkasicAI

Products and services

• AkasicIN Enterprise AI data preparation platform that collects, parses (including OCR), normalizes, and enriches enterprise data (documents, logs, system data, DB tables, images) into standardized AI-ready formats. Supports on-premises and air-gapped environments with full metadata and data lineage management. For enterprise AI teams in defense, finance, manufacturing, and telecom.
• AkasicDB Unified Graph-Vector-Relational DBMS distributed as a commercial PostgreSQL extension that integrates vector, graph, and relational database functions into a single engine. Features Triple Storage architecture, Traversal-Join (TJ) operator for unified query processing, and hybrid search combining semantic search with graph traversal and SQL conditions. Delivers up to 142x faster performance than global leading vector DBs with 99.5% search accuracy at scale.
• AkasicON Ontology-based data governance platform that defines data semantics, entity relationships, operational policies, and approval workflows within a unified framework. Enables What-if simulation to assess impact before execution, and applies governance rules automatically down to the DB layer. Establishes Entity, Relation, Policy, and Action definitions for AI control and operational accountability.
• AkasicAI Ontology-based AI decision-making engine that goes beyond generating plausible answers to proposing validated action plans. Automatically combines vector, graph, and SQL search methods based on query intent, traces causal relationships through data structures, performs RelationRAG for complete aggregation queries, and validates action plans through What-if simulation before presenting recommendations.
• Akasic Platform Full-stack enterprise AI data platform comprising four integrated modules: AkasicIN (data preparation), AkasicDB (unified DB engine), AkasicON (ontology governance), and AkasicAI (agentic decision engine). Connects the complete AI data pipeline from data ingestion to validated decision execution within a single co-optimized stack.

Quantifiable outcome

• AI response accuracy improved by up to 78% vs. existing RAG
• +7 more outcomes

Named customers7 records

Segments5 records

Ideal customer profiles4 records

Technology focussed Yes

API detail

Has API

Yes

API docs

API detail

Core technology

AI maturity

App detail

Integration5 records

AI capability10 records

Feature12 records

Partnerships

Eight partnerships are on record, tiered core, major and minor.

• Amazon Web Services (AWS)coreTechnology or Integration · 5 December 2025GraphAI obtained AWS Partner Software Path certification by passing the Foundational Technical Review (FTR) with its GraphON and VectorON solutions. Plans to list AkasicDB and related products on AWS Marketplace by early 2026 for global distribution. This is a strategic technology and GTM partnership enabling worldwide enterprise adoption through AWS's global cloud infrastructure and marketplace ecosystem.
• Hanwha OceancoreStrategic or Co-development Partner · 3 December 2025GraphAI and Hanwha Ocean formed a consortium under KISED's InnoWave League program to develop 'Process Risk Prediction AI'. The consortium won the Excellence Award. Hanwha Ocean is both a strategic partner and a customer, with GraphAI providing AI data infrastructure solutions for defense and manufacturing applications.
• LG U+coreStrategic or Co-development Partner · 25 November 2025GraphAI was selected for LG U+'s AI open innovation program 'Shift' Season 2. The partnership includes collaboration with LG U+ and global partners (AWS, Salesforce) on product refinement, MWC participation support for global market entry, and potential business collaboration with telecom operators. LG U+ is also a customer using GraphAI's AI infrastructure solutions.
• NVIDIAcoreTechnology or IntegrationGraphAI was selected as a member of NVIDIA Inception, a startup nurturing program operated by NVIDIA. The NVIDIA Inception program provides GraphAI with access to NVIDIA's technology ecosystem, GPU technology resources, and credibility within the global AI community. GraphAI's technologies (GFlux, FlexGNN, AkasicDB) leverage GPU computing for high-performance graph and AI workloads.
• KAIST InfoLabcoreStrategic or Co-development PartnerGraphAI originated from KAIST's InfoLab, which has produced world-class research in graph data processing for over a decade. CEO Minsoo Kim is a KAIST Professor and leads ongoing research collaboration between GraphAI and KAIST, producing published research at SIGMOD, VLDB, and KDD conferences. This academic-industry collaboration is the foundation of GraphAI's technology differentiation.
• Hanwha SystemsmajorStrategic or Co-development PartnerGraphAI collaborates with Hanwha Systems on enterprise AI data infrastructure solutions for defense applications. Hanwha Systems is a strategic defense technology partner and customer.
• KT CloudmajorStrategic or Co-development PartnerGraphAI collaborates with KT Cloud on AI data infrastructure solutions for telecommunications and cloud environments.
• SalesforceminorGTM or Marketing PartnerSalesforce is listed as a global partner in the context of the LG U+ Shift program collaboration, where GraphAI plans to work with LG U+ and global partners including Salesforce to refine its products and accelerate global market entry.

Scale indicators6 records

Recent moves6 records

Expansion highlights6 records

Direct peers

• Neo4j: Leading graph database vendor with vector search capabilities added to its core graph engine. Directly comparable to GraphAI's AkasicDB on graph database functionality and increasingly on hybrid vector-graph search, though Neo4j does not integrate relational DBMS into a single engine.
• Pinecone: Managed vector database purpose-built for AI applications and similarity search. Comparable to GraphAI on the vector DB component of AkasicDB, but lacks integrated graph and relational DB capabilities and the agentic/ontology governance layer.
• TigerGraph: Enterprise graph database platform focused on analytics over connected data. Comparable to GraphAI's graph DB and graph-relational integration ambitions (Chimera), targeting similar enterprise use cases in finance, defense, and manufacturing.
• Weaviate: Open-source vector database with hybrid search capabilities (vector + BM25 keyword). Directly comparable to AkasicDB's vector search and hybrid retrieval functionality for RAG workloads, though without the integrated graph and relational DB layers.
• ArangoDB: Multi-model database natively supporting graph, document, key-value, and search models in a single engine. Closest architectural analog to GraphAI's unified DB vision, though without the agentic AI and ontology governance layers or enterprise RAG focus.
• Qdrant: Open-source vector database with strong performance benchmarks and growing enterprise adoption. Comparable to AkasicDB on vector search and RAG workloads, competing in the same enterprise AI infrastructure category.
• Memgraph: Real-time graph database with streaming and in-memory capabilities. Comparable to GraphAI's graph DB offering, with similar enterprise focus on operational graph analytics and AI applications.

Broad incumbents

• Databricks: Unified data and AI platform with vector search (Mosaic AI Vector Search), graph capabilities (Spark GraphFrames), and ML/AI tooling. A broad incumbent that competes with GraphAI across multiple layers of the AI data stack and has the resources to build or acquire integrated capabilities.
• Snowflake: Cloud data platform expanding into AI workloads with vector search and Cortex AI features. Broad incumbent with strong enterprise customer base that competes for the same AI data infrastructure budget GraphAI targets.
• MongoDB: Document-oriented database with Atlas Vector Search and multi-model capabilities. Competes broadly in the same enterprise database and AI infrastructure category as GraphAI, with much greater scale and established enterprise sales motion.

Market position

Strengths4 records

Weaknesses4 records

Competitive moat5 records

Key risks6 records

Key highlights6 records

Customer concentration

• LinkedInlinkedin.com/company/graphai-io

Compliance2 records

Revenue estimate

Valuation estimate

Number of profiles

Profiles3 records

Funding overview

Funding rounds3 records

Investors9 records

Funding detail is available on the Subscription and Enterprise plan.Contact sales →

M&A

Investments

M&A and investment is available on the Subscription and Enterprise plan.Contact sales →

Frequently asked questions about GraphAI