ABEJA

ABEJA / 株式会社ABEJA

A Japanese enterprise-AI software company that provides a platform for developing, deploying, and operating AI applications in business processes.

TSE: 5574 (Growth market)
Profile as of September 14, 2026
#127 in the AsiaAI Tech Index
How the ranking works →

ABEJA is a small but profitable Japanese enterprise-AI implementation company whose revenue is accelerating as customers deploy generative AI. Its central constraint is operating in a competitive, service-intensive layer where major integrators and cloud vendors can substitute for its offering.

Key figures

Revenue FY2025: 3.585 billion Japanese yen
Revenue growth FY2025: 29.6%
Operating margin FY2025: 12.4%
Net income FY2025: 448 million Japanese yen
R&D spend FY2025: Not disclosed
Capital expenditure FY2025: Not disclosed
Market capitalization 2026-09-14: 28.894 billion Japanese yen Estimate
Gross margin FY2025: 62.4%
FY2026 revenue guidance FY2026: 4.550 billion Japanese yen
Q1 revenue growth Q1 2026: 55.9% year over year

Overview

ABEJA operates from Tokyo as a domestic enterprise-software vendor, supplying the ABEJA Platform to develop, deploy, operate, and govern AI applications in business processes. Founded in 2012, it is an independent listed company serving Japanese enterprises in sectors including retail, manufacturing, logistics, infrastructure, and financial services.

The company depends on customer data access, public-cloud or customer-managed infrastructure, third-party foundation models, and enterprise-system integration. It does not manufacture chips, servers, power equipment, networking equipment, or edge devices. Instead, it acts as an implementation partner and operating layer for specific AI deployments, competing with systems integrators, domestic AI consultancies, and cloud partners. If ABEJA stopped shipping, its customers would lose a Japanese-language implementation partner and operating layer for specific AI deployments, but most projects are likely replaceable over time by systems integrators, domestic AI consultancies, cloud partners, or internal engineering teams.

The AI angle

ABEJA's economic exposure to AI comes from implementation, operation, and application work on its ABEJA Platform rather than from semiconductor design, cloud infrastructure ownership, or foundation-model training. AI demand reaches revenue through enterprise spending on generative-AI and large-language-model deployment, particularly projects that connect models to operational data, workflow systems, and existing enterprise processes.

The company attributes its FY2025 growth and ongoing FY2026 expansion to demand for LLM-related projects and broader AI utilization through its platform. It matters most in the enterprise applications and cloud and AI platforms layers because its revenue comes from putting AI into customer workflows and operating those systems after deployment. ABEJA depends on customer data access, public-cloud or customer-managed infrastructure, third-party foundation models, and enterprise-system integration. Named suppliers of cloud compute, GPUs, foundation-model APIs, or data-center capacity are not disclosed. It does not manufacture chips, servers, power equipment, networking equipment, or edge devices.

Technology and moat

ABEJA's differentiator is not proprietary silicon, hyperscale cloud infrastructure, or a frontier foundation model. Its practical advantage is accumulated experience integrating machine learning and generative AI into enterprise processes, handling operational data, and deploying systems intended for mission-critical use. The ABEJA Platform is positioned as the common software foundation for development, deployment, operation, and continued improvement of customer AI systems.

This is difficult to copy only to the extent that a customer deployment contains reusable data models, workflow logic, governance controls, operating practices, and customer-specific domain knowledge. Such switching costs can matter for an installed customer, especially where AI is embedded in retail operations, manufacturing workflows, or enterprise decision processes. They are weaker than infrastructure lock-in because a customer can procure similar project delivery from another systems integrator or build internally. Its intellectual-property position is not quantifiable from public patent-count data because its current number of granted patents and patent applications is not disclosed.

Five-pillar assessment

Scale and market position: A small Japanese enterprise AI software vendor with FY2025 revenue of 3.585 billion Japanese yen and about 133 employees.
Technology and R&D: Defensibility rests on the ABEJA Platform, accumulated deployment experience, customer data integration, and operational workflow knowledge rather than proprietary models or silicon.
Supply-chain centrality: Partners with SAKURA Internet and serves Japanese enterprises like Mitsubishi Heavy Industries, while relying on undisclosed cloud and foundation-model providers.
Financial momentum: Accelerating revenue growth, with FY2025 revenue up 29.6% and FY2026 first-quarter revenue up 55.9%, driven by LLM-related demand.
Governance and quality: Independent listed company on the Tokyo Stock Exchange with professional management, though detailed ownership concentration is not disclosed.

Supply chain and relationships

ABEJA depends on public-cloud or customer-managed infrastructure, third-party foundation models, and enterprise-system integration. Named suppliers of cloud compute, GPUs, foundation-model APIs, or data-center capacity are not disclosed.

Its customers are primarily Japanese enterprises in sectors including retail, manufacturing, logistics, infrastructure, and financial services. Customer names and customer-level revenue shares are not disclosed in the cited company materials, though Mitsubishi Heavy Industries is a named customer for a verification project involving autonomous functions using AI agents for unmanned aircraft. Third-party reporting says the company serves more than 300 enterprise customers, but this number is not company-reported. The company's public positioning identifies enterprise customers but does not quantify revenue dependence on any named customer, making it difficult to assess whether a small number of large projects drives revenue volatility.

Customers

Mitsubishi Heavy IndustriesVerification support for autonomous functions

Partners

SAKURA InternetCEO serves as an outside director of ABEJA

Competitors

PKSHA TechnologyInferredCompetes in Japanese-language AI software and enterprise implementation
ExaWizardsInferredCompetes in enterprise AI and digital-transformation offerings
Laboro.AIInferredCompetes for custom enterprise AI development
AVILENInferredCompetes in Japanese enterprise AI implementation
NTT DATAInferredCompetes for enterprise AI integration and managed delivery
FujitsuInferredCompetes in enterprise AI systems and transformation projects
NECInferredCompetes in enterprise AI and industrial systems
AccentureInferredCompetes for large AI transformation and cloud integration

Geopolitics and risk

ABEJA does not publicly report direct dependence on advanced-chip exports or semiconductor manufacturing equipment. Its export-control exposure is indirect through cloud, model, and GPU availability.

The company is headquartered in Tokyo and appears concentrated in Japanese enterprise customers, so domestic GDP, wage inflation, and hiring conditions directly affect its addressable spending and cost base. It must also navigate Japanese AI and personal-data regulations, as enterprise deployment requires customer confidence in data handling, governance, model-risk controls, and sector-specific compliance. Failures can slow adoption or create liabilities. Furthermore, because the company handles or integrates with customer operational data and AI workflows, security incidents or data leakage could be potentially damaging to customer trust and contract renewal.

Risk matrix
Risk	Severity	Why it matters
Enterprise IT-spending cyclicality	High	Delayed digital-transformation budgets can reduce bookings and revenue conversion.
Competition from large integrators	High	Major IT vendors and cloud providers can bundle AI work into broader contracts.
Cybersecurity and data leakage	High	Security incidents involving customer operational data could damage trust and renewals.
Talent retention	High	A small number of senior AI engineers can materially affect delivery capacity.
Customer concentration	Medium	Lack of disclosure makes it difficult to assess revenue volatility from large projects.
Dependence on third-party AI	Medium	Deployments require access to external models and cloud platforms with undisclosed terms.
Japanese AI regulation	Medium	Failures in data handling or model-risk controls can slow adoption or create liabilities.
Japan economic exposure	Medium	Domestic GDP and hiring conditions directly affect its addressable spending and cost base.

Governance and ownership

ABEJA is an independent Japanese listed company, not state-owned and not state-influenced on the public evidence cited here. Its Tokyo Stock Exchange Growth listing subjects it to Japanese exchange disclosure and corporate-governance requirements.

It is led by CEO and representative director Yousuke Okada. The company's directors include Motohiro Koma as COO, Kazuki Hanabusa as CFO, Naoki Tonogi as CSO, and outside directors including Koji Asano, CEO of SAKURA Internet. The company's listed status and director disclosures provide a basic governance framework, but public ownership concentration, beneficial ownership detail, board committee structure, and independent-director ratio are not disclosed in the cited evidence.

What to watch

Whether ABEJA reports FY2026 revenue at or above its revised 4.550 billion Japanese yen forecast.
Whether operating margin remains above FY2025's 12.4% as the company expands LLM-related projects.
Whether customer concentration becomes quantifiable through future securities reports.
Whether recurring platform and operating revenue rises relative to one-off implementation revenue.
Whether the Mitsubishi Heavy Industries verification support develops into a material production deployment.
Whether ABEJA discloses named cloud, foundation-model, or semiconductor partners.

Recent developments

August 31, 2026Announced verification support commissioned by Mitsubishi Heavy Industries for unmanned aircraft.
July 15, 2026Reported nine-month FY2026 revenue of 3.369 billion Japanese yen, up 25.1% year over year.
January 14, 2026Reported FY2026 first-quarter revenue of 1.198 billion Japanese yen, up 55.9% year over year.
October 15, 2025Reported FY2025 revenue of 3.585 billion Japanese yen, up 29.6% year over year.

Coverage on AsiaAI.FYI

We haven't written about ABEJA yet. Subscribe below to get it when we do.

Guides that cover ABEJA

Japan AI Industry Map

About this profile

Compiled with AI-assisted research from company filings, market data, and published reporting as of September 14, 2026, then reviewed by AsiaAI.FYI. Figures marked Estimate are not company-reported. Check primary filings before relying on any number.

Confidence: B. Financials and corporate identity are well sourced, but R&D expense, capital expenditure, patent count, customer concentration, and supplier relationships are not fully disclosed.

Main sources: Company corporate disclosures; Tokyo Stock Exchange disclosure documents; FY2025 results presentations; Market-data sources.

All 39 sources

Position in the AI stack

Provides an enterprise software and implementation platform used to develop, deploy, operate, and govern AI applications.

Legal name: ABEJA, Inc.
Headquarters: Tokyo, Japan
Founded: 2012
Founders: Yousuke Okada
Leadership: Yousuke Okada / 岡田陽介, CEO and representative director
Employees: 133
Ownership: Publicly listed
Listings: TSE: 5574 (Growth market)
Business: Enterprise AI software and implementation platform
Sector: AI Software & Models
Website: abejainc.com/en