AI Reshapes Software Testing Industry: Large Companies Benefit, Small/Medium Companies Face Pressure

Overview

This research report conducted interviews with two uncovered software testing companies (Digital Hearts and Valtes Holdings) to cross-validate the impact of Generative AI on Japan's IT services industry. The core conclusion is that AI is reshaping the industry landscape: On one hand, it causes shrinkage or internalization of small-scale development and simple testing demand, putting short-term performance pressure on professional testing companies; on the other hand, large integrators focused on large-scale system development (such as Fujitsu, NEC, NRI) are not only less negatively affected but also enhance efficiency through AI automation, maintaining strong profitability.

Core views

Digital Hearts (3676.T) faces dual pressure from cyclical peaks and structural AI impact. The company's guidance for FY3/27 operating profit is 27.3 hundred million JPY (+4% YoY), with growth slowing significantly. The main reason is that high-margin game debugging business reached its peak with the release of Switch 2; domestic laboratory utilization is approaching 100%, making further expansion difficult subsequently; meanwhile, its Web application and enterprise testing business (AGEST), while overall stable, sees associated small-scale system development demand being withdrawn internally by clients due to the proliferation of Generative AI, with a decline exceeding expectations, and similar return trends appear in some software testing orders. Valtes Holdings (4442.T) is in the pain period of actively ramping up AI investment, while core business remains robust. Company guidance for FY3/27 operating profit is 9.7 hundred million JPY (+5% YoY). The apparently moderate growth rate is because it included approximately 4 hundred million JPY of AI-related investment (mainly for developing generative AI testing tool "TestScape" and hiring AI talent). If excluding this strategic expenditure, its actual operating profit growth reaches 14%, continuing the double-digit growth trend. Enterprise-level software testing demand is strong, especially in automotive embedded systems and logistics sectors, where the number of active engineers increased by 8% in FY3/26 and average unit price rose by 5%, but talent competition remains fierce. Regarding covered large system integrators, the report believes opportunities brought by AI outweigh challenges. Fujitsu (Buy), NEC (Buy), and Nomura Research Institute (NRI, Buy) mainly undertake large-scale system development projects, where such complex demands are not easily replaced or simplified by AI. Conversely, these companies are actively utilizing AI to achieve automation of development and testing processes, thereby improving productivity and profitability. In comparison, small and medium-sized vendors relying on small projects and standardized testing are more susceptible to erosion from the "insourcing" trend caused by AI.

Analysis framework

The research report adopted a "Read-across" analysis method, which involves analyzing upstream and downstream industry chain or non-covered entities (here, two professional software testing companies) to infer fundamental changes in covered core entities (large SIers). This method is particularly effective in the IT services industry because software testing is often a leading indicator of overall software development demand. In terms of analytical logic, the institution did not view performance fluctuations of individual companies in isolation but attributed them within the macro variable of "Generative AI technology diffusion." By comparing the sensitivity differences of different business models (large-scale vs. small-scale, complex systems vs. simple testing) to AI impact, a structured multi-long/short judgment framework was constructed: AI is not beneficial for all IT service providers; its impact depends on business complexity and irreplacability.

Methodology notes

• Industry Analysis Framework

  Read-across Method

  Refers to deducing and verifying the performance trends of covered companies by analyzing operating data of non-covered companies or industry chain-related companies. In this report, analysts use feedback from professional testing companies to assess the extent of AI impact on the entire IT services industry, filling information blind spots of single-perspective views.

• Industry Analysis FrameworkSubstitution Effect Analysis

  Distinguishing AI Substitution and Complementarity Effects on IT Services

  The report breaks down AI impact into two categories: "Substitution Effect" on simple, standardized tasks (such as basic code writing, simple testing), leading to demand returning to internal enterprises; and "Complementary Effect" on complex, customized tasks (such as large-scale system integration), helping service providers reduce costs and increase efficiency. This distinction is key to judging the investment value of IT companies in the AI era.

Asset mapping & comparison

Structured mapping from thesis to named assets (strengths, weaknesses, peers, risks).

• Fujitsu (6702.T)

  Beneficiary: Focuses on large-scale system development, less affected by demand return caused by AI, and can improve profitability via AI automation

  Strengths

  High business complexity, low AI replacement risk; Actively optimizing development and testing processes using AI

  Comparison

  Less susceptible to negative AI impacts compared to smaller testing vendors like Digital Hearts

• NEC (6701.T)

  Beneficiary: Also a large SIer, logic consistent with Fujitsu, maintain Buy rating

  Strengths

  Mainly large-scale system development, AI empowerment effects visible

  Comparison

  Placed in the beneficiary camp alongside NRI and Fujitsu

• Nomura Research Institute (4307.T)

  Beneficiary: Large consulting and system integrator, AI acts more on efficiency enhancement than substitution for its business

  Strengths

  Strong capability in high-end consulting and complex system integration

  Comparison

  Deeper moat compared to mid-tier SIers

• Digital Hearts (3676.T)

  Losing/Observation: Game debugging peaked + outsourcing returned due to AI, short-term performance pressured

  Strengths

  Switch 2 cycle previously brought high profits; Acquired HUwiZ to expand overseas

  Weaknesses

  High dependence on small-scale development and simple testing, easily replaced by AI

  Comparison

  Compared to Valtes, negative shocks from AI are more direct and lack hedging measures

  Risks

  Switch 2 demand below expectations; Accelerated customer loss due to AI

• Valtes Holdings (4442.T)

  Transformation Observation: Main business stable but needs to bear AI transformation costs, long-term whether able to successfully position for AI testing tools

  Strengths

  Rigid enterprise-level testing demand; Actively investing in AI tool TestScape

  Weaknesses

  AI investment drags profits in short term; Intense talent recruitment competition

  Comparison

  Compared to Digital Hearts, business structure leans more towards enterprise-level, stronger anti-cycle ability

  Risks

  AI tool commercialization below expectations; Rising talent costs erode profits

Key data

• Digital Hearts FY3/27 OP Guidance27.3 hundred million JPYUp 4% YoY, growth slowed due to Switch 2 demand peaking and reduced development demand driven by AI
• Valtes FY3/27 OP Guidance (Adjusted)+14% YoYCore profit growth excluding AI investment, showing enterprise testing main business still maintains double-digit growth
• Valtes AI Investment Plan1.2 hundred million JPYFY3/26-FY3/28 mid-term total plan, used for developing AI testing tool TestScape and talent acquisition
• Digital Hearts Domestic Debugging Sales Growth+11% YoY (FY3/26)Driven by Switch 2 release, but FY3/27 expected not to exceed last year's level, capacity reached bottleneck
• Valtes Software Testing Sales Growth+14% YoY (FY3/26 Q4)Strong enterprise-level testing demand, automotive and logistics sectors were main drivers

Impact & implications

The report believes that Generative AI is accelerating structural differentiation in Japan's IT services industry. For investors, this means "AI concept" cannot simply be equated with broad-based rises in the IT sector. Large integrators possessing complex system design capabilities and able to leverage AI tools to improve delivery efficiency (such as Fujitsu, NEC, NRI) have stronger risk resilience and earnings elasticity. Conversely, companies focusing businesses on low value-added areas or testing/small development fields easily automated by AI may face long-term demand contraction and margin compression unless they can successfully transform like Valtes to become providers of AI tools rather than merely replaceables.

Risks

• Generative AI leads to demand return within enterprises for software testing and development faster than expected
• Large decline in Switch 2 related game debugging demand in FY3/27, with no new projects to follow
• AI testing tool R&D fails or commercialization progress lags, resulting in inability to recover prior investments
• IT talent market competition intensifies, leading to rising labor costs or failed recruitment plans

What to watch

• Actual profit margin improvement data from AI automation in financial reports of large SIers (Fujitsu/NEC/NRI)
• Quarter-over-quarter change trend of small-scale development demand in Digital Hearts AGEsT business
• External customer adoption status and revenue contribution of Valtes AI testing tool "TestScape"
• Update frequency of subsequent software ecosystems for Switch 2 and corresponding debugging demand release rhythm