Report Interpretation

Overview

This webinar-based report examines how generative AI and AI search are changing Japanese consumer marketing and online advertising, with CyberAgent as the primary equity subject. Goldman Sachs believes the discussion supports a more constructive long-term view of CyberAgent’s search-ad opportunity, but keeps a Neutral rating because meaningful earnings from AI-search and GEO activities remain some time away.

Core views

Goldman Sachs hosted a September 17 webinar with CyberAgent’s Internet Advertising Division and istyle’s AI Innovation Strategy Office to assess how generative AI is reshaping consumer touchpoints and advertising. The institution concluded that advances in AI create substantial earnings opportunities for the internet-advertising industry and could make firms’ AI-utilization capabilities increasingly important to marketing outcomes, irrespective of company size. A key conclusion was that generative AI need not shrink traditional search advertising: panelists said conversational AI and conventional search are being used in parallel, increasing searches overall, while AI can improve the precision of search-ad delivery. CyberAgent’s position rests on AI capabilities, the scale of its data and relationships with media, according to the webinar discussion. The company began investing in AI around 10 years ago and applies proprietary AI across performance-based advertising, which it says accounts for 90% of the internet-advertising market. This enables large volumes of one-to-one ad creatives and landing pages. CyberAgent launched GEO Lab in July 2025, became an official launch partner for ChatGPT ads in Japan when they were released in June 2026, and established its AI Search Marketing Division in June 2026. That unit spans search ads, generative engine optimization (GEO) and generative-AI search from strategy through operations; headcount is planned to rise from about 150 to 250 by FY9/27. The report describes an evolving consumer journey. CyberAgent’s survey found that generative-AI search usage in Japan exceeded 50% in August 2026, initially led by teenagers and people in their twenties but also reaching a majority among people in their forties. Consumers increasingly consult AI rather than independently searching and comparing. Yet for categories such as cosmetics, where consumers still want to test products physically, AI research can be followed by greater need for customer service. The panel’s core point is that users still visit and verify actual websites, and that Google search queries recently hit a record high rather than being displaced by conversational AI. For advertisers, the report identifies a broader toolkit beyond search ads and SEO: GEO aims to have AI mention or recommend a brand, while AI search and ChatGPT ads create new ways to reach consumers. GEO may be especially useful for addressing potential customers who have not yet formed a specific product intent. In traditional search advertising, AI’s more precise interpretation of intent can support different ads for the same keyword, increasing demand for one-to-one creative production and delivery design. Return on investment remains the central criterion, but both companies said effectiveness measurement is still immature: links between sales and AI dialogue or GEO remain under verification, and istyle’s combined pre- and post-purchase data can make causality difficult to establish. The report also presents a nuanced consumer-goods implication. Large brands may be more readily recommended by AI because they possess extensive reliable historical information and larger advertising budgets. However, small and medium-sized brands can also compete if they provide detailed, distinctive content tied to consumer needs; AI recommendations may favor unique value and information rooted in actual experiences. Companies’ response speed is highly company-specific, but the panelists identified urgency among senior executives and marketing leaders, plus well-prepared and accessible brand data, as important triggers for adoption. For CyberAgent specifically, the company characterizes AI-search advertising and GEO as a once-in-two-decades shift. It expects internet advertising to take a larger share of total ad spending, even though aggregate advertising expenditure is linked to GDP. It sees scope for higher ad effectiveness from context-aware ChatGPT ads and for traditional search-ad monetization: ads currently appear on about 20% of searches, a ratio it believes could rise as AI improves ad accuracy. The new unit is intended to create a multi-layered earnings model by adding planning-consulting fees to traditional agency commissions. Goldman Sachs acknowledges that AI-search and GEO monetization is at an early stage and likely will not make meaningful earnings contributions immediately. Nevertheless, it views CyberAgent’s technology and rapid launch of a specialized unit as appropriate responses to structural change. The institution believes investor concerns about traditional search-ad contraction may be eased by evidence of parallel search usage and room for AI-driven accuracy gains. It remains Neutral because investors are likely to continue evaluating the medium-term online-advertising outlook, while noting that search ads represent about one-third of CyberAgent’s internet-advertising sales and that the outlook presented for this segment was encouraging.

Analysis framework

Goldman Sachs uses webinar commentary from CyberAgent and istyle to connect changes in consumer search behavior with advertiser demand, ad-delivery mechanics and CyberAgent’s operating response. It then assesses the implications for CyberAgent’s earnings opportunity and values the company using a sum-of-the-parts approach based on FY9/27.

Methodology notes

• Industry AnalysisUpstream-Midstream-Downstream Transmission

  Consumer-search and advertiser-marketing transmission analysis

  The report traces how AI-assisted consumer search changes brand data needs, creative production and advertising demand, and then links those changes to CyberAgent’s agency and consulting revenue opportunity.

• Valuation methodsSOTP (Sum-of-the-Parts) Valuation

  Sum-of-the-parts valuation

  Goldman Sachs states that its ¥1,580 12-month target price for CyberAgent is based on SOTP using FY9/27 as the valuation base year.

Asset mapping & comparison

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

• CyberAgent (4751.T)

  Primary covered company positioned to benefit from AI-enabled search advertising, GEO and AI marketing services.

  Strengths

  Long-standing AI investment, data scale, media relationships, one-to-one creative capability and an early specialized AI Search Marketing Division.

  Weaknesses

  Meaningful AI-search and GEO earnings contributions are still expected to take time to materialize.

  Comparison

  The report notes that both large and smaller brands can use AI marketing, while CyberAgent’s position is assessed through its advertising capabilities rather than a direct agency-peer comparison.

  Risks

  Game-title performance, ABEMA and IP-business profit trends, internet-advertising share gains, and the timing of DX-business profit contribution could differ from estimates.

Key data

• Generative AI search usage in JapanExceeded 50%CyberAgent survey result for August 2026.
• CyberAgent AI Search Marketing Division headcountc.150 to 250Planned expansion by FY9/27.
• Performance-based advertising share of internet-advertising market90%CyberAgent describes this as the market segment across which it deploys proprietary AI.
• Ads displayed when a search is conductedc.20%CyberAgent sees room for the ratio to rise with AI-driven improvements in ad accuracy.
• Search ads as share of CyberAgent internet-advertising salesc.1/3Goldman Sachs highlights this segment as central to the company’s advertising outlook.
• Target price¥1,58012-month target price based on FY9/27 SOTP valuation.

Impact & implications

Goldman Sachs believes generative AI can expand the addressable opportunity for online-ad agencies by improving intent matching, increasing creative requirements and adding GEO and AI-search services. For CyberAgent, the report sees potential for search-ad growth and additional consulting fees, although the new AI-search revenue streams remain immature and the medium-term market outlook still requires investor assessment.

Risks

• Game-business earnings may exceed or fall short of estimates depending on new-title performance and momentum in existing major titles.
• Media and IP segment profit may vary with the pace of ABEMA improvement and trends in IP businesses such as anime production.
• Internet-advertising share gains could differ from expectations.
• Meaningful DX-business profit contribution could arrive earlier or later than expected.

What to watch

• The pace at which AI-search advertising and GEO generate meaningful earnings contributions.
• CyberAgent’s progress in expanding its AI Search Marketing Division from about 150 to 250 staff by FY9/27.
• Whether AI-driven advertising accuracy raises the share of searches that display ads from the current roughly 20%.
• Evidence linking AI dialogue or GEO activity to sales and advertising ROI.
• The medium-term outlook for Japan’s online-advertising market and CyberAgent’s share gains.