High-Definition 3D Data Built for Autonomous Driving Is Rewriting the Rules of Game Development — Tokyo Game Show Exhibit Opens a New Entertainment Frontier

Executive Summary

Envalith had the opportunity to interview President & CEO Yoshimura. Dynamic Map Platform has entered a new phase, extending the high-definition 3D data (HD maps) and 3D Gaussian Splatting (3DGS) assets it has built for autonomous driving into the gaming and entertainment space. The company's objective in exhibiting at the Tokyo Game Show is to surface latent demand for a licensing-model business that leverages existing data assets as-is, and to generate leads. Entertainment stands out as an add-on domain, where broadening the applications of the company's high-definition 3D data can open up new customer segments and licensing opportunities. In the core business, Data for AI (enterprise licensing) and overseas data generation projects are expected to serve as growth drivers from FY3/27 through FY3/28.

Message from the Company

Key Discussion Points

• Tokyo Game Show Exhibit and Entertainment Market Development

  The company positions its Tokyo Game Show exhibit as a venue for building relationships with game developers and generating leads. Entertainment currently accounts for a negligible share of company-wide revenue, but beyond an existing delivery to a French indie game studio, concrete delivery projects are already underway. A build-up of inquiries in gaming would increase utilization of existing data assets and incrementally broaden the customer base and revenue opportunity for 3D data licensing.

• Distinguishing "Accuracy" for Autonomous Driving Versus Gaming Applications

  President Yoshimura breaks accuracy into two axes: geometric correctness (3D shape) and photorealism. The former is where the company's HD maps deliver direct value; the latter presumes CG craftsmanship. By converting vector data — including lane markings, gradients, and curvature — into Unreal Engine or Unity formats for import, road model construction that was previously manual can be dramatically compressed. 3DGS substitutes for on-site location scouting and serves as reference material during modeling, helping standardize quality.

• Core Business Growth Design Centered on Data for AI

  Enterprise licensing is priced on a three-axis structure: duration, coverage (geography), and use case. Use cases are defined across five categories — training, evaluation (simulation), supplementary information, certification, and safety mechanisms (ODD setting) — with a customer base spanning automakers, semiconductor companies, AI system companies, and regulators. Management identified enterprise licensing as the fastest-ramping area from FY3/27 through FY3/28.

Envalith's Perspectives

• 01.

  The Essence of the Business Model: Reuse of Existing Assets

• 02.

  A Clear Reason to Buy: Compressing the Game Development Workflow

• 03.

  Management's Ranking of Revenue Drivers

• 04.

  Challenges and Countermeasures

• 05.

  Risk Factors

Q&A Highlights

• Q

  Can autonomous-driving-grade accuracy and game-grade realism coexist within a single asset?

  A

  Accuracy needs to be separated into geometric correctness and photorealism. In racing games, the sensation of actually driving that road determines immersion, so demand for geometrically accurate 3D data is very strong. Visual beauty, by contrast, is redrawn in CG — 3DGS is not used as-is, but as reference material during production. What the company supplies is data already built for autonomous driving, delivered as a licensing model using existing assets as-is.

• Q

  Which parts of the development workflow do game studios find valuable?

  A

  First, location scouting and on-site surveying. Work that previously required staff to travel and shoot on location can be replaced by checks within 3DGS. Second, road model construction. Manually drafting lanes, elevation, gradient, and curvature can now be completed in one step by importing existing data. Third, CG modeling. Access to 360-degree photorealistic reference improves quality and standardizes the work. The net result is improvement across quality, cost, and delivery time.

• Q

  How will you measure the outcome of the Tokyo Game Show exhibit?

  A

  First and foremost, by how many leads we generate. It is a venue for direct dialogue with a large number of companies, and for us it is an entirely new domain. It is additive to revenue, and because incremental development labor is modest, we expect a profit contribution as well. We view entertainment as an incremental business opportunity to deploy high-definition 3D data we already maintain and own into a new market. At the Tokyo Game Show, we aim to expand contacts with game developers, understand customer needs, and create licensing opportunities.

• Q

  How do you think about incremental cost and margins when adding use cases?

  A

  Deals where we can use data we already own as-is carry extremely high gross margins; internally we benchmark above 80% and aim for levels approaching 100%. Where we must newly generate data for areas we do not cover, the work becomes project-based, and our policy is to accept such mandates while securing a minimum 35% gross margin with a target of around 50%. Profitability is determined less by the application field than by whether existing data assets can be leveraged.

• Q

  What is the contract structure for Data for AI enterprise licensing?

  A

  Unlike in-vehicle mass-production licensing, it is not tied to unit volumes; pricing is based on a combination of three axes — duration, coverage (geography), and use case. Use cases are defined across five categories: training, evaluation (simulation), supplementary information, certification, and safety mechanisms (ODD setting). Even within the same customer, raising the coverage level lifts the contract value, and monetization opportunities expand as use cases broaden from training to validation to edge-side deployment. The customer base also broadens in layers — automakers, semiconductor companies, AI system companies, and certification authorities.