The Frontier of AI Startups

The CVO is a division of the Innovation Growth Strategy Division, located globally in North America, Japan, and Europe. I head North America, leveraging the experience of having been in Silicon Valley for over 20 years and the network of various startups and venture capital firms I have built during the time period.
The first role of the CVO is to introduce startups invested in by the Hitachi Ventures^*1fund to Hitachi's business units (BUs) and to help the BUs realize the important products and solutions envisioned in their mid- to long-term business strategies. The second role is to strengthen cooperation between startups, the BUs, and the BUs' customers.
Specifically, the first role involves regular meetings with the BUs' innovation leaders (CEO, CTO, etc.) to understand the technology required, and the BUs and their customers' needs. Then we introduce them to startups that have connections with the CVO. When introducing startups, we offer suggestions not only from a technology perspective, but also on how they can contribute to increasing sales of the BU's products and services and on patterns of collaboration. We can also find and approach new startups that meet the needs of the BU.
In addition to sharing information with the BUs, we run marketing programs where we invite people from the BUs to meet with startups funded by Hitachi Ventures.
As for the second role, we collaborate with Hitachi's R&D Division, where I used to work. As a preliminary consideration stage of a BU's cocreation with a customer, we work with Hitachi's data scientists and other parties to verify whether the customer's problem can be solved by Hitachi and startup technologies. If we determine that we can contribute to solving the problem, we will propose and carry out Proof of Concept (PoC) and Proof of Value (PoV) for commercialization. In this process, the CVO accompanies the BU to help strengthen the relationship with the customer.

In the AI startup world, technology is being rebuilt around generative AI. This is expected to lead to a paradigm shift, including the emergence of new leaders who lead the technology in each level. To be specific, transformation is occurring at all levels, from semiconductors as the lowest layer to platforms, infrastructure, middleware, and applications as the highest layer.
Generative AI is being applied across a wide range of operational areas. Its application begins in customer operations, and then it spreads to software engineering, marketing and sales, R&D, supply chains, and then finance. Especially, in the areas of software engineering and R&D, the use of generative AI has reduced work time for developers and researchers. We are seeing new trends compared to traditional AI technologies.
Text data generation is a domain where companies are advancing R&D proactively, including major players such as OpenAI for its GPT series, Google DeepMind, Meta, and Alibaba, and startups such as Anthropic and Tabnine. Similarly, in the 3D object domain, models such as DreamFusion, NVIDIA GET3D, and Shape-E are being developed. Due to the high cost of developing large language models (LLMs), many startups opt to customize open-source LLMs for their applications, as only a few companies can afford to build their own from scratch.
In addition, new attempts to utilize generative AI have begun to improve the operational efficiency of IT infrastructure and autonomous control in the robotics domain. However, their efforts are still in the early stages, and specific results are expected in the future.
The venture capitals in Silicon Valley are following these trends and launching hundreds of millions of dollars in funds in each area of generative AI application. The amount of their funding ability is estimated from $10 billion to $12 billion a company.

Hitachi Ventures has invested in or is considering investing in several startups that are focused on developing next-generation technologies. In the field of applications, three companies are developing innovative technologies: Archetype AI, Xaba and Makersite.
Archetype AI develops multimodal generative AI models that analyze image, audio, and other data for retail stores, distribution warehouses, and other service locations. This technology supports the decision-making of site supervisors at manufacturing and logistics sites, and is expected to improve safety and efficiency.
Xaba aims to improve productivity of the manufacturing industries by developing software to automate the programming and engineering of industrial robots and 3D printers.
Makersite develops software called “Product Lifecycle Intelligence” for the manufacturing industry that integrates CAD^*2, PLM^*3, ERP^*4, and BOM^*5data with supplier information and uses machine learning to analyze and optimize in product design and material selection.
On the other hand, there are some startups that should be noted as security companies.
Trustwise AI has developed technology to detect hallucinations and prevent data breaches by generative AI to enhance enterprise data protection.
StrikeReady develops security software called “Threat Intelligent Engine” that provides integrated monitoring of multiple layers of security and enables rapid response in the event of an incident, helping to proactively defend and avoid crises.
Archetype AI, Xaba, and Makersite apply generative AI technology in the industrial domain, which I think is highly compatible with Hitachi's existing business and customers and has great potential to create new value. I also believe that Trustwise AI's and StrikeReady's technologies are important for safety in the mission-critical domain (core business functions that are so critical that they cannot be interrupted and must be available 24/365).
As next-generation technologies, I'm focusing on microcontrollers and autonomous vehicles, which are key components of service systems in industrial areas. In terms of the semiconductor stack, I'm also looking at a company for its technology that improves LLM performance on the CPU by applying the understanding of the structure of the human brain nerve system onto CPU architectures.

The CVO's role is supporting the BUs in their efforts to strengthen their customer base and the popularization of products and services. To this end, it is important to improve the customer experience and expand the products and services.
If we provide customers not only individual products, but also combining multiple products and services to offer service value that enhances their business value, then customers will recognize Hitachi as a strategic partner.
As Hitachi expands its products and services, the CVO will support collaboration with startups specializing in generative AI. Collaboration with Hitachi can benefit startups as well, including access to Hitachi's wide customer base, so I hope that we can build mutually beneficial relationships.
Generative AI may seem like an emerging field in domestic business in Japan, but I believe it is important to strategically identify investment areas for the Hitachi's next Mid-Term Management Plan toward 2027.
To this end, the CVO will analyze trends in technology and startups, as well as trends in competitors, and will actively provide information to customers and Hitachi. Through these activities, we hope to contribute to improving the Hitachi's corporate value.