From real-time operating systems to full-stack embedded platforms, eSOL enables mission-critical software across vehicles, industry and medical devices—turning connected machines into reliable smart systems.
To begin, how would you define the current state and core strengths of Japan’s IT industry in this moment—when the world is moving rapidly toward AI and cloud computing, yet Japan has historically ranked low in digital competitiveness?
We actually addressed this at our private annual conference at the end of June, where I gave the keynote titled, “Where Japan’s Software Is Now and Where It’s Going: A New SDV for Japan for Every Vehicle.” The starting point is a recent Digital Economy Report from METI—Japan’s Ministry of Economy, Trade and Industry—published in early April. The report is blunt: in enterprise IT, Japan has essentially lost control of the platform layer. The operating systems, hyperscale clouds, and databases our companies rely on are almost entirely governed by non-Japanese players—Microsoft, AWS, and others.
In that world, many Japanese IT firms are just integrating someone else’s platform. Most developers don’t even know the hardware under the hood. With the arrival of AI, the nature of work itself is changing: activity-heavy, month-long efforts will see 80% of the tasks produced by AI in minutes. If you don’t adopt this, your competitors will—and they will outpace you.
The differentiator in AI isn’t the model; foundational models are increasingly open and widely shared. China, for example, contributes heavily to open-source AI. The real competitive battleground is data—especially closed, proprietary data. In our business, that means design documents, source code, verification reports, and design reviews—the accumulated know-how. Before large language models, only humans could digest and apply that knowledge. It didn’t scale. Now LLMs can generate an 80% solution in minutes—work that might take a human a month. For safety-critical systems we still need human experts to take the output from 80% to near-100%, but the productivity step-change is undeniable.
Here’s the catch: you only capture and govern data if you control the platform. Applications read from and write to the platform. If you don’t own the platform, you are effectively donating your data to whoever does. That’s Japan’s problem in ICT: we’re building on external platforms, so we don’t control the data, and as AI accelerates, application development itself gets automated away. In short, the ICT game is basically over for us unless we change direction.
But there’s another domain: cyber-physical systems—the software that runs in the physical world: cars, robots, industrial equipment, social infrastructure. This domain has no global winner yet. It’s harder than enterprise IT because it’s mission-critical. If your phone freezes for two seconds, it’s annoying; if your brake controller freezes for two seconds, people can die. That’s the essence of real-time, safety-critical software.
Our society today is only halfway to true cyber-physical integration. We instrument the physical world with IoT and analyze data in the cloud, but human operators still interpret results and take action. With extreme weather, traffic, and logistics dynamics, human-in-the-loop responses are often too slow. The opportunity is to close the loop—sense, decide, and actuate autonomously—safely and reliably at scale. That’s difficult, but it is precisely where Japan can lead.
Japan remains one of the world’s manufacturing powerhouses. We’re no longer the undisputed number one—China is formidable—but Japan is still second or third in many categories, and our political environment is relatively stable. Critically, consider scale in automotive: global vehicle production is roughly 90 million units per year. If you aggregate Japan’s twelve OEM groups, their combined production is about 31 million vehicles—around 35% of the global market—larger than any single country or region. Software has a near-zero marginal cost of replication, so platform economics reward reuse and scale. If our OEMs align on shared software platforms, Japan has a real shot at defining the de facto standard for software-defined vehicles.
If we fail to build these platforms, we repeat the smartphone and PC story: Japan had a strong handset industry that vanished when platform control shifted to iOS and Android. The same will happen in “monozukuri”—the physical products—because everything is becoming software-defined. Lose the platform, lose the data, and over time you even lose product competitiveness because you can’t learn from how customers use your products. METI’s report underestimates the severity in embedded and industrial software; our on-the-ground view is more sobering. Even under a base-case scenario, the ten-year digital deficit in monozukuri—combining the cost of imported software platforms and lost product competitiveness—could approach the scale of Japan’s pre-COVID trade surplus. If we simply sit still, we decline.
So my message is: change course now. Unlike in past eras, we cannot “rebuild from zero” later. Digital complexity and the pace of compounding improvements make catch-up from a cold start unrealistic. We must build cyber-physical platforms, at scale, together.
Image of eSOL Full Stack Engineering (FSE), ©eSOL Co.,Ltd.
Turning to eSOL specifically: your company is known for a full-stack approach—bringing together the operating system, middleware, services, and applications. Why did you choose this model, and how does it create superior value and competitiveness?
We don’t build the vehicle or robot itself; our role is to enable manufacturers to build a robust, reusable software platform that fits the realities of embedded systems. A car’s platform is not AWS, iOS, or Android. If it were, the hyperscalers would already own the market. Automotive and robotics demand distributed, real-time computing with hard safety constraints. “Real-time” isn’t just about speed—it’s about bounded, consistent response every time. If the brake must respond within, say, 100 milliseconds, it must do so a million times in a row. That’s our core expertise, developed over decades of building real-time operating systems and safety-critical software.
Our model is to provide the technology and reference blueprints so each manufacturer can design its own platform—yet with enough commonality that future alignment across OEMs is feasible. Today, a Honda engineer may privately understand that, eventually, Honda will need to partner with Toyota, and vice versa. When that day comes, if both have built on similar architectural skeletons, integration becomes realistic. That’s why we call what we do “full-stack engineering”: foundation OS, middleware, and a services layer built on a service-oriented architecture (SOA), which is the same architectural paradigm that underpins PCs, smartphones, and clouds.
In practice, the platform provides safe APIs for capabilities like automated driving or remote vehicle control. A mobility operator—today’s equivalent of an Uber or a parking operator—doesn’t need to reinvent safety or real-time control; they call the API and the platform handles safety, timing, and fault tolerance. This opens the ecosystem to many more application developers without compromising safety.
Equally important, platform development turns monozukuri software from one-off, project-based efforts into a continuous, evergreen capability. Historically, a new car program would lock hardware and electronics, then throw software over the wall to “finish before SOP.” With software-defined vehicles (SDV), software evolves continuously like iOS—new features, improvements, and safety updates over the life of the product. That requires full-stack competence: OS, middleware, services, and application know-how, so we can help customers transform their current applications into reusable services—often 80% of existing app code becomes services in a platform model. That’s how we accelerate the industry’s shift to true platforms.
Do you envision a future where Japan’s twelve major OEMs become largely fabless—selling a basic EV skateboard and focusing mainly on software, with manufacturing outsourced?
In some niches, perhaps—but not universally. Toyota’s greatest strength isn’t design; it’s production. Their manufacturing system remains world-class and is the core source of profitability. If you give up production, you surrender a decisive competitive asset. The real risk is the opposite: if a manufacturer fails to master software platforms, it could end up as a contract manufacturer—excellent at making someone else’s design, but on thinner margins, like a Foxconn. Many Japanese firms have outstanding factories; if they only produce and don’t own the platform and product experience, the business shrinks and profitability erodes. The goal is to combine production excellence with software platform leadership.
Your partner network spans semiconductors to middleware and beyond. How do your engineering services and partnerships enhance value for clients? Are you seeking new international partners?
Full-stack engineering covers a vast territory. We’re about 500 people—AI boosts productivity, but we still need partners to cover the entire landscape. Our philosophy is “one team”: eSOL, our partners, and the customer, aligned on a shared goal. Today we have roughly 100 partners. That’s powerful, but not fully efficient. We’re re-profiling at least half into strategic, longer-term collaborations—think multi-year agreements rather than ad-hoc, project-by-project engagements—so both sides can invest and plan.
The strongest form of partnership is acquisition. In August, we acquired Kyoto Microcomputer, a small but technologically formidable, 40-year-old Japanese products company. Their strength is in compilers—absolutely fundamental tools that translate human-written source code into binary instructions executed by chips like ARM or x86. The quality of the compiler directly affects system performance and efficiency, so this capability is elemental to building great computers—and thus great platforms. We are also in discussions on additional acquisitions. Broadly speaking, there are two types: technology-driven tuck-ins and scale-oriented deals. We’ve just done the former; you can infer that we’re also evaluating the latter. International acquisitions are very much on the table.
On international strategy—beyond your Paris office—where are you focusing now, and which markets look most promising?
Given today’s geopolitical context and the urgency of Japan’s own digital deficit, our top near-term focus is Japan. Pushing aggressively overseas is less efficient than it should be because many regions prefer local suppliers for non-technical reasons. An exception is China, where we see strong technology appetite and alignment with global standards like AUTOSAR. China’s leadership understands that demographic headwinds will shrink the domestic market, so they’re investing to compete globally, which means embracing international standards. They’re currently the largest AUTOSAR adopters, precisely to smooth overseas expansion. That creates opportunities for collaboration with advanced customers, which, in turn, sharpens our own technology. Japan’s unique position as a trusted counterpart to many regions also helps. But practically, to build the foundational platforms we’ve discussed, saving and transforming the Japanese market—the largest for us—is existential. That is where we must lead.
Looking ahead to 2030, what do you want eSOL to have achieved? What’s your personal ambition as president?
Our six-year plan, “eSOL Reborn 2030,” runs through the end of 2030. The vision is to make a true cyber-physical society possible by delivering full-stack engineering platforms to monozukuri customers. On people and partnerships, we’re institutionalizing the “one team” model and deepening strategic alliances and M&A.
Financially, translated roughly at ¥100 to the dollar for simplicity, last year’s revenue was about $120 million—a record for revenue and profit. We aim for at least $250 million by 2030—more than double. This year is already tracking above last year. Longer term, I’ve publicly said the company has a credible path to $1 billion, even if that milestone might land slightly after 2030. Achieving the platform vision across Japan’s manufacturing base would make that outcome unsurprising. Organic growth is our core engine, but inorganic growth will be necessary to scale at the speed the market demands.
Your customer mix shows automotive at roughly 47–50%. Will that weighting change as you grow areas like medical, industrial robotics, and aerospace?
Today, automotive is about 50%. Consumer electronics—historically our largest segment—has declined as that industry matured; it’s around 20% and stable to slightly down as our overall revenue grows. Industrial is roughly 15%, and we also serve medical and aerospace customers, though some prefer not to be named publicly.
I don’t manage to fixed percentage targets. Twenty years ago, consumer electronics was 50% and automotive much smaller; now it’s flipped. With the rise of software-defined vehicles, automotive is the most logical beachhead for platformization: massive volumes, high unit value, stringent safety and cost requirements, and continuous software evolution. Once you master the automotive platform, you can port 70% of that architecture into adjacent verticals—industrial robots, drones, AGVs, marine vessels, medical and surgical robots, even social robots—because the engineering patterns are similar. So our strategy is to win the platform in automotive, then extend horizontally. Over time, automotive’s share may decrease as other segments expand—but that would be because the platform has successfully propagated across verticals, not because automotive has weakened.
Thank you for such a comprehensive perspective—both on Japan’s digital deficit and on the opportunities ahead if the OEMs align. Any final thoughts?
Only that the window is open now, not later. If Japan’s manufacturers align on shared, safety-critical software platforms—built for real-time, distributed control—we can retain production excellence and pair it with software leadership. That combination will define the next era of monozukuri. Our mission at eSOL is to make that practical—full-stack, reusable, and scalable—so Japan can lead again in the software-defined physical world.
For more information, visit their website at: https://www.esol.com/
0 COMMENTS