Artificial Intelligence and Japan form a combination that, just a few years ago, would have seemed unlikely to a lot of people.
The country famous for keeping fax machines in government offices and rubber stamps as official signatures has turned, almost overnight, into one of the hottest markets on the planet for AI technology. This transformation didn’t happen by accident, and it wasn’t planned with the kind of precision typical of Japanese corporate planning. It was driven by a real, urgent need with no easy solution: a shortage of people to do the work.
And at the center of this shift is an unexpected character: Devin-kun, the affectionate nickname Japanese engineers gave to Cognition AI‘s software agent. The suffix kun in Japanese is reserved for close colleagues, usually younger or of similar rank. In other words, Japan’s tech community isn’t treating this AI agent as a tool — they’re treating it as a teammate. According to Russell Kaplan, president of Cognition AI, there was even a debate in the community about which honorific was appropriate for Devin, and the consensus landed on this endearing nickname. That says a lot about what’s happening over there 🇯🇵
But behind that charming nickname lies a much more complex story, one that blends demographic collapse, aging digital infrastructure, and a billion-dollar bet by American companies on the Asian market. To understand how Japan got here and what the adoption of Devin-kun reveals about the future of work in tech, you need to look at some numbers that would scare any planning ministry.
The problem no amount of hiring can fix
Japan currently has the most aged population in the world, with nearly 30% of its residents over 65. And the outlook isn’t encouraging at all: the working-age population is expected to shrink by more than 30% between now and 2060. This phenomenon, known as the shrinking workforce, is already being felt across virtually every sector of the Japanese economy, but in tech the impact is even more visible because demand for skilled professionals keeps rising while the number of people available to fill those roles keeps falling. It’s no exaggeration to say Japan is facing an unprecedented workforce crisis in the digital age.
The Japanese government’s own numbers put the scale of the challenge into perspective. In 2023, the Ministry of Economy, Trade and Industry (METI) estimated the country would face a shortage of 789,000 software engineers by 2030. That’s a massive gap in a sector that won’t stop growing, and no hiring campaign, no matter how aggressive, can produce qualified professionals fast enough to close it.
Receive the best innovation content in your email.
All the news, tips, trends, and resources you're looking for, delivered to your inbox.
By subscribing to the newsletter, you agree to receive communications from Método Viral. We are committed to always protecting and respecting your privacy.
To make matters worse, a large portion of Japan’s digital infrastructure still runs on legacy systems dating back to the 1980s and 1990s. Banks, insurance companies, government agencies, and even major retailers operate on outdated architectures that need to be maintained, updated, and eventually migrated to modern environments. This work requires professionals experienced in languages and platforms that younger programmers simply don’t learn anymore. The gap between what the market needs and what the available workforce can deliver has become a chronic problem the Japanese government has been trying to solve for years through incentives, educational reforms, and reskilling programs — all yielding partial results that fall short of the challenge’s magnitude.
It was in this context that Artificial Intelligence went from being a technological curiosity to a concrete answer to an existential need. As Kaplan put it, the needs are real, especially in critical infrastructure and government, with the country running on aging systems and a declining workforce. And when Devin, from Cognition AI, hit the market as an agent capable of autonomously executing software development tasks, the reception in Japan was enthusiastic in a way that surprised even the product’s own creators.
Cognition AI and the agent that became a coworker
Cognition AI is a San Francisco-based startup founded in 2023 that quickly gained attention by introducing Devin, described by the company as a true software engineering teammate. The product isn’t just a coding assistant that auto-completes lines of code or suggests functions, like tools such as GitHub Copilot. Devin was built to understand a problem, plan the solution, write the code, test it, identify bugs, and deploy the result autonomously — all within the tools the engineering team already uses day to day. It was one of the first examples of what many started calling an AI employee, an agent integrated into platforms like Slack that you can assign tasks to without having to feed it command after command.
The Japanese market’s appetite for this technology is staggering. According to Kaplan, Japan was the first or second most popular country in terms of overall user engagement for the company. That’s why Cognition AI chose Japan as the first step in its Asian expansion, opening an office in Tokyo in April, with plans to make Singapore the Asia-Pacific headquarters later this year.
The practical results help explain the excitement. Facing a national IT compliance mandate, the city government of Sapporo needed to modernize more than one million lines of legacy code — a job that, according to Kaplan, would normally require about 200 months of engineering effort. Using Devin, Sapporo’s engineers completed the task in roughly a quarter of that time. It’s this kind of efficiency gain that makes Japan look like the ideal proving ground for AI-powered software engineering.
When this agent arrived in the Japanese market, something interesting happened. Instead of being met with the resistance that many Western technologies encounter when trying to break into Japan’s work culture — known for its rituals, hierarchies, and very specific modes of collaboration — Devin was absorbed into the teams’ emotional vocabulary. The nickname Devin-kun isn’t just a cute cultural tidbit. It signals that these professionals are genuinely integrating the agent into the team’s workflow, assigning it a relational role, not just a functional one. And when a group of people starts treating a technology as a team member, adoption tends to be far more organic, lasting, and deep than any corporate rollout campaign could ever produce. 🤖
Japan’s bet on American AI
The Cognition AI case isn’t an isolated one. Japan has become the preferred gateway for American AI companies eyeing global expansion. Both OpenAI and Anthropic chose Tokyo to open their first international offices. Giants like Microsoft and Alphabet have committed billions of dollars to data centers in the country. Microsoft, for example, announced a 10-billion-dollar investment in AI and cybersecurity infrastructure on Japanese soil.
While countries like South Korea and Singapore have made so-called sovereign AI a national priority, Japan seems more comfortable betting on American AI, a result of the country’s investments in and close relationships with U.S.-based AI labs. As Kaplan noted, Japan has invested disproportionately in working side by side with American companies to influence these products’ roadmaps and address local domestic needs. It’s no coincidence that one of OpenAI’s biggest investors is SoftBank, the Japanese telecom giant led by tech enthusiast Masayoshi Son.
There’s also a positive side effect of this integration: AI can help Japan connect its digital systems with the rest of the world. Kaplan pointed out that low English proficiency has historically created a degree of isolation for some Japanese companies. But AI’s multilingual nature is tearing down that barrier. A Japanese engineer can work entirely in Japanese with Devin while collaborating, through the agent, with teams on the other side of the globe.
What changes for people who code
One of the most common questions when it comes to Artificial Intelligence agents that can write code is the most obvious one: will programmers lose their jobs? In the Japanese context, that question comes with an extra layer of complexity because the problem isn’t too many programmers — it’s exactly the opposite. The tech talent shortage in Japan is so severe that tools like Devin-kun didn’t show up to replace anyone; they showed up to fill roles that simply don’t have enough human candidates to go around.
In other markets, though, the picture is more tense. For some investors, Cognition AI‘s coding tools represent an existential threat to programmers and software engineers, especially in countries like India, a traditional hub for back-office work. The idea of AI agents doing that same work for a fraction of the cost rattled the market, and shares of giants like Infosys, Wipro, Tata Consultancy Services, and HCLTech dropped between 30% and 40% over a twelve-month period.
Kaplan, however, doesn’t seem worried about India’s ability to adapt. According to him, an engineer’s job can become more fun and impactful, with professionals who once worked solo on specific pieces of a project getting promoted to lead entire teams of AI agents. The companies working with Cognition, he said, are using the productivity gains to become more ambitious.
The company’s expansion across Asia has also revealed surprising markets. Malaysia, with its capital Kuala Lumpur, has become a regional software engineering hub, driven by a large English-speaking talent pool, lower operational costs, and proximity to the rest of Southeast Asia. Kaplan described the engineers his team found there as some of the most skilled in the world at managing AI agents, so much so that Cognition launched an applied AI engineering program there to identify and train top professionals. South Korea and Australia are also on the company’s radar as potential expansion markets.
In practice, what Japanese engineers are reporting is that the Cognition AI agent is taking over the most tedious and repetitive day-to-day development tasks — like refactoring old code, writing automated tests, documentation, and fixing simple bugs — freeing up human professionals to focus on the parts of the job that actually require creativity, contextual judgment, and deep business knowledge. This model of dividing labor between humans and AI agents isn’t new as a concept, but Japan is one of the first markets to apply it at real scale, on production projects, with documented results and in technically demanding contexts.
Another important aspect of this transformation is the effect it has on the concept of a tech workforce. If an AI agent can competently handle a significant chunk of the tasks a junior or mid-level developer would do, what changes isn’t just the number of people needed to deliver a project — it’s also the profile of those people. Japanese companies at the forefront of this adoption are already starting to redesign their hiring, training, and performance review processes to account for AI agents as a permanent part of the workflow. It’s no longer about hiring someone who knows how to write code — it’s about hiring someone who knows how to work with an agent that writes code. That shift in perspective is subtle, but it has enormous implications for how tech careers will evolve in the coming years, not just in Japan, but around the world.
It’s worth noting that this adoption frenzy is happening amid explosive growth for Cognition itself. In late May, the startup raised over 1 billion dollars in a funding round that valued the company at 26 billion dollars, more than doubling the valuation from a previous round. The company’s annualized revenue jumped from just 37 million dollars to 492 million in a single year. And since processing power — the famous compute — is a finite resource with demand doubling every seven weeks, having teams spread around the globe has even become an operational advantage: when it’s a workday in Japan, it’s the middle of the night in New York, allowing the company to tap into processing capacity during off-peak hours.
What Japan is experiencing today with the adoption of Artificial Intelligence in the tech sector is, in many ways, a natural experiment the rest of the world will be watching closely. The country’s unique conditions — the demographic crisis, legacy infrastructure, collaborative work culture, and a surprising openness to technologies that solve real problems — have created an environment where the integration of humans and AI agents is advancing organically and at a rapid pace. And the fact that the engineers over there chose to call this agent Devin-kun might be the most honest indicator that this integration has already moved past the experiment phase and into the coexistence phase. 🚀
