![\"Japan](\"https:\/\/www.europesays.com\/japan\/wp-content\/uploads\/2026\/07\/1784132168_116_0x0.jpg\")
<\/p>\n<\/p>\n
CHIBA, JAPAN – OCTOBER 17: Visitors attend the latest innovations at CEATEC 2025 in Chiba, near Tokyo, Japan, on October 17, 2025. The event, themed “Innovation for All,” highlights technologies shaping the future of everyday life, from AI and smart mobility to digital health and sustainable living. (Photo by David Mareuil\/Anadolu via Getty Images)<\/p>\n
Anadolu via Getty ImagesSolving the shortage of babies with an AI robot army<\/p>\n
For decades, Japan has held a special place in the Western imagination as the country where the future arrives first \u2014 bullet trains, the Walkman, robot dogs, toilets with more computing power than the Apollo program. Then, somewhere in the 1990s, the future seemed to decamp for Shenzhen and Palo Alto, and Japan settled into a long period of being politely admired and quietly underestimated.<\/p>\n
Tokyo has apparently decided it’s done being underestimated.<\/p>\n
On June 30, Japan’s Ministry of Economy, Trade and Industry formally commissioned<\/a> a new company called Noetra \u2014 backed by SoftBank Corp (9434.T), Sony Group (6758.T), NEC (6701.T), and Honda Motor (7267.T) \u2014 to build a homegrown foundation model for \u201cphysical AI\u201d: artificial intelligence that doesn’t write your emails but instead runs factories, warehouses, and eventually restaurants and hospitals. The government will provide up to \u00a51 trillion<\/a>, about $6.2 billion, over five years, with \u00a5387.3 billion out the door in year one.<\/p>\n The number attached to that money is the revealing part. Industry Minister Ryosei Akazawa wants 10 million AI-equipped robots<\/a> deployed across 18 sectors by 2040. That is not a research target. It is a mandated headcount. <\/p>\n Japan’s fertility rate has sunk to a record low, its over-65 population has passed 36 million, and by METI’s own math the country will be short 3.39 million AI and robotics specialists by 2040. A shortage of the very people needed to build the robots meant to replace all the other people Japan doesn’t have. When your labor market has a hole that Japan\u2019s immigration policy won’t fill and birthrate campaigns can’t begin to touch, you must build your workers instead.<\/p>\n The Plan Behind the Plan<\/p>\n Readers of my column on China’s blueprint<\/a> for 2030 will recognize some common themes. Japan has been watching its neighbor’s five-year-plan industrial policy machine with a mix of alarm and, evidently, note-taking.<\/p>\n Noetra is one piece of a 14-year national growth strategy, unveiled in June, targeting \u00a5370 trillion<\/a> (roughly $2.3 trillion) in combined public and private investment across 17 strategic sectors by 2040. Semiconductors claim the largest share at \u00a568 trillion, anchored by Rapidus<\/a>‘s audacious attempt to leapfrog back into advanced 2 nm logic manufacturing. Physical AI gets \u00a510.5 trillion. Biopharma gets \u00a520.8 trillion, and nuclear fusion \u00a53 trillion. Quantum computing, space, shipbuilding, defense, and, delightfully, anime and film round out the list, the last with a target of \u00a520 trillion in annual content exports, an amount comparable to Japan’s car exports. METI’s own budget rose roughly 50%<\/a> this fiscal year to fund it all.<\/p>\n To its credit, Tokyo is being unusually honest with itself about where it can win. Nobody in Kasumigaseki is pretending Japan will out-train OpenAI on frontier LLMs. The internal logic runs instead through the things Japan never stopped being good at. <\/p>\n The country still produces about 38%<\/a> of the world’s industrial robots, and its factories sit on decades of proprietary operational data that no Silicon Valley lab can scrape from the internet. The wager is that AI’s next act will be fought in the physical world \u2014 where hardware, sensors, precision manufacturing, and factory-floor data are the ammunition. \u2014 This is terrain where Japan enters the fight already armed. Nvidia (NVDA) evidently agrees; Jensen Huang has spent the past two years courting Japan’s robot makers as enthusiastically as its politicians.<\/p>\n There’s also refreshing discipline in the fine print. Noetra’s funding comes with annual stage-gate reviews<\/a> \u2014 only the first two years are locked in, and if milestones slip, Tokyo has designed itself a quiet exit. That’s not a detail one typically finds in industrial policy, which usually operates on the sunk-cost principle.<\/p>\n Fixing the Capital Allocation Culture<\/p>\n Here is where the story gets interesting for capital markets professionals, because Japan’s innovation ambitions have long been strangled by a very particular financing pathology.<\/p>\n Japan produces world-class science and engineering, but its venture ecosystem has historically run out of oxygen exactly when companies need it most: the late-stage growth rounds where a promising startup becomes a global contender. The domestic fix has been the state-backed Japan Investment Corporation, which runs growth-stage funds, a dedicated vehicle for pre-IPO secondaries, and in March published a \u201cmodel term sheet\u201d<\/a> designed to coax institutional LPs into Japanese venture funds \u2014 a document that amounts to a polite official admission that the existing fund terms were scaring the grown-up money away.<\/p>\n The encouraging news is the grown-up money is starting to arrive anyway. Khosla Ventures, NEA, and Bessemer have all begun writing checks into Japanese startups directly. Andreesen Horowitz<\/a> just opened a local office in Japan to source deals. A new legal instrument<\/a> taking effect this spring even lets startups pledge intangible assets \u2014 know-how, customer relationships \u2014 as loan collateral, a small revolution in a country where bank lending has traditionally required something you could photograph.<\/p>\n Tokyo Stock Exchange Raises the Bar, and the Exits Narrow<\/p>\n Meanwhile, the Tokyo Stock Exchange has been performing surgery on the other end of the pipeline. For years, Japan’s listing culture ran on what might be called the IPO-as-graduation-ceremony model: venture investors, lacking other exits, pushed portfolio companies to list early and small, with market caps of a few billion yen, after which the stocks drifted into institutional obscurity. The TSE, deciding that \u201clisting is not the finish line, but the starting point,\u201d has been tightening standards<\/a> on its Growth Market since 2022, with genuine teeth: 161 companies were out of compliance as of last November, and delistings began<\/a> in March.<\/p>\n The result is a paradox worthy of a haiku. Japanese equities spent the first half of 2026 near record highs \u2014 yet the country logged just 17 to 18 IPOs, the fewest since 2011<\/a>. Over the same six months, Hong Kong raised nearly $44 billion in equity capital, much of it from the Chinese AI listings I chronicled in my January column<\/a>. Japan is minting better companies and fewer listings simultaneously.<\/p>\n For a certain kind of Japanese company \u2014 too ambitious for a small domestic float, too capital-hungry to stay private \u2014 the arithmetic increasingly points to a splashy IPO across the Pacific.<\/p>\n The PayPay Precedent<\/p>\n Which brings me to why I’m genuinely excited about this September.<\/p>\n In March, PayPay (Nasdaq: PAYP), Japan’s dominant mobile payments platform and a SoftBank Group (9984.T) portfolio company, listed on Nasdaq<\/a> in the largest U.S. IPO by a Japanese company in nearly a decade \u2014 raising $880 million at a $10.7 billion valuation, with cornerstone orders<\/a> from Visa (V), the Abu Dhabi Investment Authority, and Qatar Holding, and a healthy first-day pop for good measure. It priced into a down market during a geopolitical flare-up and traded up anyway. Historically, Japanese companies on Nasdaq were micro-caps that struggled to get an analyst’s phone call returned. PayPay arrived at a scale institutions could take meaningful positions.<\/p>\n