What JAFCO Announced

JAFCO Group has finalized the speakers for JAFCO SEED 2026, its third conference for founders around company formation and the seed stage. The free event will run from 12:15 to 6:00 p.m. on Thursday, August 6 at Bellesalle Iidabashi First in Tokyo. Attendee registration closes August 3.

The program has two main teaching tracks. A macro session brings together the Ministry of Economy, Trade and Industry, the Tokyo Stock Exchange and Japan Finance Corporation. It will cover startup policy, public startup loans and the Growth Market’s revised continued-listing standard. An AI session brings Anthropic Japan, LayerX and Amazon Web Services Japan together to examine foundation models, infrastructure, implementation, go-to-market design and durable competitive advantage.

The named speakers are Shunta Sato of Japan Finance Corporation, Tonami Uzuyama of the Tokyo Stock Exchange, Yuta Sakaki of METI, Taishi Okada of Anthropic Japan, LayerX CTO Yuki Matsumoto and Tomoyasu Mamiya of AWS Japan. The pitch final will be judged by ANRI general partner Anri Samata, East Ventures partner Tsuyoshi Kaneko and JAFCO partner Yutaro Saka.

Precision matters. This announcement concerns a conference, networking program and pitch final. It does not say every attendee will pitch, receive funding, obtain Anthropic access or enter an accelerator. The pitch application process began in March and included document review, interviews and pitch coaching. General attendee registration is a separate process.

August 6Conference and pitch final in Iidabashi, Tokyo.
FreeAttendance fee; registration closes August 3.
Nearly 400Cumulative participants across the first two editions, JAFCO says.
3rd yearJAFCO SEED’s 2026 edition.

Why Put an Exchange in a Room of Seed Founders?

A founder with a prototype may regard an IPO as a distant abstraction. JAFCO’s lineup argues the opposite: public-market discipline begins with the first cap table, first customer contract and first management report. Choices made at seed—how revenue is recognized, who owns intellectual property, how options are granted, whether customer concentration is measured—can determine whether later diligence is routine or painful.

Tokyo’s Growth Market was created in the 2022 market restructuring for companies with high growth potential. The exchange has since tightened its message. From March 2030, the continued-listing criterion will require market capitalization of at least ¥10 billion five years after listing, replacing a ¥4 billion threshold after ten years. Initial listing standards are not being raised in the same way, but the new maintenance rule tells small issuers that going public is not the finish line.

An IPO is not graduation from building a company. It is a change in who can inspect the homework.

Japan has long been criticized for “small IPOs”: companies list early, gain liquidity, but remain too small for major institutional investors and sometimes lose growth urgency. The TSE change seeks to encourage scale, mergers and repeat entrepreneurship. A seed founder should therefore ask not merely “Can this company list?” but “Can it become large, liquid and strategically important after listing?”

Anthropic: The Model Is Powerful—and Rentable by Everyone

Anthropic represents the foundation-model layer. Its Claude systems give small teams access to capabilities that once required a large research laboratory: software generation, document analysis, customer support, research assistance and agentic workflows. Anthropic opened a Tokyo office in 2025 and has expanded Japanese enterprise partnerships, including a 2026 collaboration with NEC.

That access is a gift and a trap. If every competitor can call a similar model API, “we use AI” is not a moat. Model quality changes quickly, prices move, context limits expand and customers can switch tools. A startup built as a thin interface over a third-party model may discover that its supplier or a larger platform can reproduce the feature.

Possible AI advantage What makes it defensible
Workflow Deep integration into a painful sequence of work, not one isolated prompt.
Proprietary data Legally obtained feedback or outcomes that improve the product over time.
Distribution Trusted access to a specific profession, industry or customer channel.
Reliability Evaluation, human review, security and auditability for high-stakes use.
Switching cost Configuration, history and team habits that create genuine continuing value.
Speed of learning A system that converts customer use into better product decisions faster than rivals.

A responsible founder must also know the dependency map: which model processes what data, where it is stored, how output is evaluated, what happens during an outage, and whether the product can switch providers. Model strategy belongs beside product and risk strategy, not inside an engineering footnote.

LayerX: The Difficult Middle Between Demo and Business

LayerX represents implementation. Founded in 2018, the company moved from blockchain-era experimentation into enterprise workflow products. Its Bakuraku services address invoices, expenses, cards and other back-office work; its newer AI businesses aim to place agents inside corporate processes. The company reported a cumulative Series A raise of about ¥10.2 billion in 2023 and has since pursued larger-scale growth.

Its relevance is not that every founder should copy expense software. It is that enterprise AI succeeds by understanding the unglamorous middle: approvals, exceptions, master data, permissions, audit trails, procurement, security review and the employee who must fix the machine’s mistake on Friday afternoon.

A demo optimizes the happy path. A product must survive missing fields, contradictory rules, unusual customers and old systems. A business must then acquire customers for less than their lifetime gross profit, implement the service without consuming all margin, retain accounts and expand use. These layers—not the cleverness of one model response—decide whether AI becomes revenue.

The Capital Ladder: Equity Is Only One Tool

The presence of Japan Finance Corporation alongside venture investors teaches another useful distinction. Equity capital buys time and risk tolerance but permanently changes ownership. Debt preserves equity but creates repayment obligations. Grants and subsidies may fund defined work without dilution, but impose eligibility, documentation and timing constraints. Customer revenue is the healthiest financing when available, yet early technical work may precede a sellable product.

Capital source Best suited to Core trade-off
Founder money / revenue Early validation and disciplined growth Limited scale and personal risk.
Angel / VC equity High uncertainty with potential for very large outcomes Dilution, governance rights and exit expectations.
Public startup loan Defined spending with a credible repayment path Cash repayment and underwriting requirements.
Grant / subsidy R&D or policy-aligned projects Restricted use, applications and reporting.
Corporate partnership Market access, validation and co-development Slow decisions, exclusivity risk and dependence.

The right question is not “How much can we raise?” It is “What milestone will this capital buy, and will reaching it make the company more valuable before the money runs out?” Seed finance should turn unknowns into evidence: a working product, repeat use, willingness to pay, a regulatory path or a technical breakthrough.

How a Seed Pitch Works

A pitch is a compressed model of a company. JAFCO’s published selection lens emphasized the clarity and impact of the problem, why now, market growth and scalability. Those criteria can be translated into a chain of evidence.

Pitch question Evidence stronger than a slogan
What hurts? Observed workflow, quantified loss and interviews with the actual user.
Why now? A specific technology, rule, cost shift or behavior change.
Why this product? A demonstration tied to the user’s existing process and alternatives.
Why this team? Unusual access, insight, technical ability or earned trust.
Why can it be large? A reachable first market plus credible expansion, not a giant top-down statistic.
What has been learned? Retention, paid pilots, usage, failed experiments and changed assumptions.
What does funding buy? A milestone, budget, runway and next financing logic.

The pitch is not the company. Presentation skill can create attention, but diligence should test the claims. Still, attention has economic value. A credible public story can attract hires, customers and investors at the same time. That is why pitch events can create momentum—and why founders must not confuse momentum with product-market fit.

JAFCO and the Birth of Japanese Venture Capital

JAFCO’s own history mirrors Japan’s risk-capital development. It was established in 1973 as Japan Associated Finance, a joint venture of Nomura Securities, Nippon Life and Sanwa Bank. Japan’s high-growth era had just ended in the first oil shock. Venture investment was still embryonic and finance remained centered on banks and large corporate groups.

In 1982 JAFCO established what it describes as Japan’s first venture-capital partnership fund. Overseas investment followed. Japan later created a legal foundation for investment limited partnerships in 1998, making fund structures more recognizable. The Mothers market opened in 1999 to give growth companies an earlier public route. Independent VCs, university funds, corporate venture capital and government-backed vehicles expanded through the internet, mobile and deep-tech eras.

JAFCO reports more than ¥1 trillion in cumulative fund commitments and more than 1,000 portfolio-company listings since 1973. These are company-reported cumulative figures, not a promise about the outcome of JAFCO SEED participants. Venture portfolios follow a power law: a small number of exceptional outcomes must compensate for many losses and modest returns.

Year Milestone Meaning
1973 JAFCO founded Institutional venture investing emerges during Japan’s shift from high growth.
1982 JAFCO No. 1 partnership Japan’s first VC partnership fund, according to JAFCO.
1998 Investment Limited Partnership Act A legal foundation for modern Japanese VC funds.
1999 Tokyo Stock Exchange Mothers opens An earlier public-market route for high-growth companies.
2022 TSE restructuring and Startup Five-year Plan Growth Market replaces Mothers; government targets far more startups and investment.
2030 New Growth Market standard takes effect ¥10 billion market cap required five years after listing.

Japan’s Policy Bet: From More Startups to Bigger Outcomes

The government’s 2022 Startup Development Five-year Plan set an extraordinary ambition: increase annual startup investment from roughly ¥800 billion to ¥10 trillion by fiscal 2027, and eventually create 100 unicorns and 100,000 startups. By 2023, official figures said Japan had about 22,000 startups, up from 16,100 in 2021, while investment was ¥754 billion.

Those numbers expose the challenge. Creating more companies is not the same as producing globally important businesses. Japan needs later-stage capital, international sales, experienced executives, acquisitions and markets willing to reward long-term investment. It also needs to tolerate failure without treating it as permanent social disqualification.

JAFCO SEED’s public-policy and TSE session reflects a transition from “increase the count” to “increase the scale.” Founders should use policy as leverage—not as a substitute for customers. Subsidies can accelerate experiments; they cannot manufacture demand. A favorable listing route can supply capital; it cannot create competitive advantage.

What a Founder Should Take Home

A useful conference changes the questions a founder asks the next morning. From Anthropic: which capabilities are becoming commodities, and where can the company create ownership? From LayerX: what operational exception will break the demo? From AWS: what will inference, storage and security cost at ten and one hundred times usage? From Japan Finance Corporation: can some milestones be financed without dilution? From METI: which rules or programs change the timing of the market? From TSE: can governance and growth survive public scrutiny?

Founders should arrive with a one-sentence customer problem, three pieces of evidence, a dependency map and one sharply framed question for each kind of speaker. Networking is not collecting business cards. It is finding the person who can falsify an assumption, unlock a customer or explain a constraint before it becomes expensive.

JAFCO SEED 2026 is a stage, not a funding result. Its value will depend on whether founders convert the afternoon into better decisions. The deepest lesson of the lineup is that the seed stage is not small. It is simply early: the architecture of technology, ownership, governance and ambition is already being set.

Sources and Further Reading