2028年3月までに10兆円。昨年の実績は7,613億円だった

In November 2022 the Kishida cabinet published a document with an unusually specific promise in it. By March 2028, investment into Japanese startups would reach ¥10 trillion, “more than 10 times the amount in 2021,” alongside 100 unicorn companies and 100,000 startups (JETRO, summarising the Startup Development Five-Year Plan; the Cabinet Secretariat keeps the plan’s own portal here).

That deadline is now nineteen months away. It is worth looking at the scoreboard, because the gap between the target and the number is not a rounding error, and because what the gap is made of is more interesting than the gap itself.

The number

Japanese startups raised ¥761.3 billion in 2025, excluding debt, which is almost unchanged from ¥779.3 billion the year before (Speeda, Japan Startup Finance 2025, via Japan FinTech Observer, 2 February 2026).

The target is ¥10 trillion. The market is at roughly seven and a half percent of it, and it has been flat for two years. The 2022 peak, when the plan was written, was over ¥990 billion, which means the line has gone slightly down since the promise was made, not up.

No amount of policy is closing a thirteen-fold gap in nineteen months. That much is arithmetic, and everyone involved knows it.

Two caveats are worth stating before anyone builds an argument on the number. The Speeda figure excludes debt, so it is a measure of equity going into startups rather than of all capital reaching them, and Japan’s policy banks do a meaningful amount of their work through lending. And the plan’s own ¥10 trillion is a target for investment, a word the plan does not tie to a single published methodology, which means the government is not obliged to score itself against this particular tracker. Neither caveat changes the shape of the gap. A denominator argument moves a number by tens of percent; this one needs to move by an order of magnitude.

What the flat line is hiding

The aggregate is the least useful number in the report. The composition underneath it has changed a lot, and not in the direction the plan intended.

• The median round shrank. ¥77.6 million to ¥62.4 million in a year, across 3,743 deals.
• Capital moved late. Series C was the largest single stage at ¥192.4 billion.
• Capital moved old. Companies founded seven or more years ago took 45.6% of the money. Companies under three years old took 18.2%.

Read those together and the flat headline resolves into something sharper: a smaller number of larger cheques going to older companies, while the typical early round gets smaller. That is a market becoming more selective, which is a normal thing for a venture market to do after a correction, but it is the opposite of an ecosystem widening at the base, which is what “100,000 startups” describes.

It also explains why founders and the government can look at the same ¥761 billion and disagree about whether anything is working. If you run a Series C company, the capital is there. If you are raising your first ¥50 million, it left.

The plan survived a change of government

The more surprising finding is institutional rather than financial.

The Startup Development Five-Year Plan was a product of the Kishida government’s “New Capitalism” apparatus. That apparatus is gone: the New Capitalism Realization Council was wound up in late 2025, and its work moved to the Japan Growth Strategy Headquarters under the current government (Cabinet Secretariat).

Ordinarily that is how a Japanese policy package quietly dies — the council that owned it is dissolved, the branding changes, and the targets are never mentioned again.

That is not what happened. A startup policy subcommittee was stood up under the new headquarters and met four times between February and May 2026, publishing a package on 20 May titled スタートアップ総力創出パッケージ — roughly, an all-hands startup creation package, subtitled “generate, grow, deploy” (Cabinet Secretariat).

So the substance was carried across an administration and a rebrand, with two years left on the clock. For anyone deciding whether to build a Japan strategy on the assumption that startup policy continues past the current prime minister, that is a more useful data point than the funding total. Targets get missed everywhere. Institutional continuity across a government change is rarer, and it is the thing that determines whether the 2028 date is followed by a second plan or by silence.

The reform that will actually bite is on the exit

While the funding target dominates the coverage, the Tokyo Stock Exchange has been rewriting the rule that shapes Japanese startup behavior more than any subsidy does.

Under the revision, a Growth Market company’s market capitalization must be ¥10 billion or more five years after listing. The current standard is ¥4 billion after ten years. It applies as of the end of the first fiscal year ending on or after 1 March 2030 (Japan Exchange Group).

Japan’s structural peculiarity has long been that its startups list early and small, an IPO at a valuation that would be a Series B elsewhere, after which the company stops behaving like a startup and starts behaving like a small listed company with a quarterly reporting burden. The old maintenance bar was low enough to make that a viable end state.

Raising the bar to ¥10 billion at five years removes the soft landing. A company that lists early now has to grow into an institutional-scale valuation on a fixed clock, or move to the Standard Market. The intended effect is to push companies to stay private longer and raise larger private rounds — which is, not coincidentally, exactly the behaviour that would move the ¥10 trillion line.

The plan’s money target and the exchange’s listing rule are pointed at the same problem from opposite ends. Only one of them has an enforcement mechanism.

What to watch

Three things worth tracking between now and March 2028:

1. Whether the median round recovers. The aggregate is dominated by a handful of large late-stage deals; the median is the honest read on whether the base of the ecosystem is widening.
2. Whether the May 2026 package converts into deployed capital rather than announced capital. Committed government money and money that has reached a founder’s balance sheet are separated by a long lag in Japan.
3. How Growth Market listings behave as 2030 approaches. Companies that would have listed early in 2027 have a reason not to now. If listings slow and private rounds get larger, the exchange rule is working as designed and the funding line will move for structural reasons rather than promotional ones.

A note on the unicorn target: counts differ substantially between trackers depending on how each defines and verifies a private valuation, and the published figures for Japan are not reconcilable to a single number. The funding total is the cleaner scoreboard, which is why it is the one used here.

---

Tech for Impact Summit 2027 takes place in Tokyo on 18–19 May 2027, about ten months before the plan’s deadline, which makes it a good vantage point on whether any of this landed. T4IS is an invitation-only executive summit and a partner event of SusHi Tech Tokyo, convening leaders across business, policy and capital around technology aimed at the world’s most urgent problems.

If you are building, funding, or regulating in this space, you can join the 2027 waitlist at tech4impactsummit.com/apply.