Japan's PTS Cap at 10%, Its Actual Market Share at 19.5%: The AI Rally Broke a Regulatory Rule Before Anyone Noticed

The Nikkei 225 touched 70,000 for the first time in June 2026. That is the headline. But the number that changes the analysis is smaller and has been invisible until now: in June 2026, Japan's alternative trading systems handled 19.5% of total listed-share trading - nearly double the 10% cap the Financial Services Agency set to prevent market fragmentation.

The rule was already broken. The FSA is only noticing because the AI chip rally forced the volume past a threshold the regulator couldn't ignore.

Decomposition

Three structural shifts are running simultaneously, and the market is conflating them into one narrative.

The AI rally is the catalyst. The Nikkei broke through 50,000 in October 2025, reached 60,000 in April 2026, and touched 70,020 in June - covering three major psychological levels in under eight months. Kioxia Holdings... became Japan's largest listed company by market cap by June 12, overtaking Toyota in 18 months. SoftBank's FY2025 net profit exceeded ¥5 trillion ($31 billion), a first for a Japanese company, driven by AI investment returns. AI and semiconductor names carry outsized weight in the index, which is why the headline number looks so dramatic.

The retail revolution is the capital source. As of late June 2026, Japanese individual investors held over ¥16 trillion ($99 billion) in brokerage accounts. Margin buying balances on the Tokyo and Nagoya exchanges climbed to ¥6.39 trillion as of May 29 - the highest reading since records began in December 1994. Retail investors now account for 25% of Japan's total equity trading value in fiscal 2025, the largest share in 12 years. Margin trading volume rose 39% year-over-year in FY2025 to approximately ¥618 trillion ($4 trillion). Younger traders using tax-efficient investment accounts drove a meaningful portion of that growth.

But retail money doesn't trade evenly. The margin concentration is heavily clustered in semiconductor and electronics names - Kioxia, Murata Manufacturing, Ibiden. That is the vulnerability. When margin buying clusters in a handful of stocks, the unwind risk is not distributed across the market. A sharp move in one name forces margin calls, which trigger forced sales, which push prices lower, which trigger more margin calls. The July 28 flash crash - when the Nikkei briefly plunged by more than 2,000 yen... targeting artificial intelligence and semiconductor-related stocks - was a dress rehearsal for exactly this dynamic.

The PTS shift is the plumbing nobody was watching. PTS (Proprietary Trading System) is a privately run market for shares listed on the Tokyo Stock Exchange. Japan has three PTS operators: Japannext, JAX (launched in January 2025 by a Rakuten-led consortium), and Osaka Digital Exchange. Cboe Japan ended its Japanese stock PTS in August 2025.

But in June 2026, PTS trading reached 76.2 trillion yen - 19.5% of total listed-share trading. Both figures are record highs. JAX alone holds 59% of the PTS market share as of October 2025. JAX handles over 7% of total Japanese market volume, according to JPX and JSDA data.

The FSA currently limits PTS platforms to no more than 10% of total Japanese stock trading value. The limit was designed to prevent market fragmentation - the concern being that if too much trading moves off the main exchange, liquidity fragments, price discovery degrades, and the TSE loses its function as the central market. The limit has not been enforced at the hard-cap level because the regulatory framework separates PTS from full-fledged exchanges, and the FSA has encouraged PTS growth to ease trading concentration at the TSE. In practice, the 10% figure functions as a policy ceiling the system has quietly exceeded.

The Gap Between Narrative and Structure

The competitor framing - Japan considers raising the trading cap as AI stocks surge - makes it sound like a regulatory response to growth. The structural reality is that the PTS system was already expanding before the AI rally became dominant. JAX launched in January 2025. The PTS infrastructure was being built. The AI rally just drove enough volume through it to make the old rule impossible to ignore.

The FSA will explore changing the rule that separates proprietary trading systems from full-fledged exchanges, Nikkei reported on August 6. A study group of outside experts is expected to form as early as this fiscal year.

But the question is not just whether to raise a cap. The question is what the cap change means for market structure. If PTS volume is allowed to grow further, the TSE's share of total trading will continue to shrink. The night session - which runs from 5 p.m. to as late as 11:59 p.m. on PTS platforms - will become a larger fraction of the day's price discovery. Margin trading, currently restricted to hours overlapping with the TSE during night sessions, could expand.

Each of these changes moves Japan closer to the US model, where alternative trading systems handle a substantial share of equity volume and night/extended-hours trading is normalized. It also moves the market further from the auction-based price formation that the TSE currently uses for its opening and closing prices.

The contrast is instructive. In the US, retail-driven momentum rallies in 2020–2021 created similar structural questions around dark pools, ATS fragmentation, and retail leverage. The regulatory response was enforcement actions against specific brokers and platforms, not a redesign of venue caps. Japan's approach - adjusting the structural cap itself - suggests the FSA is thinking about market architecture rather than behavioral policing.

The Concentration Risk

Here is the part the headline doesn't capture. Daily average trading value for the TSE Prime Market reached ¥6.7 trillion in FY2025, and hit ¥9.13 trillion in March 2026. That is a 36% increase in three months. Derivatives trading value reached ¥3,994 trillion for the full FY2025 and ¥587 trillion in March 2026 alone - both second-highest records on the full-year basis, and a record for the monthly figure.

Volume is surging. But it is not distributed evenly across the market. The AI-driven rally reshaped the Topix index composition, edging out smaller firms as semiconductor equipment and chip designers gained weight. The margin concentration in Kioxia, Murata, and Ibiden means the same handful of names that drive the index also carry the most leverage.

When the Nikkei plunged 2,000+ points on July 28, the sell orders were almost entirely AI and semiconductor names. The market recovered, but the mechanism - concentrated positions, leveraged retail, thin night-session liquidity - has not changed. If a correction comes, PTS night trading could amplify it, not dampen it. A stock that drops sharply after only a small number of shares change hands in the thin overnight session sets the opening price the next morning before anyone has read the full picture.

What to Watch

• The FSA study group's mandate and timeline. Will it focus on raising the cap, or will it address PTS night-session liquidity, margin rules, and the gap between PTS and exchange price discovery? The scope of the study determines whether this is cosmetic or structural.
• PTS night-session margin rules. Currently, margin trading during night hours is limited to TSE-overlapping hours. If that restriction is lifted, leveraged trading on thin overnight liquidity becomes the norm - and the July 28 flash crash mechanism becomes a repeatable risk.
• Retail margin balances relative to TSE Prime Market volume. If margin balances hold above ¥6 trillion while total market volume remains elevated, the system is structurally leveraged. A sustained drop below ¥5 trillion would signal de-risking.
• JAX market share trajectory. At 59% of PTS volume and 7% of total market volume, JAX is the dominant alternative venue. If it continues gaining share, the practical effect of any FSA cap adjustment flows disproportionately through Rakuten's infrastructure.
• Semiconductor margin concentration. The risk is not in the aggregate margin number. It is in which names carry it. If Kioxia, Murata, and Ibiden remain the top three margin concentration points, a sector-specific correction will look like a systemic event.