TSMC (() surged more than 5.49% during ) trading today the 27th, breaking through the 2300 level and edging close to its own stock code 2330. On April 23, the Financial Supervisory Commission announced that, starting April 24, it would loosen the domestic stock concentration limits for domestic equity mutual funds and actively managed ETFs that invest in a single company, raising the cap from 10% of the fund’s net asset value to 25%.
The term “TSMC clause” comes from the fact that among the more than 1,000 constituent stocks in Taiwan’s weighted index, TSMC is the only individual stock that meets the target: TSMC’s weight in the broader market exceeds 10%, and its weight in Taiwan stocks is as high as 44.3%. There are 138 domestic equity mutual funds investing in Taiwan stocks, with assets totaling NT$48.1B, while actively managed ETFs currently total 14 in number with a combined size of NT$231 billion. This article attempts to break down the impact of ETF buy orders on Taiwan stocks from the perspective of TSMC’s trading volume.
In the past, due to the holding limit for ETFs, they could only take a detour to buy TSMC.
To understand the significance of this relaxation, you can look back at the “00981A wrapper controversy” that was just hotly debated in January 2026. At the time, the actively managed ETF issued by Uni-President Securities actively made money from Taiwan stocks growth (00981A), because it bought large quantities of passive ETFs such as Yuanta Taiwan 50 (0050) and Fubon Technology (0052). It was criticized by well-known trader “Giant J”—calling it “ETF wrapper,” “asking a Michelin chef to serve McDonald’s,” and “peeling investors’ skin in two layers.”
(00981A profit of 61% per year yet still got slammed? Giant J attacks ETF wrapper, peeling off 2 layers! Experts, however, have a different view)
But at the time, researchers Yu Zhe’an and Freddy pointed out that the essence of this operation is to buy TSMC via a detour: under the 10% limit, if you want to maintain exposure close to the broader market, allocating to 0050 and 0052 with high content concentration is actually a legitimate, and also the lowest transaction-cost shortcut.
This view was directly confirmed by Vincent, co-founder of Man News and former executive director of asset management at J.P. Morgan: “The holding limit is a pain point for many funds. In my prior work experience, our team also very often tried to find ways to make fund performance keep up with TSMC’s stock price trend. They know they should buy more TSMC, but because of the holding limit, they have to endure the pain of trimming.”
In other words, before 4/24, the entire industry did everything it could to take detours—buy passive ETFs to build up content exposure, buy ADRs, and use structured products for exposure. Previously, fund managers effectively “chased momentum through content” by buying high content-concentration holdings like 0050 and 0052. After 4/24, they can buy TSMC openly and directly, and can buy up to 25%.
How big is the impact of ETF buy orders on TSMC?
As for how much impact there will be specifically, market-circulated research estimates that with existing scale at the end of March 2026, the available room for immediately realized additional allocation is actually: conservative NT$4.7 billion, mid estimate NT$30.4 billion, and optimistic NT$95.0 billion. Even if you extend the timeframe to 12 months and assume fund size continues to grow at a CAGR of 20–30%, the most optimistic additional allocation room would only reach NT$159.7 billion.
One conclusion of the study is: “In the optimistic scenario, total additional allocation after 12 months would be NT$159.7 billion, which is less than 0.3% of TSMC’s total market value (NT$56.66 trillion). The real significance of the relaxation lies in ‘removing institutional constraints,’ not ‘creating buy demand.’”
Switching to a trading-volume perspective, the story is completely different. According to official Taiwan Stock Exchange data, TSMC’s full-year trading value in 2025 was NT$10.30 trillion, with a daily average of about NT$41.9 billion. After the stock price pushed into the NT$2,000 range in 2026, January’s monthly trading value reached NT$1.46 trillion, and on April 24 alone, the trading value was about NT$112.9 billion. At the current price range, the daily average trading value is about NT$75–80 billion, and the estimated trading value for April alone could exceed NT$1.5 trillion.
Comparing this baseline with various additional-allocation valuation assumptions reveals the true magnitude of the impact: the mid estimate NT$30.4 billion derived from objective data equals the trading value of 0.43 of a normal trading day, or 2.1% of a single month’s trading value. This is the scale managers can “hide in volume.” The optimistic NT$95.0 billion is 1.33 trading days, or 6.6% of monthly volume; the NT$200 billion estimated by institutions is 2.8 trading days, or 13.8% of monthly volume.
If these buy orders are released in a concentrated way within 1–2 months after the relaxation, then under the optimistic scenario, the optimistic case could consume 6–13% of one month’s trading value.
What’s more, the buy orders come not only from funds. “Retail investors who hear the news” are actually the real multiplier effect. The market expects that after the information ferments over the weekend, the “expected psychological buy demand” next Monday could be more intense than the buy demand from investment trust funds after they revise their allocations.
Where does the money come from? Stocks other than TSMC would likely be bled
When everyone’s money crowds into TSMC, these funds must come from positions in other stocks. The research also decomposed the impact in the same direction: if the relaxation is implemented immediately and no new capital enters, then in the mid estimate scenario there would be an average -3.0% deallocation pressure on other stocks. The most directly hurt are the stocks in the AI supply chain that funds were originally overweighting.
After the relaxation, it is estimated that the “stuck group” of mutual funds and actively managed ETFs with a total size of NT$921 billion, accounting for 76% of the parent population—these 93 funds/ETFs have long underweighted TSMC due to the 10% limit, and now only have room to fill in.
The research believes the top 5 additional-allocation funds are:
Allianz Taiwan Technology (size NT$141.3 billion, estimated additional allocation NT$4.39 billion)
00981A Uni-President Active Taiwan Equity Growth (size NT$99.8 billion, additional allocation NT$3.11 billion)
Cathay Small Dragon (size NT$24.1 billion, additional allocation NT$1.88 billion)
Cathay Cathay Fund A (size NT$10.8 billion, additional allocation NT$1.14 billion)
Allianz Taiwan Great Dam A (size NT$53.5 billion, additional allocation NT$1.04 billion)
And in terms of the holdings ranking among these funds that apply to the new rules, the top 5 non-TSMC holdings are all AI supply-chain stocks:
Wangsyn (6223) holds NT$68.9 billion
ASE Technology? / TAI 光電 (2383) NT$68.88 billion
Delta Electronics (2308) NT$59.42 billion
Qisda? / 奇鋐 (3017) NT$51.70 billion
InnoLux? / 穎崴 (6515) NT$48.76 billion
In the mid-estimate scenario with immediate realization, each of these stocks could face an individual relative deallocation pressure of between NT$1.4 and 2.1 billion.
The research also proposed a buffering mechanism: over a 12-month time dimension, if fund size can grow by a CAGR of 20% or more, then new money entering the market (conservatively estimated at NT$241.4 billion) would be sufficient to fully absorb all incremental allocation needs, without having to sell other holdings.
The time dimension will determine whether it is a “real bleed.” If, during the relaxation implementation period, fund size continues to receive inflows, the deallocation pressure on other stocks would be diluted by new money. But if there is market agitation early in the relaxation period and investment trusts do not receive new capital inflows, then the pressure to sell the AI supply-chain holdings to reallocate into TSMC would become directly apparent.
If you hold strong positions that are not TSMC, you need to assess the risk that fund managers may prioritize selling those positions to switch into other allocations—especially AI supply-chain stocks that have performed relatively weaker. If you have idle cash on hand, high-quality stocks besides TSMC could appear as relatively cheap entry points during this bleeding cycle. Having cash ready and waiting for fund selling pressure to hit a short-term “wrong-side” mistake may actually be the opportunity for a contrarian layout.
From the perspective of institutional estimates, NT$200 billion in new capital liquidity is negligible when viewed by market value; from objective data, the real “4/24 incremental” is only NT$30.4 to 95.0 billion. From a trading-volume perspective, this level of magnitude is already enough to push the stock price upward in the short term.
This article “TSMC’s stock price breaks through the 2300 mark and closes in on 2330! Deconstructing the real impact of the TSMC clause ETF buy orders” first appeared on Lian News ABMedia.
