NextFin News - Hong Kong's equity fundraising machine just logged its strongest summer on record, powered by a rush of artificial-intelligence listings and a reform package that has repositioned the exchange as the default venue for China's technology capital raise. Initial public offerings in the city doubled in the first nine months of 2026 to US$48.4 billion, the highest nine-month haul since records began in 1980, with AI and hard-technology companies accounting for a dominant share of the largest deals and the pipeline still building into year-end.
The rally is not a single-theme spike. Twelve companies spanning the AI value chain - from chip designers to generative-model platforms - listed in the December-January window alone, and the pace held through the summer months, when quarterly capital raised across IPOs, placements and block trades reached US$45.5 billion, just short of the all-time quarterly record set in 2021. The question for investors is whether this is a cyclical catch-up from a depressed base or the first leg of a structural shift in where China's technology companies choose to price their equity.
The Numbers: A Record Built on Breadth, Not One Megadeal
The headline figure carries the weight of history. Through the first nine months of 2026, 112 companies raised US$48.4 billion on Hong Kong's main board, with two further listings on the Growth Enterprise Market, according to market data released by the London Stock Exchange Group. By deal count, the combined total of 114 new listings rose 72 percent year on year - the busiest nine-month period since 2018, the last time the city's IPO engine ran this hot.
The quarterly rhythm shows the acceleration was sustained, not front-loaded. The first quarter alone produced 40 new listings raising HK$110.4 billion, the second-highest first-quarter fundraising volume in the exchange's history, with technology, media and telecommunications - dominated by AI - accounting for 55 percent of total funds raised. Hong Kong hosted eight of the ten largest TMT IPOs globally in that quarter. The first half closed at HK$209.9 billion across 85 IPOs, a 92 percent jump from the same period a year earlier and the strongest first-half performance in five years. The third quarter added another US$45.5 billion across primary and secondary issuance, leaving the full-year trajectory within striking distance of the HK$380 billion forecast from PwC.
Two comparisons put the scale in perspective. First, the nine-month total already exceeds the city's full-year fundraising in each of the three prior years: roughly US$6 billion in the 2023 trough, about US$11 billion in 2024, and US$37.4 billion across 119 listings in 2025. Second, the market is broadening behind the blockbusters: the ten largest offerings in the first half accounted for 44 percent of proceeds, a concentration that has been falling as mid-tier issuance fills in - a healthier footprint than a market leaning on a handful of anchor deals.
"New AI companies are going public in Hong Kong, [and] investors can now - for the first time - look beyond AI proxy stocks to direct investment opportunities in frontier companies from across the China AI value chain."
Johnson Chui, the exchange's head of global issuer services, said after the year's first wave of AI listings. The roster bears him out. Chip designer Shanghai Biren Technology raised HK$5.58 billion in the city's first listing of 2026 and closed its debut 76 percent above its HK$19.60 offer price. Generative-model platform Zhipu AI priced its January offering at US$558 million and returned to the market in July with a share sale that raised about US$4 billion. Alibaba-backed MiniMax followed a day later with a US$618 million raise - about US$711 million with the over-allotment option fully exercised - that more than doubled its shares on the first trading session.
The breadth matters as much as the size. Thirteen specialist-technology companies listed in the first half of 2026 alone, more than the eight that listed across the previous three years combined, and 24 A-plus-H dual listings were completed in the same period - already more than the 19 recorded across all of 2025. Mainland companies raising in Hong Kong contributed HK$121.7 billion of the HK$209.9 billion first-half total.
The Reform Engine: What Changed on July 24
Beneath the deal flow sits a rulebook rewrite that took effect on July 24, 2026, when the exchange published the conclusions of its Listing Framework Competitiveness Review. The reforms are technical in the details but strategic in intent: they lower the friction for exactly the kind of issuer that has been sitting on the sidelines.
Three changes matter most. First, the market-capitalization threshold for secondary listings of companies without weighted-voting-rights structures was cut from HK$10 billion to HK$6 billion, while preserving an alternative route at HK$3 billion for issuers with a five-year track record. That reopens the door for large overseas-listed Chinese companies that fell between the old thresholds. Second, the exchange codified its discretion to accept weighted-voting-rights structures with a minimum underlying economic interest as low as 5 percent of issued share capital - replacing the old "minimum shareholding percentage" test and excluding treasury shares from the denominator. Third, the confidential-filing process was converted into a formal "non-public filing" regime, giving applicants a clearer path to test the market without triggering immediate public disclosure, along with a new right to voluntarily publish an updated application proof.
The exchange also streamlined the guidance around changing listing status - moving from a secondary to a dual-primary or primary listing - after acknowledging that its existing rules on the migration, primary-conversion and overseas-delisting routes were "unduly complicated." A second phase of consultation is already earmarked for continuing obligations, further overseas-issuer facilitation, and reform of the Growth Enterprise Market and the SPAC framework.
The reforms did not go as far as some market participants wanted - the exchange held the line on financial thresholds for weighted-voting-rights and secondary listings despite calls for further loosening - but the direction is unambiguous: the venue is deliberately making itself cheaper, faster, and more forgiving for technology issuers.
Cyclical Wave or Structural Shift?
Here is the judgment the market has not fully priced. The AI fundraising surge is a cyclical wave riding on top of a structural shift - and confusing the two is how investors lose money in IPO cycles.
The cyclical leg is the easier call. IPO volume is mean-reverting by nature, and this rally follows one of the deepest troughs in the exchange's modern history. The 2023 trough, at roughly US$6 billion, was one of the weakest annual totals the market has seen in decades; 2024's US$11 billion was a partial recovery; 2025's US$37.4 billion compounded it; and 2026's pace builds on all three. Valuation multiples for AI assets have expanded globally, liquidity has returned to Asian primary markets, and a backlog of issuers that could not price in 2023 and 2024 is clearing in a narrow window. Every one of those drivers is transient. When the backlog clears and the multiple cycle turns, volume will fall. That is not a bear case; it is arithmetic.
The structural leg is different, and it is the more important call. The companies listing are not cyclical issuers timing a window - they are Chinese AI and hard-technology companies that have been systematically priced out of the United States by geopolitical and regulatory constraints. The exchange's own data shows the rerouting: twelve AI value-chain companies listed in the December-January window, specialist-technology listings in the first half of 2026 outnumbered the prior three years combined, and 145 technology companies sat in the public filing pipeline as of late June. This is a rerouting of the primary listing destination for an entire sector, not a timing trade. It is reinforced by the reform package, by the city's policy push toward national strategic sectors including AI, semiconductors and quantum technology, and by the dual-listing pattern that embeds Hong Kong into the capital strategy of mainland technology issuers rather than leaving it as an overflow valve.
The mechanism, then, is two-channel: a cyclical liquidity-and-backlog channel that will revert, and a structural venue-shift channel that will not. The market is currently pricing the cycle; the structural channel is the underappreciated variable.
The Second-Order Bet: Hong Kong as the Pricing Venue for China's AI
The first-order effect of the AI listings is obvious: more deals, more fees, a higher global ranking. The second-order effect is what actually matters for cross-asset investors. If Hong Kong becomes the default pricing venue for China's AI sector, it becomes the reference market for the entire asset class outside the United States - and that pulls capital, research coverage, and derivative liquidity into the city's ecosystem in a self-reinforcing loop.
That loop has a concrete transmission channel. Direct AI listings replace "proxy stocks" - the listed suppliers, distributors, and partial-exposure names that investors previously used to get AI exposure - with pure-play comparables. Once pure plays exist, analysts can build sector-specific multiples, index providers can construct AI baskets, and options markets can price single-name volatility. Each step lowers the cost of capital for the next issuer in the pipeline. The exchange's own data shows the first step already landed: TMT fundraising reached 55 percent of the first-quarter total, and eight of the ten largest global TMT IPOs in that quarter cleared in Hong Kong.
But the second-order effect cuts the other way, too. A crowded cohort of AI issuers pricing in the same window creates correlation risk. If post-listing performance disappoints across the cohort - if the debut pops fade and the stocks trade down in unison - the entire sector's cost of capital rises together, and the pipeline that underpins the 2026 record thins out faster than a single-name miss would suggest. The market is not yet pricing that correlation; it is underwriting each deal on its own story.
The Counter-Thesis: A Front-Loaded Pipeline and a Mean-Reverting Base
The strongest case against the structural read is straightforward: 2026 is a catch-up year built on a depressed base and a finite backlog, and the numbers will look very different once the AI darlings have been absorbed. The bear case rests on three observable facts. First, the nine-month record still trails Nasdaq's US$144 billion over the same period, a gap widened by a single transaction - the roughly US$86 billion SpaceX offering in June, the largest IPO in history. Second, the third quarter's US$45.5 billion across IPOs, placements and block trades fell just short of the US$47.6 billion quarterly record set in 2021, at the peak of the last tech mania - suggesting the rally is approaching a ceiling, not breaking through one. Third, the reform package deliberately held the line on the most significant loosening that issuers had demanded, signaling that the exchange is managing supply, not unleashing it.
The bear case is internally coherent and deserves its weight. But it conflates volume with venue. Even if 2027's total fundraising halves from 2026's pace - a plausible cyclical reversion - the question of where the next Chinese AI company lists is already answered by the venue shift embedded in the listing data. A thinner market can still be a structurally different market. The counter-thesis would be proven right not by lower volume alone, but by a reversal of the venue choice: if Chinese AI and specialist-technology issuers begin choosing Shanghai's STAR Market or Shenzhen's ChiNext over Hong Kong for their primary listings again, the structural call fails.
What to Watch: The Signals That Settle the Debate
The forward view splits cleanly by time horizon, and each horizon has its own falsifying signal.
In the short term - the next two quarters - the market is liquidity-driven. Watch the debut-day performance of new AI listings: if the median first-day gain falls below 10 percent for three consecutive months, the window is closing and the cyclical leg is turning. The third-quarter data already shows the pace moderating relative to the 2021 peak, so the direction of travel matters more than the absolute level.
In the medium term - through 2027 - the base case is a lower but still elevated volume, with full-year 2026 fundraising tracking toward the HK$380 billion PwC forecast, followed by a cyclical step-down. The upside case is that the reform package's second phase - continuing obligations, overseas-issuer facilitation, and Growth Enterprise Market and SPAC reform - unlocks a second wave of dual-primary conversions and international listings. The downside case is that a coordinated drawdown across the 2026 AI cohort freezes the pipeline and pushes the next meaningful wave into 2028.
In the long term, the structural question resolves on one metric: the share of Chinese AI and specialist-technology primary listings choosing Hong Kong. The falsifying signal is specific - if that share falls below 50 percent for two consecutive years, or if fewer than ten Chinese AI companies list in Hong Kong in any full year after 2026, the structural-shift thesis is wrong and this was a cycle after all. Conversely, if the share holds above 70 percent through 2027 even as total volume reverts, the venue shift is durable and the current record is merely the first data point.
The HK$380 billion full-year forecast is achievable but leaves no margin for a risk-off turn in global liquidity; it should be read as a ceiling case for the cycle, not a floor for the structural trend.
The bottom line: Hong Kong's record summer is real, but the record is the least interesting part of the story. The durable change is that China's AI sector now has a home market for equity capital outside the United States, and that shift will outlast the deal cycle that revealed it. Investors underwriting the next AI listing on the strength of 2026's volume are betting on the cycle; investors watching the venue-share metric are betting on the structure. Only one of those bets survives the mean reversion that is already underway.
Explore more exclusive insights at nextfin.ai.
