🇭🇰Hong Kong: HKEX Listing and Dim Sum Bonds

Concept

The Hong Kong exchange is not one door but six parallel listing routes with market capitalisation thresholds running from HK$150 million to HK$20 billion, and the route determines both the cost of the deal and whether the founder still controls the company after it.

The direct answer for anyone deciding now. A profitable company goes through the Main Board profit test: profit of ≥HK$35 million in the most recent year and ≥HK$80 million in aggregate over three years, with market capitalisation ≥HK$500 million (thresholds in force since 1 January 2022). A large company without profit goes through the market cap / revenue test: revenue ≥HK$500 million and market capitalisation ≥HK$4 billion. Pre-revenue biotech goes through Chapter 18A at ≥HK$1.5 billion. Other pre-revenue technology goes through Chapter 18C: ≥HK$4 billion with revenue of at least HK$250 million, or ≥HK$8 billion without it (the reduced thresholds run from 1 September 2024 to 31 August 2027). A company already listed on NYSE, Nasdaq or the LSE goes through Chapter 19C: from HK$3 billion with five years of compliance, or from HK$6 billion with two. The realistic all-in budget (estimate) runs from 8.2–12.4% of proceeds on an HK$800 million deal to 1.4–2.9% on an HK$12 billion deal. The cycle is around seven months from kick-off to listing, and four to six months for an A+H dual listing on the accelerated track. The debt alternative in the same city is a dim sum bond in offshore renminbi: three to six weeks and 25–50 basis points all-in.

What follows is the detail, and one warning that governs all of it: almost everything written about HKEX before 2026 is obsolete. The rules have been rewritten three times in thirteen months.

Map of HKEX routes: the threshold matrix

A listing route is a set of alternative qualification tests under Chapter 8 of the Listing Rules plus the special Chapters 18A, 18C, 19C and 18B, and the issuer picks between them rather than the exchange picking for it.

Sources for the thresholds: PwC, HK Listing Rules 2025 (PDF); Charltons on the profit test increase from 1 January 2022; Morgan Lewis on the GEM reform effective 1 January 2024; Skadden on the 18C and SPAC threshold cuts from 1 September 2024; HKEX announcement of 24 July 2026; Charltons on the rules that took effect on 24 July 2026.

Main Board: what the exchange actually tests

The Main Board is the baseline Chapter 8 route with three alternative financial tests, and passing any one of them is enough.

The thresholds are an entry filter, not the substance of the review. Applications are not turned away for missing profit by HK$5 million; they are turned away on three other grounds. The first is business sustainability: the exchange tests whether track-record revenue was manufactured by one-off transactions, related parties or customer concentration that will fall apart after listing. The second is continuity of ownership and management: three years of continuous management and the most recent audited year of continuous ownership. Since 24 July 2026 that rule has been codified more softly — a change of controlling shareholder within the track record period is acceptable provided there was no material change of management influence in the last audited year before listing. The third is suitability as a free-standing discretion: a business model that depends on a single regulator's permission, litigation and tax tails, opaque pre-IPO restructuring.

Accounts are accepted under HKFRS, IFRS or, for PRC issuers, CAS; since 24 July 2026 US GAAP is permitted more widely, for subsidiaries of US-listed groups and companies with substantial US operations, subject to reconciliation and disclosure of differences. The board needs at least three independent non-executive directors making up at least one third of its members, plus a mandatory audit committee and remuneration committee. The latest reporting period must be no more than six months old at the date of the listing document, which is exactly why a preparation process that drags on triggers a re-audit.

Jurisdiction of incorporation: Hong Kong, the PRC, Cayman and Bermuda are recognised by default, alongside a list of Acceptable Jurisdictions each with its own Country Guide (SFC/HKEX Joint Policy Statement). The standard construction for a founder who is not a PRC resident is a Cayman or BVI holding company; the choice of vehicle is covered in the material on holding structures.

Chapter 18A and 18C: companies with no revenue

Chapters 18A and 18C are two special routes for companies that pass none of the Chapter 8 financial tests, where instead of profit the exchange assesses R&D, product stage and the presence of professional investors.

18A (biotech) is the more road-tested of the two. Market capitalisation of HK$1.5 billion, working capital covering 125% of the next 12 months' costs, a core product past concept stage, and at least one Sophisticated Investor that does not exit at the IPO. The route works: 2025 saw 14 pre-revenue biotech listings against 4 in 2024 (KPMG, 10 December 2025), and a cumulative 88 biotech and specialist technology companies had listed in Hong Kong under the two chapters since 2018 (count as at 22 December 2025) (HKEX, 22 December 2025).

18C (specialist technology) launched on 31 March 2023 and was written off as dead for three years: eight listings in total across 2023–2025. The break came in 2026. By 31 March 2026, 14 companies had listed under Chapter 18C raising HK$28.4 billion (HKEX Insight, "18C, Explained"), and the first half of 2026 alone produced 13 Chapter 18C listings — more than the previous three years combined and 14% of all funds raised on the exchange (KPMG, June 2026). Three causes: the temporary threshold cut to HK$4 billion / HK$8 billion from 1 September 2024, the launch of the TECH channel (Technology Enterprises Channel) on 6 May 2025 with a dedicated support team and a presumption that applicants meet the innovative company requirements, and an accumulated pipeline of Chinese hardware and AI.

One practical point to keep in the calendar: the reduced 18C thresholds expire on 31 August 2027. Companies valued between HK$4 billion and HK$6 billion face a hard filing deadline. Chapter 18C also imposes post-IPO disposal restrictions keyed to market capitalisation: at ≥HK$15 billion, 20%/25% of the holding; at HK$15–30 billion, 15%/20%; above HK$30 billion, 10%/15%. The mechanics are the same as for the ordinary controlling shareholder lock-up, set out below.

GEM and SPAC: two routes that do not work

GEM and SPAC are formally live HKEX routes, both reformed and both failing to deliver the deal flow that was expected of them.

GEM became markedly friendlier after the reform of 1 January 2024: quarterly reporting abolished, the annual report deadline extended from three to four months, the controlling shareholder lock-up cut from 24 to 12 months, the compliance officer role made optional. An alternative R&D test was introduced: market capitalisation ≥HK$250 million, revenue ≥HK$100 million over two years with growth, and R&D ≥HK$30 million running at no less than 15% of opex in each year. A streamlined transfer to the Main Board was opened as well: three full years on GEM, a clean compliance record and average daily turnover of ≥HK$50,000 on at least half the trading days in a 250-day reference period — with no sponsor and no prospectus. None of this restored deal flow: on 24 July 2026 HKEX announced a second phase of consultations with GEM reform carved out as a separate item.

SPAC (since 1 January 2022) requires gross proceeds of ≥HK$1 billion at a share price of ≥HK$10, a minimum of 75 professional investors of whom 20 institutional investors hold 75% of the shares placed, and a promoter with an SFC Type 6 or Type 9 licence holding at least 10% of the promoter shares. De-SPAC: announcement within 24 months, completion within 36 months of listing. An independent PIPE investment is mandatory on a sliding scale from 25% for a deal below HK$2 billion to 7.5% at ≥HK$7 billion, and since 1 September 2024 it is the lesser of that percentage and HK$500 million (Charltons on the SPAC regime).

Why it never took off. The Hong Kong SPAC was designed from the start as the opposite of the American one: before de-SPAC the security is closed to retail, so there is none of the speculative demand that created promoter economics in the US. The mandatory PIPE effectively hands valuation validation to an institutional investor, meaning the promoter has to bring in the same investor who would have subscribed to an ordinary IPO — but with two layers of cost and a hard time limit. Finally, the 2022–2024 window coincided with the worst period the Hong Kong market has seen, and by the time it turned in 2025–2026 an ordinary IPO had become faster and cheaper than a SPAC. The upshot: SPAC reform has also been pushed into the second phase of the 2026 consultations.

Secondary listing, dual primary and weighted voting rights

Chapter 19C is the route for a company already listed on a Qualifying Exchange (NYSE, Nasdaq, LSE Main Market Premium), letting it enter Hong Kong without the full Chapter 8 cycle.

The reform of 24 July 2026 cut the thresholds by roughly half across every branch: Criteria B without WVR fell from HK$10 billion to HK$6 billion at two years of compliance; with WVR, Test A fell from HK$40 billion to HK$20 billion, and Test B from HK$10 billion plus HK$1 billion of revenue to HK$6 billion plus HK$600 million. Guidance on converting a secondary listing into a primary one was simplified at the same time (Morgan Lewis, August 2026).

For a founder, the WVR block matters more. The cap on the voting ratio stays at 10:1, but rises to 20:1 where market capitalisation is ≥HK$40 billion. The minimum stake for a WVR beneficiary drops from 10% of the issue to 5%, provided the economic interest is ≥HK$4 billion. The innovation test has been rewritten into two routes: Route A is technological (substantial R&D, successful IP monetisation, or a market capitalisation disproportionate to tangible assets), Route B runs on the business model (a new model, revenue CAGR ≥30%, a notable industry position, profitability metrics). Route B is genuinely new in that it opens super-voting rights to non-technology companies. Chapter 18A and 18C applicants are presumed to satisfy Route A.

The bedrock Chapter 8A safeguards survive intact: a WVR beneficiary must be an individual who is a director at listing and retains an executive role; the rights lapse on death, departure from the board, incapacity or failure to meet the director requirements; WVR shares cannot be transferred to third parties; and on reserved matters (appointment and removal of independent non-executive directors, voluntary winding-up, changes to the constitutional documents, change of auditor, variation of class rights) strict one-share-one-vote applies. The ticker carries a "W" suffix. At the end of 2025 there were only 31 WVR issuers — 1.2% of the 2,686 companies listed on the exchange (21 primary and ten secondary listings) — though they accounted for 14.7% of market capitalisation (HKEX consultation paper, March 2026, PDF).

What changed in 2025–2026: a dated summary

The regulatory delta is the set of changes to the Listing Rules and adjacent regimes that took effect after August 2025 and are not reflected in most available guides.

The listing parameters are cross-checked against HKEX Consultation Conclusions cp202412cc (PDF), August 2025; Latham & Watkins, August 2025; Charltons on the new ongoing float requirements from 1 January 2026; HKEX, 17 December 2025; HKEX, 21 August 2026; HKEX Consultation Paper, March 2026 (PDF); and GovHK on the stamp duty rate effective from 17 November 2023.

Public float, free float and lock-up

Public float is the proportion of shares in public hands required at listing and on an ongoing basis; free float is the narrower figure that excludes any shares subject to disposal restrictions.

The scale at listing, in force since 4 August 2025: market capitalisation ≤HK$6 billion — 25%; HK$6–30 billion — the greater of 15% or shares worth HK$1.5 billion; above HK$30 billion — the greater of 10% or HK$4.5 billion. For A+H issuers, H shares in public hands must be at least 10% of the H share class or HK$3 billion of market value. A separate initial free float requirement was added: at least 10% of the class with a value of ≥HK$50 million, or ≥HK$600 million outright; 5% for A+H. The key consequence: cornerstone investors' shares and the controlling shareholder's locked-up shares do not count toward free float, so you cannot sell half the issue to a cornerstone pool and still expect to clear the minimum free float.

The controlling shareholder lock-up under Rule 10.07: an absolute bar on disposals for the first six months from the listing date, then six months in which disposals are permitted but not to the point of losing controlling shareholder status. Twelve months in total, and the same for GEM since 1 January 2024. Chapter 18C adds to that: 12 months for commercial companies and 24 months for pre-commercial companies, key persons included. Cornerstone investors are locked up for six months; the proposal to allow staggered release was rejected in the 2025 consultation conclusions. The general mechanics of lock-up, including the structure of carve-outs and organised secondary sales, are covered separately in the materials on lock-up and secondary share transactions.

The IPO calendar and cornerstone investors

The listing calendar is a cycle of roughly 210 days from kick-off to listing, split into four phases with fixed regulatory windows inside them.

The indicative timetable is benchmarked against A&O Shearman, Indicative Timeline for an IPO in Hong Kong. Realistically (estimate): an A+H on the accelerated track takes four to six months, a Chapter 18C listing from a standing start nine to fifteen. Where CSRC approval is needed, the Hong Kong accelerated track runs into the mainland regulator's own timetable.

The sponsor must be appointed no later than two months before the A1 filing and must hold an SFC Type 6 licence. The full team: sponsor, issuer's legal counsel, underwriters' legal counsel, reporting accountant, valuer (a property valuer where real estate is material, an independent business valuer for 18A and 18C), industry consultant (de facto mandatory for 18C), tax adviser, receiving bank, share registrar, IR agency, financial printer, D&O insurer. An A+H adds PRC counsel and a CSRC filing. The requirements around audit in Hong Kong are a separate layer of preparation, and usually the one that dictates when the calendar really starts.

Cornerstone investors sign up before the roadshow, take a fixed amount at the offer price (they take no part in price discovery), are named in the prospectus, receive a guaranteed allocation and accept a six-month lock-up. The typical share is 20–50% of the issue, closer to the upper bound in the large 2025–2026 deals; the de facto ceiling is set by the rule requiring 40% in the bookbuilding tranche and sits around 50–55% of the issue (both figures are market estimates). In Shein's listing on 1 September 2026 the cornerstone pool was about US$383 million — up to 22.5% of proceeds — with Boyu Capital, Tiger Global, General Atlantic, Tencent, Greenwoods, Taikang Life and UBS (SCMP, 24 August 2026). The 2025–2026 trend is the return of international long-only money after a four-year absence: BlackRock, Temasek, Qatar Investment Authority, Fidelity International (SCMP on the cornerstone comeback).

The full IPO budget: what a listing actually costs

The cost of an IPO is the sum of exchange fees, professional fees and underwriting commission; the exchange fees are negligible within it, and the economics are decided by the other two.

Only the first line is public. The initial listing fee is paid up front at the A1 filing on a scale keyed to the value of the securities to be listed: HK$150,000 for an offering of ≤HK$100 million, HK$250,000 at HK$500 million, HK$300,000 at HK$750 million, HK$350,000 at HK$1 billion, HK$550,000 at HK$3 billion and HK$650,000 above HK$5 billion (the cap). The annual listing fee starts at HK$145,000 a year for a nominal value of ≤HK$200 million and rises on a scale (HKEX Main Board Fees Rules, PDF).

Everything else is undisclosed. Neither HKEX, nor the Big Four, nor the compilers of cross-border listing guides publish adviser fees. The ranges below are market estimates assembled from observed deal budgets; the spread within each line comes down to three factors: the complexity of the pre-IPO restructuring (number of jurisdictions, VIE structures, historical cleanliness of title), the volume of re-audit work (moving from a local standard to IFRS or HKFRS doubles the reporting accountant's budget), and whether the issuer is running a parallel mainland track with a CSRC filing.

The arithmetic reads one way only: professional costs are close to fixed, so the cost ratio falls in inverse proportion to deal size. Below HK$500 million of proceeds the costs eat a double-digit share, the offering stops making economic sense, and that — not any rule threshold — is the real minimum entry point for HKEX.

The other half of the equation is the collapse in underwriting commissions, which almost nobody writes about. The average base commission on Hong Kong IPOs in 2025 was 1.5% — the lowest since at least 2000, against a 25-year average of 2.3% and 2.4% across Asia-Pacific. For A-to-H deals the average base commission is 0.9% against 2.2% for ordinary IPOs; CATL paid a base commission of 0.2% plus a 0.6% incentive fee on proceeds of US$5.3 billion. The total fee pool nonetheless reached US$489 million in 2025, the highest since 2021: banks are earning on volume, not on rate (China Daily HK, 11 December 2025). The claim that "an IPO costs 5–7%, as in the US" is off by an order of magnitude for a large Hong Kong deal.

One budget line is regularly forgotten at the decision stage: the recurring cost of being public — annual listing fee, HKFRS audit, independent director fees, IR, D&O insurance, legal support for disclosure and connected transactions. The realistic order of magnitude is HK$8–20 million a year for a mid-sized company (market estimate), and since the reform of 31 January 2026 monthly float returns have been added to it.

Stock Connect and the A+H wave: the economic case for Hong Kong

Southbound Stock Connect is the channel through which mainland investors buy Hong Kong securities, and it is the main economic argument for listing in Hong Kong rather than Singapore or London.

The eligibility conditions are strict and mechanical. The base rule is inclusion in the Hang Seng Composite indices: LargeCap, MidCap, and SmallCap where market capitalisation is ≥HK$5 billion. For A+H dual-listed companies, the H shares are eligible if the corresponding A share is within the Northbound perimeter. WVR companies and pre-revenue biotech under Chapter 18A do not enter the perimeter automatically: they must satisfy additional criteria set out in the SSE and SZSE rules rather than by the Hong Kong side (HKEX, Stock Connect Information Book for Investors, PDF). Exit works the same way: if average monthly market capitalisation over the 12 months before an index review falls below HK$4 billion, the stock moves to sell-only status. Reviews are tied to the semi-annual HSSI review, so a new IPO does not enter the perimeter immediately — the actual lag runs from a few weeks to six months (HKEX Stock Connect FAQ, version of 6 July 2026, PDF).

Volumes: Southbound ADT in the first half of 2026 was HK$123.1 billion against HK$111.0 billion a year earlier (+11%), Northbound RMB 345.3 billion against RMB 171.3 billion (+102%) (HKEX Interim Results, 19 August 2026, PDF); net Southbound inflows in the first quarter of 2026 exceeded HK$220 billion (HKEX, Q1 2026 Hong Kong Market Update). How the mutual access channels look from the mainland side is described in the market access section of the China material.

The A+H wave is the market's defining story. 2025 saw 19 A+H deals on KPMG's full-year recount, delivering half of all funds raised (HK$136.5 billion); the first half of 2026 alone produced 24 deals — more than the whole of 2025 — and 58% of funds raised, with eight of the ten largest deals of the half-year being A+H. As of June 2026 the pipeline held 116 A+H applications out of 443 filed publicly. The causes stack up: the CSRC's registration (filing) regime replacing the approval regime from 31 March 2023; the narrowing of the historical H-share discount to A shares through the 2025–2026 rally; mainland champions needing freely convertible capital for overseas capex and M&A; a base commission of 0.9% against 2.2%; the HKEX/SFC accelerated track for A-share issuers with market capitalisation above HK$4 billion; and, finally, the A+H float rule — 10% of the H class or HK$3 billion — which allows a listing with almost no dilution of control. The mainland leg of this route, the SSE and the STAR Market, is covered separately in the material on listing in Shanghai.

The scale of the venue as it stands: Hong Kong finished 2025 first in the world by IPO proceeds for the first time since 2019 — 100–119 listings and HK$272–286 billion depending on the counting methodology; as at the end of July 2026 there were 2,763 listed companies with a market capitalisation of around HK$46.8 trillion, and HK$328.2 billion raised in January–July (+154% year on year) (China Daily HK, 7 August 2026). In the first half of 2026, however, HKEX ranked second globally behind Nasdaq: a single SpaceX IPO of US$85.7 billion (per the Nasdaq newsroom) accounted for 45% of global half-year volume. That is a statistical artefact rather than lost momentum, but anyone deciding on the strength of a "number one in the world" argument needs to understand the difference.

Who Hong Kong suits, and who it does not

Choosing a venue means weighing three things: where the buyer of your stock sits, what entry costs, and which disclosure obligations you are willing to carry for life.

Hong Kong beats Nasdaq where the business, its revenue and its counterparties sit in Greater China and Asia-Pacific; where an A-share listing exists or is planned (in which case A+H is practically the only option); where the company qualifies under Chapter 18A or 18C and values the TECH channel; where Southbound access is critical; where there is sensitivity to US disclosure, PCAOB inspections and HFCAA risk; and where the deal exceeds HK$3 billion, since commissions are a multiple below the American 5–7%.

Nasdaq or NYSE win where you need depth of long-term capital for a loss-making technology story with no China angle, high secondary market liquidity, and the ability to pay for US M&A in your own stock. The LSE has been effectively out of contention for an Asian issuer since 2023–2025 on liquidity and valuation — the Shein case is instructive, having passed over New York and London to list in Hong Kong. SGX has lower entry thresholds and suits REIT and business trust structures and family offices better, but it is radically weaker on liquidity and offers no channel into mainland capital; the corporate and tax comparison of the two jurisdictions sits in Hong Kong versus Singapore.

Hong Kong is a poor fit for deals below HK$500 million, where costs consume the proceeds; for companies with no Asian history and no coherent China narrative, which risk becoming illiquid orphans in the order book; and for anyone who needs immediate Southbound inclusion, since that arrives no earlier than the next index review.

The Russian angle: the route is closed by compliance, not by rules

The constraint on companies with a Russian connection is not regulatory but a compliance matter: HKEX does not formally prohibit such listings, but no private participant in the chain will accept one.

Formally, Russia was recognised by HKEX as an acceptable jurisdiction of incorporation in January 2016, with a published Country Guide; a Russian-incorporated company may list only depositary receipts on the Main Board, whereas a Hong Kong, Bermuda or Cayman holding company with Russian subsidiaries may list shares directly (Charltons, 15 January 2016).

In practice the route has been closed at gatekeeper level since 2022. A sponsor holding an SFC Type 6 licence is always an arm of a global or major Chinese bank with dollar clearing, and it will not sign off on such a deal; without a sponsor an application is impossible in principle. Big Four auditors will not issue the accountants' report. Receiving bank, share registrar and settlement infrastructure all run into the OFAC, EU and UK perimeters. In April 2026 the EU sanctioned 60 third-country entities, Hong Kong companies among them, for facilitating Russian access to technology, which sharpened local providers' caution even about indirect links. Market precedent: Rusal (ticker 486) remains listed, but it is incorporated in Jersey and listed back in 2010, and in 2018 it lost roughly 40% of its market capitalisation in a day on OFAC sanctions and was dropped from the indices.

Tax for the founder

The tax side of a Hong Kong listing is defined by the territorial basis of profits tax, the absence of capital gains tax and a low stamp duty on share transfers.

The tax treatment is cross-checked against PwC, Hong Kong Tax Facts and Figures 2026/27 (PDF).

The issue of new shares in an IPO carries no stamp duty — the charge arises on transfers of existing securities. Watch the definition of Hong Kong stock: it turns on whether the transfer must be registered in Hong Kong, so a Cayman holding company with a Hong Kong register falls within the charge.

The main trap is the FSIE (Foreign-Sourced Income Exemption) regime, in force since 1 January 2023 and extended to disposal gains from 1 January 2024. It deems certain foreign income, including gains on the disposal of assets, to be Hong Kong sourced and chargeable to profits tax where the recipient is part of a multinational group and fails the economic substance, nexus or participation exemption tests. In other words, the proposition that "Hong Kong has no capital gains tax" holds for a resident individual selling shares but is not automatically true for a holding company inside a group. The exit structure has to be designed around FSIE before the transaction, not after; the adjacent questions are Hong Kong tax residence, economic substance and the general logic of capital gains tax.

Dim sum: the mechanics of an issue

A dim sum bond is a bond denominated in offshore renminbi (CNH), issued and settled outside mainland China and predominantly in Hong Kong; legally it is an ordinary eurobond, just in a different currency.

The governing law is usually English or New York, the documentation follows ICMA standards, settlement runs through the CMU or through Euroclear and Clearstream, and no filing with any Chinese authority is required. The issue itself is unregulated — the market is offshore, and regulation enters at exactly two points: listing and settlement.

There is no restriction on the issuer's nationality. The actual roster in 2025–2026: Chinese technology corporates (the main driver), Chinese state banks and their overseas branches (mostly certificates of deposit), provincial and municipal people's governments (Shenzhen, Hainan, Guangdong), China's Ministry of Finance with benchmark CNH CGB issues that build the curve, the PBOC with bills as an offshore liquidity management tool, the multilateral development banks AIIB and ADB, sovereigns and quasi-sovereigns (Indonesia with a 2025 debut, the Development Bank of Kazakhstan, KazMunayGas) and international corporates — Nestlé, Chubb, Temasek.

Listing is formally optional, but it is needed for tax purposes and for the mandate restrictions many institutional investors operate under. The route is Chapter 37 of the Main Board Rules, the professional-investors-only regime.

The debt-listing requirements are cross-checked against HKEX, Summary of Chapter 37 Listing Requirements (PDF), and HKEX Main Board Fees Rules (PDF). Post-listing obligations are minimal: respond to exchange enquiries, publish material announcements, and file annual and interim accounts.

The economics of an issue. Exchange fees are negligible; the real costs are underwriting and management fees of 15–35 basis points of size for investment grade, issuer's and dealers' counsel at US$150,000–400,000 on a standalone issue, ratings from S&P, Moody's or Fitch, and trustee, fiscal agent and listing agent fees. Realistically: a standalone issue takes three to six weeks from mandate to settlement at an all-in cost of 25–50 basis points; a drawdown under an existing MTN programme takes three to seven business days at 10–20 basis points (market estimate based on observed transactions). That is an order of magnitude faster and cheaper than an IPO, and faster than a panda bond. The HKMA also subsidises part of the cost of issuing: the Digital Bond Grant Scheme for qualifying tokenised issues and the Green and Sustainable Finance Grant Scheme for green and transition paper, with the grant size set by the schemes' current terms (Hong Kong Budget 2026–27, February 2026, PDF).

Settlement runs through the CMU. The Central Moneymarkets Unit has been operated since early 2025 by CMU OmniClear Limited, an HKMA subsidiary. At the end of 2025 assets under custody stood at HK$5.2 trillion (a record, +9%), of which debt securities were HK$4.1 trillion (+19%); new issuance for the year was HK$1.7 trillion (+10%, an all-time high); and the CMU's share of global new dim sum issuance is 95% (HKMA Quarterly Bulletin 01/2026, PDF). In December 2025 HKEX became a strategic shareholder in CMU OmniClear Holdings, aiming at unified post-trade infrastructure across equities and debt; membership has been widened to foreign regulated institutions, sovereigns and supranationals, and in February 2026 the Central Bank of the UAE joined the CMU. Holding and recording securities on the investor side is covered in the material on securities custody.

The investor base: Hong Kong and Asian banks with CNH treasury books, asset managers with CNH mandates, Asian insurers, private banking and family offices in Hong Kong and Singapore (a visible segment for three to five-year paper from known names), mainland institutions via Southbound Bond Connect, and central banks in Ministry of Finance and MDB paper.

The dim sum market in 2025–2026: volumes, rates, liquidity

The capacity of the dim sum market is bounded by the offshore renminbi pool — the mass of CNH sitting outside China's capital controls — which is why pricing is built off CNH HIBOR rather than onshore rates.

The numbers demand care, because headline statistics mix different things together. On HKMA primary data, new issuance of renminbi debt securities in Hong Kong in 2025 was RMB 1,100.9 billion (US$157.2 billion), a rise of only 2.7% year on year; excluding certificates of deposit it was US$112.4 billion, the corporate segment (that is, excluding sovereign and sub-sovereign issuers) was US$90.5 billion, and the corporate segment excluding CDs was US$45.6 billion. Meanwhile the amount outstanding at the end of 2025 rose 27.6% to RMB 1,607.3 billion (US$229.6 billion) (HKMA Quarterly Bulletin 01/2026, "The Hong Kong Bond Market in 2025", PDF). Those are two different stories: the stock is growing on longer tenors, not on annual flow. The widely quoted "+180% year on year in March 2026" and "around RMB 400 billion in the first quarter of 2026, +14%" both include a single PBOC transaction of RMB 60 billion; strip it out and March growth is +72% (Caixin, 2 April 2026). An issuer sizing capacity for its own deal should look at the corporate segment excluding CDs.

The rate detail. TMT issuance rose 82% in 2025 to RMB 31 billion: Alibaba RMB 17 billion (November 2024), Baidu RMB 14.4 billion in two tranches (2025), Tencent RMB 9 billion (September 2025) including RMB 1 billion at 30 years subscribed 8.6x, Meituan RMB 7.1 billion (November 2025), Kuaishou RMB 3.5 billion (January 2026). The pricing benchmark to anchor on: JD.com placed RMB 10 billion in April 2026 — RMB 7.5 billion at five years with a 2.05% coupon and RMB 2.5 billion at ten years at 2.75%, rated A− by S&P (Caixin, 21 April 2026). International issuers grew 28.5% in 2025: Nestlé RMB 2 billion (May 2025), Temasek RMB 5.5 billion (July 2025), Chubb RMB 4.5 billion (August 2025), the Development Bank of Kazakhstan RMB 2 billion (September 2025), KazMunayGas RMB 1.25 billion (October 2025).

Pricing: the base is CNH HIBOR (the TMA fixing in Hong Kong), not onshore SHIBOR. Average dim sum yields in May 2025 were 2.29% against 1.64% onshore at an average duration of 3.6 years against 5 years — meaning offshore pays a premium of around 65 basis points for lower liquidity and shorter tenor. Primary market spreads have compressed to 10–20 basis points, a sign of excess demand. The frequently quoted "average renminbi funding cost of around 0.7%" describes the short, high-grade end — the level of CNH HIBOR and certificates of deposit — rather than the whole curve: yields on outstanding dim sum paper run roughly three times higher. Against roughly 2.5% in alternative currencies the saving is still large, and issuers raising CNH and swapping into dollars have saved up to 40 basis points against issuing in dollars directly. There is a structural refinancing driver behind the demand: a third of Chinese corporates' dollar bonds, out of some US$750 billion outstanding, matures within the next two years.

The regulator's tool for managing that risk is the HKMA's renminbi facilities, and here almost all available commentary is out of date. The chronology: February 2025, the RMB Trade Financing Liquidity Facility launched; October 2025, it is replaced and widened into the RMB Business Facility (RBF) at RMB 100 billion, with eligible uses extended from trade finance to capex and term working capital loans; 26 January 2026, a doubling to RMB 200 billion from 2 February 2026 is announced after the previous quota was fully allocated across 40 participating banks; announced on 7 July 2026 and effective 10 July 2026, the size is raised to RMB 500 billion, tenors extended to 9 months, 2 years and 3 years, and by July more than 90% of the quota had been allocated (HKMA, 7 July 2026, PDF; Government Information Services, 26 January 2026). The model is hub-and-spoke with Hong Kong as the hub, funded off the swap line with the PBOC. Also under study: a tender mechanism for seven-day offshore renminbi liquidity in repo form, and HKMA issuance of its own short-dated offshore renminbi instruments to build out the curve. How this liquidity translates into the price of credit for a borrower is covered in the material on the cost of money in China.

Southbound Bond Connect turns dim sum from a purely offshore instrument into a channel to mainland liquidity. It launched in 2021; in July 2025 the eligible investor base was widened from banks to brokers, fund managers, insurers and wealth management companies outside the QDII quotas, after which issuance roughly doubled to around RMB 102 billion over August–September 2025; in June 2026 the largest Chinese insurers began buying dim sum through the widened channel; and on 7 July 2026 the annual quota was raised from RMB 500 billion to RMB 800 billion, the perimeter extended to Hong Kong dollar bonds and derivatives of renminbi paper as well as to the Macau market, with repo against Southbound holdings introduced (The Standard, 7 July 2026). The widening of the investor base in 2025 and of the quota in 2026 is precisely why spreads compressed to 10–20 basis points.

Dim sum versus panda bonds

Dim sum and panda are two ways of borrowing in renminbi: the first offshore in CNH with no Chinese regulation, the second onshore in CNY with mandatory registration and capital controls.

• Speed. Dim sum by a wide margin: three to six weeks for a standalone issue and three to seven business days for a drawdown, against months of onshore registration procedures.
• Use of proceeds. Dim sum gives full freedom, like any eurobond; panda proceeds are constrained by onshore regulation, and the mechanics of those constraints are set out in the material on panda bonds.
• Headline rate. Panda is cheaper: typically around 100 basis points over the CGB curve. Dim sum pays a premium of roughly 65 basis points on average yield.
• All-in cost. Dim sum is generally cheaper: no local rating and no local guarantee wrapper, lower legal costs, and an MTN programme that can be reused.
• Depth of demand. For short, high-quality paper of a few billion renminbi the two are comparable. For long tenors, sub-investment grade, and any issuer that needs the money outside the PRC, dim sum has no competition, and the gap widens as Southbound expands.
• When panda makes sense. When the money is genuinely needed inside the PRC — capex for a Chinese subsidiary, local refinancing — or when the political and reputational effect of entering the onshore market is itself the objective.

Here panda is treated only as an alternative to dim sum for the same issuer: the onshore route in full — registration, reporting, repatriation and capital controls — is the subject of a separate material.

Equity or debt: the fork by funding purpose

The choice between an IPO and a bond is decided not by the cost of capital but by what you are buying with the money: permanent capital and public status, or term money with no dilution.

The practical sequence most Asian issuers follow is the reverse of the familiar one: dim sum first, as a way to build a credit history, public reporting and a rating, and only then an IPO — so that by the time the A1 is filed the company already has tested accounts, a working IR function and an external rating, which is a meaningful share of the preparation budget. The wider map of Hong Kong infrastructure, from a bank account to a family office under the FIHV regime, is set out in the jurisdiction overview.