Algorithmic pricing has become a focal point of antitrust enforcement worldwide. The 货拉拉 (Huolala) rectification marks Beijing’s first major public move to bring algorithmic pricing within the scope of antitrust enforcement. More importantly, however, it signals a broader shift in China’s antitrust regulator approach: using the Anti-Monopoly Law not only to scrutinise algorithms, but also to flag its growing focus on platforms’ exercise of buyer-side market power. On 18 June 2026, China’s State Administration for Market Regulation (SAMR) announced a sweeping antitrust rectification of Huolala, the country’s largest intra-city freight-matching platform, concluding a nine-month supervisory process that began with a regulatory interview in September 2025.11 The rectification extends well beyond algorithmic pricing, requiring Huolala to overhaul its pricing algorithms, dismantle platform rules with exclusionary effects, reduce commission rates, strengthen protections for drivers, and establish a more robust antitrust compliance framework. Huolala responded to SAMR’s requirements on the same day with a commitment to comply.22 SAMR’s rectification measures (red) and Huolala’s corresponding responses (blue) What insight can be drawn from SAMR’s Huolala case? Takeaway 1: Algorithmic pricing carries a genuine antitrust risk Although algorithmic pricing remains a relatively new area of enforcement, regulators have made clear it is not exempt from antitrust scrutiny. The OECD’s recent report on algorithmic pricing and competition, surveying work across G7 antitrust authorities, signals that this concern is converging globally rather than confined to any single jurisdiction.33 The report organises these concerns around two theories of harm: collusion and unilateral conduct. On collusion, the underlying worry is that algorithms can produce co-ordinated pricing outcomes through channels well beyond a traditional cartel agreement. These can take varying forms: algorithms may simply be used to execute or monitor a conventional price-fixing arrangement; they may create hub-and-spoke dynamics or vertical restraints; or they may give rise to tacit co-ordination through autonomous learning. On unilateral conduct, a dominant firm’s use of pricing algorithms can itself amount to an abuse of market power. Pre-calibrated, non-transparent pricing mechanisms may enable predatory pricing, rebate structures or price discrimination designed to lock in users, deter switching and ultimately harm consumers. The Huolala case shows this second theory playing out in practice – SAMR’s objection was that Huolala used its substantial market position to run an opaque, multi-factor pricing algorithm that suppressed rates paid to drivers. The remedy followed the theory: disclosure of pricing rules, a reduced share of multi-factor pricing and a return to transparent per-order and per-kilometre rates. It is a concrete illustration that China, like European and other regulators pursuing unilateral-conduct theories, now treats an opaque pricing algorithm as sufficient grounds for antitrust enforcement. Businesses should note that using AI or algorithmic pricing tools does not reduce their responsibility under competition law. Companies are expected to understand how their pricing systems work – including the data inputs, the logic driving pricing outcomes and whether similar tools are used by competitors. This is particularly important where AI capabilities are added incrementally to existing pricing systems without a reassessment of antitrust risks. In short, algorithms used for pricing, personalisation, market co-ordination or influencing consumer choice remain fully subject to existing competition laws, regardless of whether they are rule-based or self-learning. Takeaway 2: SAMR’s focus signals growing scrutiny of platforms’ buyer-side market power The Huolala case sits within a broader campaign. In 2026, SAMR has intensified its “反内卷” (anti-involution) push across the platform economy, targeting the low-price, race-to-the-bottom dynamics that algorithmic pricing can generate for those on the supply side. What makes Huolala notable is who the rectification protects–not consumers but drivers, the supply side captured by the platform. Viewed alongside SAMR’s parallel actions this year against food-delivery and travel platforms, the case points to a clear enforcement priority: reining-in large platforms’ buyer-side power over the merchants and workers who depend on them, not just their conduct toward end users. This mirrors a broader global shift toward treating buyer-side market power as a genuine antitrust concern. US antitrust authorities have increasingly incorporated labour market considerations into enforcement. The Department of Justice successfully challenged the proposed merger between Penguin Random House and Simon & Schuster, arguing that it would substantially lessen competition in the market for acquiring publishing rights from authors and reduce authors’ compensation.44 Also, the 2023 DOJ/FTC Merger Guidelines expressly recognise that mergers may harm competition in labour markets.55 In the EU, the European Commission’s (EC) new draft Merger Guidelines issued this May identify that reduced labour-market competition may constitute a theory of harm,66 and its recent Delivery Hero/Glovo decision demonstrates an increasing willingness to enforce against labour-market collusion,77 a priority increasingly echoed by national competition authorities such as those in Portugal, Italy and France. In Hong Kong, the Competition Commission (HKCC) reached a resolution with Keeta in late 2025 after finding that contractual restrictions preventing restaurants from joining or promoting themselves on competing platforms risked locking in merchants and preventing multi-homing. The common thread across jurisdictions is that a platform’s market power does not need to harm consumers directly to attract scrutiny–exploiting its supply-side is now independently actionable. For business, the practical implication is to assess market power symmetrically: antitrust risk reviews should cover not only pricing and terms offered to end users, but also the leverage exercised over the counterparties–drivers, merchants and riders–who depend on the platform for their livelihood, as this becomes an active enforcement priority both in mainland China and among regulators worldwide. Conclusion The Huolala case is best read not as an isolated platform-economy fine, but as a signal on two fronts converging worldwide: algorithmic pricing is now squarely within antitrust regulators’ reach, and buyer-side market power (monopsony) is increasingly as actionable as harm to consumers. Businesses should take this opportunity to review both areas proactively, rather than waiting until a similar regulatory intervention reaches their own doorstep.

Imagine spending years building a virtual asset fund management business, only to watch your departing Portfolio Manager walk through the door of a direct competitor and then discover that the non-compete clause you relied on is unenforceable. That is precisely what happened to the employer (the “Company”) in the recent case of Pando Finance Limited v Ng Ean Kiam [2026] HKCFI 1046. The Company sought an interlocutory injunction to stop its former Portfolio Manager (the “Employee”) from joining a rival firm (the “Competitor”) and allegedly using the Company’s confidential information, including its Securities and Futures Commission (SFC) filings and exchange-traded fund (ETF) product applications, to compete against it. But the court refused. Post-termination restrictive covenants (PTRs) are a critical line of defence for employers wishing to protect trade secrets, client relationships and competitive advantage after an employee departs. But in Hong Kong, PTRs are presumptively void. Courts will only uphold them if the employer can prove that the restriction is no wider than reasonably necessary to protect a legitimate business interest and is reasonable in all the circumstances. The bar is high and this case shows just how easy it is to fall short of it. The facts of this case serve as a cautionary tale: a PTR that is too wide in scope, too long in duration and unrestricted in geography is not just imperfect, it is worthless. Employers who invest in attracting and developing senior talent should ensure their contractual protections are equal to that investment. What happened The Employee was a Singaporean national and a seasoned fund manager with over 30 years of experience in financial services. He held Responsible Officer status under the SFC for regulated activities Types 1, 4 and 9. He joined the Company as Portfolio Manager in February 2024 and was confirmed as permanent staff in May, at which point he signed a confirmation letter containing a 12-month non-compete covenant. He resigned on 9 January 2025 and left on 8 April after serving his three-month notice period. Within weeks, he joined the Competitor as Managing Director and Senior Portfolio Manager. His SFC licence was transferred to the Competitor on 29 April, a publicly available update on the SFC’s website. The Company alleges it only discovered the move in approximately July 2025, when the Competitor launched virtual asset ETF products that appeared similar to those of the Company. The Company issued a cease-and-desist letter on 5 August and commenced proceedings on 22 August, some four months after the Employee had already joined the Competitor. In an earlier skirmish, the court had already refused to grant an interim-interim injunction, finding that the Company had failed to prove urgency and that the balance of prejudice did not favour restraint. The Company’s application for an interlocutory injunction suffered the same fate. Because the 12-month restriction period would expire in April 2026, before any trial could realistically take place, granting the injunction would have amounted to granting final relief. The Company therefore faced a heightened threshold. It needed to show that it had good prospects, or better prospects of success than the Employee. Simply showing a “serious issue to be tried” would not be enough. The Company could not clear that bar. Drafting failures that doomed the clause The court identified three fundamental flaws in the non-compete clause, each of which independently undermined its enforceability. Employers should treat each as a checklist item when auditing their own PTRs: No geographical limitation—worldwide reach, zero enforceability. The clause contained no geographical restriction and purported to bind the Employee globally. The Company tried to justify a worldwide restraint on the basis that virtual asset ETF products launched overseas would be “easily accessible” to Hong Kong investors. The court rejected this reasoning: the stated purpose of the clause was to protect confidential information relating to the Company’s SFC applications—information that had little relevance to ETF products launched in other jurisdictions under the oversight of different regulators. The court also noted that a worldwide restraint was particularly difficult to justify given that the Employee was a Singaporean national with prior employment ties to Singapore. Absent a compelling justification, courts will not hesitate to strike down a clause of this kind. No evidence to justify 12 months—duration must be grounded in fact. The Company adduced no evidence that the confidential information it sought to protect—primarily its SFC applications and related filings—had a commercial shelf life warranting a 12-month restriction. There was no evidence of operational strategies or special intelligence that would continue to be sensitive for that period. The lesson: duration must be calibrated to the actual lifespan of the information or interest being protected, not chosen as a round number. Scope too wide—captured the entire industry, not just competitors. The clause extended to “any other business or entity … which directly or indirectly competes with the business of the Group.” Because the Company’s business was not expressly limited to virtual asset management, the court found the clause would have prevented the Employee from joining even traditional asset management firms with no connection to virtual assets whatsoever. The Company argued the clause should be read narrowly to cover only virtual asset competitors, but the court found that construction was not supported by the language of the clause or its surrounding context. If you intend to limit the clause to specific sectors or activities, say so expressly. The court also found that the Company had failed to demonstrate good prospects of establishing a breach. Despite alleging that the Employee had divulged confidential information, including SFC filings and email communications with regulators, the Company could not identify the specific information it sought to protect with sufficient particularity. Critically, certain of the Company’s ETF-related documents (such as prospectuses) were publicly available on the SFC’s website, meaning they could not be characterised as confidential. The court also noted that the Competitor’s draft prospectuses had been prepared before the Employee joined, which undermined any inference that he had used the Company’s confidential information to assist the Competitor. The broader principle is that an employer cannot restrain a former employee from deploying skill and knowledge acquired during employment for a new employer’s benefit. Employers must be specific and accurate about what information they are truly protecting. On the balance of convenience, the court found in favour of the Employee on every count. The Company failed to demonstrate any real risk of irreparable damage, particularly given that the Employee had already provided confidentiality undertakings (which the Company acknowledged were enforceable as court orders). By contrast, the Employee faced serious prejudice from real risk of losing his employment, lasting reputational stigma from an abrupt termination and impeded career development. This was made more acute by the fact that he had been unemployed for around a year before joining the Company, demonstrating difficulties he faced in the job market. The delay compounded the Company’s difficulties. The Employee’s SFC licence transfer to the Competitor had been publicly available on the SFC’s website since 29 April 2025. The Company’s own CEO had been informed in February 2025 that the Employee intended to take up new employment in “mid-April 2025”. News of the move was also known among the Company’s own staff shortly after the SFC update appeared. Even accepting the Company’s case that it only learned of the new employment in approximately July 2025, there was still an unexplained gap of around four to six weeks before proceedings were commenced. As the court noted, in a case where granting interim relief would amount to granting final relief, delay is a weighty factor, and one that employers ignore at their peril. The Company also did not offer the Employee alternative employment during the restriction period, which further tipped the balance against it. The court dismissed the Company’s application and ordered costs against it. What employers should do now This case is a practical reminder that non-compete clauses are not self-enforcing shields. Here is what employers should review immediately. Add a geographical restriction—and justify it. A worldwide restraint will rarely be enforceable unless the employer can specifically demonstrate why global coverage is necessary to protect the identified interest. Match the geography of the restriction to the geography of the business the employee is actually involved in. Tailor every clause to the employee and interest you are protecting. A one-size-fits-all PTR is a false economy. The scope of the restriction must be proportionate to the employee’s actual role and the specific confidential information or client relationships at stake. Conduct a clause-by-clause review and consider whether your existing PTRs remain fit for purpose for key employees whose roles have evolved since their contracts were signed. Act fast—and monitor public registers. Delay can be fatal to an injunction application. Employers should have a process for monitoring the SFC register and other publicly available licensing databases or platforms so they are not caught off-guard when a former employee moves to a competitor. If you suspect a breach, take legal advice and move immediately. Weeks of unexplained inaction could cost you the right to relief entirely. Review and update PTRs at key career moments. PTRs are interpreted as at the date the contract was entered into—a clause drafted for a junior hire may offer far less protection once that person has been promoted to a senior role with access to sensitive strategy, client relationships or regulatory know-how. Build a practice of revisiting PTRs on promotion, role change or on the expiry of a significant period of employment. The judgment is available here. For a more detailed discussion on the relevant legal principles on PTRs, please refer to our earlier legal update at: https://www.jsm.com/publications/2024/hong-kong-restrictive-covenants/