Unprofitable AI firms face public market test under Star Market’s 5th rule

China’s technology sector is approaching a critical inflection point as regulators and early investors prepare for a wave of initial public offerings from artificial intelligence and robotics firms. The Shanghai Stock Exchange has recently provided guidance under the Star Market’s fifth listing standard, explicitly opening a path for unprofitable AI large-model companies that possess strategic technologies but have not yet achieved meaningful revenue or profits. In parallel, Unitree Robotics, a developer of humanoid robots, has advanced further through the Star Market process, offering a concrete test case for how the public market will price a high-growth robotics company against the backdrop of China’s humanoid robot boom. Together, these two examples reveal a larger structural shift: public markets are about to impose discipline on private market narratives that have been fueled by capability-based fundraising rather than financial fundamentals. The Star Market’s fifth standard, designed for ‘strategic technology’ firms, removes the traditional profitability hurdle but demands rigorous disclosure and investor scrutiny. For AI large-model companies, this means that while they may access public capital earlier than conventional tech firms, they must also confront valuation models that rely on revenue multiples, growth trajectories, and addressable market estimates rather than product demos or policy alignment. Unitree’s case is more conventional in some respects—it has reported rapid revenue growth and even profits—yet its more recent filings indicate margin compression, signaling that even established robotics players may struggle to sustain the valuation premiums once commanded in private rounds. The implications for late-stage investors are profound. Private markets in China’s AI and robotics space have relied on proxy metrics: chip access, founder credentials, government policy relevance, and proof-of-concept performance. These signals are difficult to translate into public market multiples. Once IPOs establish public comparables, they will ‘travel backwards’ into the private domain, forcing boards and valuation committees to justify last-round marks. Funds that invested at valuations predicated on a perpetual narrative of technological superiority may discover that the IPO window, far from lifting their stakes, narrows the gap between storytelling and evidence. This disciplining effect could trigger down rounds or forced markdowns in the pre-IPO secondary market. From a sector-wide perspective, the co-occurrence of these two tests for distinct sub-sectors—general AI platforms and embodied AI (robotics)—is not coincidental. It reflects a deliberate regulatory push to channel capital toward strategic technology areas while simultaneously imposing market-based accountability. The Star Market’s architecture, with its tiered listing standards, is designed to balance national industrial policy goals with investor protection. The fifth standard, in particular, was created to attract innovation-heavy firms that might otherwise list overseas; but it also exposes them to the full glare of quarterly reporting, analyst coverage, and institutional investor demands. This environment stands in stark contrast to the earlier funding environment, where rounds were often closed on the basis of a single technical milestone or a well-connected lead investor. Unitree’s trajectory also highlights the challenges of scaling humanoid robotics. While the global market for humanoid robots is projected to grow substantially, the path to mass adoption remains uncertain, plagued by high unit costs, safety regulations, and limited commercial use cases beyond niche industrial settings. Unitree’s early financial disclosures suggest it has captured some demand, but the reported margin pressures raise questions about whether unit economics can scale favorably. This mirrors broader debates in the AI sector, where inference costs, chip shortages, and intense competition can erode the profitability of even the most advanced large-language models. For the broader financial ecosystem, these IPO test cases will likely set a precedent for the valuation frameworks applied to China’s entire next-generation tech cohort. Investment banks, auditors, and regulatory bodies are watching closely. A successful pricing and aftermarket performance could catalyze a wave of public listings, unlocking liquidity for venture capital and providing new benchmarks for private funding rounds. Conversely, a disappointing reception—characterized by sharp first-day declines or sustained trading below the offer price—would not only chill the immediate IPO pipeline but also reverberate through the venture capital and growth equity markets, potentially triggering a broader repricing of technology assets across Asia. Looking ahead, the months leading up to the first Star Market listing of a large-model company will be characterized by intense valuation debates. Investors must balance the promise of strategic technology against the hard realities of public market comparables. Boards will be forced to reconcile the lofty visions presented in pitch decks with the granular metrics required by prospectuses. In this environment, the ‘rude surprise’ for China’s tech firms may not be that they cannot go public, but that the public market’s verdict on their worth is far more sober than the numbers assigned in the last private funding round.