Recently, AgiBot was reported to have launched a new round of financing ranging from CNY 1 billion to CNY 1.5 billion, with a post-money valuation of approximately CNY 50 billion, making a Hong Kong stock listing unachievable within this year. On the same day, the S-4 filing of Agility Robotics took effect, with a shareholder vote scheduled for November 2. While one is slowing down its process, the other is accelerating.
01. Two Issues Holding Back AgiBot
Looking at the financials of the two companies, the differences are quite specific: AgiBot's revenue in 2025 was CNY 1.05 billion, while Agility's was USD 1.782 million. Scale is not the issue for AgiBot—the 10,000th unit rolled off the production line in March, and the 20,000th unit was delivered to Chimelong in September. What is holding it back are two issues of a completely different nature.
The first is procedural. According to the rules of the Hong Kong Stock Exchange, if new financing is completed after the application proof is submitted but before the listing, the company must wait for 120 days after the final tranche is received before it can be listed; this rule is automatically triggered. When the initiation of the listing process was announced in July, the prevailing expectation was that the listing could happen as early as August. The new financing has pushed this milestone to 2027.
The other is attitude. According to relevant media reports, regulators are informally guiding through window guidance to suspend the listing of some humanoid robot companies, requiring them to prove sustainable revenue capabilities or a path to narrowing losses. AgiBot and X Square Robot were specifically named, with the focus pointing to the commercial substance of the orders from the data collection centers.
In September 2025, AgiBot pre-won a humanoid robot procurement project in Hubei Optics Valley with a bid of CNY 31.0161 million. The tenderer was Hubei Optics Valley Dongzhi, in which AgiBot holds a 15% stake. Such orders have been publicly disclosed to accumulate over CNY 40 million; in January 2026, the training center, in turn, sold thousands of hours of training data to AgiBot.
In February this year, AgiBot spun off its data business into Mifeng Technology to promote independent financing, but it remains the controlling shareholder and has not deconsolidated it, so the audit pressure has not been lifted.
02. Agility's Ledger Shows Only USD 1.782 Million
The S-4 filing marks Agility's first public disclosure of its financials: net sales of USD 1.782 million in 2025, of which 64% came from related parties, a gross margin of approximately -151%, and a net loss of USD 138 million. The auditor's opinion carries more weight than the numbers: without considering this transaction, the company has determined that there is substantial doubt about its ability to continue as a going concern. If the SPAC deal falls through, the cash on hand will not last a year.
Another set of figures is even more thought-provoking. Digit 5 has over USD 300 million in multi-year committed orders, but according to the document disclosure, this money comes from a single customer for 1,000 v5 units under a three-year Robot-as-a-Service agreement, subject to delivery and product milestone conditions—it is neither recognized revenue nor an unconditional performance contract.
Therefore, what is highly concentrated is not the customer list, but the source of orders. Four customers have been disclosed: GXO, Schaeffler, Toyota Motor Manufacturing Canada, and Mercado Libre, along with nine committed deployment sites, accumulating over 65,000 hours of real-world operation.
Its strengths must also be clarified: RoboFab has a planned annual production capacity of 10,000 units and is a launch partner for NVIDIA Halos. This is the confidence it has to step into the public market with such a ledger.
03. Does Validation Count as Revenue?
The biggest difference between the two humanoid robot manufacturers is not the valuation multiple, but the accounting standards.
Agility treats a large number of real-world deployments as R&D activities and does not recognize current revenue. It has 65,000 hours of operation, USD 300 million in committed orders, and four industrial customers, yet the revenue column only shows USD 1.782 million. AgiBot does the opposite: equipment sales from data collection centers are recognized as revenue. The revenue looks good on paper, but auditors will question the commercial substance: did the money come from independent third parties, or did it just make a round trip through a joint venture in which it holds a 15% stake?
This creates a counterintuitive paradox: one company desperately tries to prove it can make money, only to be asked where the revenue comes from; the other voluntarily chooses not to recognize revenue, yet manages to secure USD 300 million in order commitments and a total fundraising amount of USD 620 million.
The so-called two sets of yardsticks differ right here—the timing of recognizing "validation" in the accounting books is different on both sides. In the A-share and Hong Kong stock systems, scenario validation must be immediately converted into revenue, preferably into profit on the spot; the U.S. system allows it to remain in the R&D account for the time being.
04. Different Yardsticks, Different Timelines
Unitree has just paid the tuition for the lesson on going public. It listed on August 19 with an issuance valuation of approximately CNY 61 billion and an issuance P/E ratio of 219.23 times, compared to the industry average static P/E ratio of 38.56 times; its market cap on the first day once reached CNY 444.9 billion, and is currently around CNY 170 billion, with JPMorgan Chase issuing a sell rating.
Yet, its revenue in 2025 was CNY 1.699 billion with a net profit excluding non-recurring items of CNY 591 million, and in the first half of 2026, revenue was CNY 1.152 billion, a growth of 48.54%. A company that is already profitable and growing at a decent pace still saw 60% of its market cap wiped out in two months. This highlights the most crucial point: the tightening in the A-share market is not directed at the income statement, but at the expectation gap.
The leniency in the U.S. stock market is not truly lenient either. The going concern doubt is a formal disclosure and can equally be used as ammunition for short sellers. Once orders fail to convert into revenue, the narrative of losses will backfire at an even faster pace. The SPAC merely delays this layer of risk.
05. What Matters Are Scenario Orders That Can Pass the Audit
Whoever secures large orders and enters real-world scenarios should be recognized by the capital market. The comparison between AgiBot and Agility perfectly corroborates this point. But there is a hidden premise here: orders must be convertible into revenue.
First, revenue must withstand exclusion. Frost & Sullivan reported AgiBot's revenue in the first half of 2026 at CNY 610 million with a 16.1% market share; this caliber explicitly excludes related-party transactions. Under this stricter caliber, it still ranks second globally, indicating that real demand is not weak; the issue lies in the structure, not the total volume.
Second, a single customer cannot be allowed to fuel all the imagination; Agility's USD 300 million order comes from just one customer.
Third, deployments must be convertible into revenue. If the 65,000 hours remain in the R&D account for a long time, they are merely PR materials.
AgiBot holds a decent report card; it just needs to re-explain its revenue structure. As for Agility, it exchanged USD 1.782 million in revenue for a ticket to NASDAQ. The ticket is real, and so is the price: post-closing, the market will scrutinize it using operational logic, and its report card is still missing the "revenue" column.
This article is compiled based on publicly available online information and is for reference only, not constituting investment advice.