Global Mofy AI sells 3D digital assets and virtual production services out of China, and the company just ran a second reverse split and a fresh equity raise within one week of each other. That pairing is a listing defense, not a growth announcement.
The mechanics are visible in the filings. The company bought back its own warrants by letting holders waive five million of exercise rights. Then a special shareholders meeting preapproved a further 16-for-1 consolidation to trigger if the share price breaks below the Nasdaq bid price within six months.
Fiscal 2025 revenue of 55.9 million masked a net loss that ran far above the operating line. Most of the gap was a warrant issuance charge, not an operating cost.
The balance sheet is now dominated by an amortizing asset base. The question for the next two quarters is whether revenue conversion outpaces that clock, or whether the stock simply keeps repricing at 0.538 per new share until the split math runs out.
Global Mofy AI operates two lines of business from Beijing and a handful of PRC subsidiaries. The virtual technology service line is fixed-price, milestone-billed 3D virtual content production for film, TV, advertising and games, and the digital asset development line licenses use rights from a bank of more than 150,000 high-precision 3D assets. The company cites Frost & Sullivan to call itself one of the leading digital asset banks in China, and names L'Oreal and Pepsi as repeat customers. The tilt shifted in fiscal 2025: the virtual service line contributed 59.2 percent of revenue, and the asset line the remainder. That is a reversal from a year earlier when assets were the larger share. The services business is growing faster, but the asset base is what the market is really trying to value.
The peer set is thin. The nearest listed comparators are Chinese VFX and virtual production houses, most of them private or Beijing-listed, plus a scatter of small-cap American digital content and asset licensing names, and the NVIDIA Omniverse ecosystem that the company has wired its Gausspeed platform into. Against that set, Global Mofy is small, at roughly 55.9 million of annual revenue, and it carries an asset base that looks more like an IP shop than a services shop. The strategic positioning is therefore a two-sided one: a services cash engine that funds a growing, self-amortizing asset library, with the AI narrative layered on top.
The ownership structure matters as much as the product. Chairman and CEO Haogang Yang holds 2,897,805 Class B shares through James Yang Mofy Limited, each carrying twenty votes, and the CFO and other executives hold Class B shares that together give officers and directors about 61.9 percent of total voting power on a post-split share count. The Class B structure is the reason the 16-for-1 preapproval passed with near-unanimous voting power in favor: the split is controlled by the very people who benefit from keeping the listing.
A related party note completes the picture. The principal shareholder has pledged financial support if cash needs outstrip resources, a soft backstop of the usual kind. The annual report discloses that the company relies on short-term bank loans of 3.4 million. It also leans on the goodwill of its largest customer, which accounted for 16 percent of fiscal 2025 revenue. The strategic context is a controlled micro-cap whose listing status is now the primary corporate objective.
The Mofy Lab platform is the named technology stack: 3D rebuilt technology that scans physical objects into movie-precision 4K digital twins, and AI interactive technology that lets customers assemble virtual scenes without deep technical skill. The asset bank sits on top, with 150,000 plus 3D assets of scenes, characters and props, licensed under copyright agreements that transfer production, adaptation and sublicense rights. The moat is scale of library plus customer familiarity, not any defensible algorithm, and the company says as much in its growth strategies, which emphasize adding assets and deepening existing relationships rather than patenting new capabilities.
Gausspeed is the AI layer. It is a generative AI solution built on NVIDIA Omniverse and RTX GPUs that shortens production time for virtual content, and the annual report positions it as the bridge between the asset bank and faster, cheaper content generation. The strategic intent is clear: turn a labor-heavy services business into a semi-automated pipeline where the library is the input and the AI is the multiplier. Whether that multiplier shows up in margins is the unresolved question, because fiscal 2025 gross margin fell while the AI spend went up.
The asset library is the real product in valuation terms. Intangible assets on the balance sheet jumped to 59.7 million in fiscal 2025, funded by heavy cash purchases. Depreciation and amortization climbed to 11.9 million over the same stretch, more than double the prior-year level. That is an asset base that eats roughly a fifth of revenue every year, and it means the library has to keep generating licensing and services revenue at a growing clip just to stand still. The amortization schedule is the quiet cost of the moat.
The honest reading is that the moat is thinner than the AI label suggests. Digital asset libraries are replicable, the customer base is concentrated, and the technology depends on NVIDIA's roadmap rather than its own. What the company does have that most small-cap AI content names lack is a three-year revenue history and a named enterprise customer list, and that combination is what the valuation section prices.
Fiscal 2025 revenue reached 55.9 million, and the service line drove that print on the back of the film and TV recovery. The service line grew 58.4 percent, but cost of revenue grew 62.6 percent, faster than revenue itself. Gross margin fell from 50.3 percent to 40.2 percent over the same period. That is a ten-point decline that tracks the outsourcing-heavy cost structure of the services business: as the project book re-filled, the company leaned on external production partners, and the partners kept most of the incremental dollar.
The net loss of 19.3 million is the number that needs decomposing. The operating story was already weak, with operating income down to 3.0 million on higher G&A and R&D, before the below-the-line items even landed. The swing below the line was dominated by a warrant issuance loss of 22.3 million, partly offset by a 5.0 million gain from extinguishing old warrants. Strip out the warrant charge and the year is a modest operating loss, not a 19.3 million collapse. The warrant math is the artifact, and it comes from the April 2025 financing. The company sold 2,030,460 Class A shares with matching warrants, and the exercise price reset to 24 percent of the then-market price within a week of signing. That reset forced the full liability through the income statement at issuance. Some holders simply waived their warrants, producing the extinguishment gain on the other side of the same capital structure.
Cash flow is the cleanest line in the statement. Operating cash flow was 22.1 million, the best in company history, but it is flattered by non-cash amortization plus stock-based compensation. Investing cash flow was negative 33.4 million, almost entirely the asset purchase, and ending cash fell to 4.2 million. The 6.7 million of customer advances on the balance sheet represents unbilled project deposits rather than earned revenue. The working capital position is thin for a company paying vendors in advance. Working capital was 1.5 million, and the accounts receivable balance had only 47.2 percent collected by the filing date. Advances to vendors, the prepayments that fund the outsourced production, grew to 10.2 million on the balance sheet, the single largest operating working capital drain in the year.
The capital structure tells the financing story better than any line item. The November 2024 reverse split compressed the share count, and four equity rounds followed it. The latest registered direct raised 2.0 million at 0.538 per share. Each round repriced the stock lower or at a discount to the prevailing price, and each round diluted the pre-existing holders. The shareholders meeting in September then ratified the 50-for-1 split. The split had taken effect in June 2026, and the ratification closed the loop on it. It also preapproved a further 16-for-1, the structural response to that repricing spiral. The split is the cheapest listing defense available, but it is also a confession that the share count, not the business, is the binding constraint on the stock.
The immediate execution risk is the compliance clock. The 16-for-1 consolidation is preapproved to fire on the tenth trading day after the share price fails to meet the Nasdaq bid price requirement during the six-month window that just opened. The stock has already traded through a 52-week low of 1.80. Before the split arithmetic, the same stock had traded near 114. A split does not change the business, but it does change the arithmetic of who can hold the stock, and a second split within a year signals that management sees the bid-price test as a live scenario, not a formality.
The second variable is the amortization race, and it is the one that decides the margin story. With an 11.9 million amortization run-rate, the library consumes about 1.0 million of pre-tax income per month. The asset line only grew 11.6 percent in fiscal 2025. The execution requirement is for asset licensing and service revenue to grow faster than the amortization schedule, and the fiscal 2025 print shows the opposite, with gross margin down ten points. A year of sub-10 percent asset growth against a 30 percent amortization run-rate is a structural margin headwind that management has not yet answered.
The third variable is customer concentration and collection, and it is the one that decides the cash story. The top customer at 16 percent of revenue and four customers holding 73 percent of accounts receivable mean that a single production delay or a single credit event shows up in both the income statement and the cash flow statement in the same quarter. The annual report discloses a 2.3 million increase in the provision for doubtful accounts, and the receivable collection rate of 47.2 percent by the filing date is the early warning.
The fourth variable is financing dependence. Cash of 4.2 million sits against a bank loan rollover of 3.4 million. A tax payable of 2.7 million completes the near-term obligation stack. A capital program that burned 33.2 million a year leaves the company structurally dependent on the public markets for working capital. The recent registered direct raise is the latest proof. The forward question is not whether management can execute the technology roadmap, but whether the compliance and amortization clocks allow enough runway for the revenue growth to arrive.
The bear case is a three-part failure. First, the bid-price test is breached inside the six-month window and the 16-for-1 split executes, resetting the share price toward a tenth of a dollar and pushing the float below the level at which retail and index coverage can function. Second, the amortization schedule outruns asset revenue for a second consecutive year, dragging gross margin into the low 30s and forcing the services business to subsidize the library. Third, the customer concentration produces a credit event that converts the receivable concentration into a write-off, and the 4.2 million cash balance is insufficient to cover the gap between the write-off and the bank loan rollover. In that scenario the listing is defended by further equity issuance at progressively lower prices, and the Class B holders, who control 61.9 percent of voting power, absorb the dilution in economic value while retaining control.
The base case is a managed drift. Revenue grows in the high teens on the film and TV recovery, the asset line holds flat to mid-teens growth, and gross margin stabilizes near 40 percent as the most expensive outsourced projects roll off. The 16-for-1 does not trigger because the stock stabilizes above the bid price. If it triggers, the shock is absorbed by a second repricing round similar to the December 2025 pipe. The net loss narrows as the warrant issuance artifact stops recurring, and the company survives on operating cash flow plus periodic raises, with the library continuing to amortize faster than it grows. The stock trades in a narrow band near the middle of the post-split range, and the equity is a dilution-adjusted call on the Chinese content cycle.
The bull case requires the AI multiplier to show up in the numbers. Gausspeed cuts production costs enough that the services line gross margin recovers to the mid 40s, asset licensing revenue doubles off the 11.6 percent base, and the library stops being a drag because the amortization schedule is outrun. In that case the 59.7 million asset base is worth more as an IP shop than as a cost line, and the reverse split arithmetic becomes a non-issue because the bid-price test is passed comfortably. The catalyst is a quarter where asset revenue grows faster than 30 percent and gross margin inflects up.
A counterargument to the whole framing deserves space. The 19.3 million net loss is not a 19.3 million economic loss. The warrant issuance charge is a financing cost booked through the income statement that does not consume cash, and the operating cash flow of 22.1 million is the real number. The company is, on a cash basis, generating more than it loses on the P&L, and the asset purchases are an investment, not an expense. The criticism of the reverse split also carries a counter: a split at 2.00 is cheaper to execute than a delisting, and the preapproval mechanism means the company does not have to run a proxy fight in the middle of a price decline. The risk is real, but it is a risk of dilution and of a lower-priced float, not a risk of insolvency.
The framework has to start from the constraint, not the multiple. At a share price near 2.13, the market capitalization is roughly 60 million on a post-split share count. That is about 1.1 times fiscal 2025 revenue, and it is negative on net income. The multiple analysis is therefore a question of what the asset base and the cash flow are worth, not of what the P/E should be, because the P/E is undefined. The balance sheet is the anchor. Total assets stand at 78.0 million, of which 59.7 million is intangibles. Total liabilities run to 16.1 million. Tangible net worth is negative, which means the equity value is a function of the library's earnings power, not its book value.
The quantified scenarios are built on revenue and margin, not on a multiple. The bear case assumes revenue of 55.0 million, with the asset line flat and gross margin of 38 percent. The operating loss after amortization is 4.0 million, plus a 3.0 million financing cost on a further dilution round. That produces an equity value near the 20 million cash-plus-current-assets mark, or roughly 0.40 per post-split share.
The base case assumes revenue of 65.0 million. The asset line grows 15 percent, and gross margin holds at 41 percent. Operating income of 3.5 million, with no further dilution, produces an equity value near 45 million. That is below the most recent raise price of 0.538. The outcome is the honest reading of the framework: the market is pricing this equity as a compliance option rather than a business. The bull case needs the Gausspeed margin story to print in at least two consecutive quarters, with the asset line re-rated from a cost line to an IP asset. In that case operating income reaches 8.0 million, and the equity value near 120 million is the only scenario where the current price clears the value.
The multiple context is thin but real, though the anchors are. Chinese VFX and virtual production peers trade, where listed, at revenue multiples in the 0.5 to 1.5 range. American digital asset licensing micro-caps have traded at 2.0 to 4.0 times revenue in years when the metaverse narrative was live. Global Mofy at 1.1 times revenue sits in the middle of that band, but the middle is only defensible if the margin trajectory is flat to improving. The fiscal 2025 print shows margin declining. The valuation is a function of the gap between the amortization run-rate and the revenue growth rate, and until that gap narrows, the 1.1 times revenue number is a ceiling, not a floor. The bear and base cases both sit below the current price, which is the honest reading of the framework.
The quarter that Global Mofy reported, and the two weeks it has operated since, define the stock more than the fiscal print does. The company took a 22.3 million non-cash warrant loss in that quarter, and it grew revenue 35.3 percent in the same window. It generated 22.1 million of operating cash flow, the cleanest line in the print. Then, within three months of the annual report, it ran a 50-for-1 split. It raised 2.0 million in that round. The company also preapproved a 16-for-1 consolidation to follow. The sequence is a listing defense, and the listing defense is the business now. The technology roadmap, the asset library, and the AI platform are the reason the stock has any value at all, but the near-term price action is driven by the compliance clock and the dilution schedule, not by the film and TV recovery that drove the revenue.
The central strategic initiative is the Gausspeed platform and the asset library expansion. What the company is positioning itself around is the gap between the two: the AI is meant to convert a 59.7 million amortizing asset base into a growing revenue engine, and that conversion is what separates the bull case from the base case. The fiscal 2025 print, with the asset line up 11.6 percent and gross margin down ten points, is the baseline against which that conversion is measured, and the baseline is unfavorable. The 2.0 million raise at 0.538 is the cost of buying the time for that conversion to show up in a quarter.
The monitoring variables, in order of importance, are the bid-price test against the six-month window that just opened, the asset line growth rate relative to the 11.9 million amortization run-rate, and the gross margin trajectory in the virtual technology service line. The receivable collection rate against the 16 percent top-customer concentration, and the next financing event, complete the list. The December 2025 pipe came in at 0.31875. The August 2026 direct followed at 0.538. That pattern shows that financing is a recurring feature, not an exception. Each of those variables resolves within the next two quarters, and the resolution of the first one, the bid-price test, carries more weight on the share price than the resolution of the other four combined.