XD Inc. (Trans): Growth Peaked; Betting on AI Creator Ecosystem

Dolphin Research summary of $XD INC(02400.HK) FY26 interim earnings call Trans

I. Key takeaways

1) Approx. RMB 50 mn in tax catch-up in H1; effective tax rate to normalize at 15%–25%.

a) Management said the catch-up is not company-specific and reflects tighter enforcement across the market. Companies need to adjust; whether it repeats in H2 is hard to predict, but tighter enforcement may be a trend.

b) Different subsidiaries face different statutory rates: 15% for High & New Tech Enterprises and 25% for regular entities. Overseas units largely fall within that band.

c) Historically, the group’s effective rate was low as some subsidiaries offset profits with prior losses. As those units turn profitable, offsets diminish, and the effective tax rate should normalize at 15%–25%.

d) On whether the preferential rate for high-tech enterprises could be removed, management said this is an industry-wide matter. Any change would apply to all enterprises equally.

2) AI-related opex rose by ~RMB 70 mn in H1; a notable decline will be recognized in H2.

a) Costs have two buckets: company-wide internal access (aggressive H1 rollout) and compute subsidies for creators via TapTap Maker.

b) Internal costs should fall sharply in H2 as open-source and commercial models get cheaper and staff become more proficient. On Maker, per-task and per-user costs continue to decline.

c) The monthly subsidy cap for Maker in H2 is expected to be only ~10% of H1 or even lower. Management expects subsidies can be phased out economically in the near future.

3) No new buyback plan; H2 capital return depends on operating performance.

Management said YTD buybacks are already sizable. Any new program will ‘depend on circumstances’, with H2 allocation decided by performance, and no amount, cadence, or timetable disclosed.

4) Apparent weakness in ADN revenue is an accounting presentation issue; underlying scale still growing.

ADN (TapTap Ad Network) looked down or flat HoH in H1 mainly because TapTap’s own placements took a higher share within ADN. This internal traffic is not booked as third-party revenue; management said overall scale is still growing and expects growth for the year.

II. Details from the call

2.1 Executive remarks

1) TapTap platform

a) This is TapTap’s 10th year. Management sees a milestone vs. its decade-ago goals, while acknowledging the platform may enter a bottleneck or stable phase.

b) Game supply on the platform surged in H1, and the supply logic is shifting. More individuals and players themselves are becoming developers, rather than only studio-led output.

d) Management welcomes the shift and sees opportunities, and will continue to explore this direction in H2. Future growth will come from new content and new vectors, not just extensions of the existing biz.

2) AI creation tools

a) The company launched AI game-creation tools for all creators in H1 and kept investing. TapTap Maker has been live for ~6 months and remains early, with focus on biz model exploration and core capabilities.

b) Internally, the company took a relatively open and aggressive stance on AI tools, granting access to all employees in H1. The aim was to help staff build habits quickly.

c) CD is even earlier. It began as an internal, one-stop, out-of-the-box AI agent for all staff, and is now external with token sales as the commercialization format.

d) The two are complementary: Maker is a pure online, cloud-based ‘computer in the cloud’ tool, while CD targets local development environments.

3) Self-developed games

a) ‘Dream Town’ completed global launch in H1. Management called it a relatively successful publishing effort.

b) The self-developed pipeline is progressing, with new projects entering R&D. Three new projects were added this period.

4) On industry change

a) Historically, studios followed a predictable path: ship a game, keep investing, scale the team, then build a bigger next title through a normal lifecycle.

b) AI breaks that certainty. Creation methods, game formats, and platform forms may all change, expanding the company’s optionality.

c) With AI, the industry turned cautious on large-scale investment. Management had anticipated a downcycle a year or two ago.

2.2 Q&A

Q: Has TapTap hit a growth ceiling? How do you view mobile headroom?

A: Without a major product or a new growth curve, growth will flatten. Management acknowledged this, but still sees upside on the developer side. This underpins investment in AI tools to help smaller developers monetize and bring more diverse games to users.

Cost reductions will show up first, while user-side uplift may lag by one or two cycles. The 10th year of TapTap means future growth must come from new content and new directions.

Q: H1 tax catch-up was ~RMB 50 mn. Will it recur in H2, and how to think about the tax rate?

A: Tighter enforcement could be a trend; the effective rate to normalize at 15%–25%. The catch-up was widespread across the market, and H2 is hard to predict. For structure and drivers, see Section I, item 1.

Q: AI costs rose ~RMB 70 mn in H1. Will they keep rising in H2, and will efficiency gains cut headcount?

A: AI costs will fall meaningfully in H2, but reducing headcount is not the primary goal. Management gave no quantified efficiency metrics; for cost composition and the path lower, see Section I, item 2.

Some roles will adjust as new tools, productivity, and workflows land. But the goal is to achieve more with fewer people and deliver better outcomes, while still attracting top talent rather than simply shrinking headcount.

Q: Any changes after Maker’s MCP integration and CD’s launch? What is CD’s positioning?

A: Both remain early; no deployment or user metrics disclosed. Maker has focused its first six months on biz model and foundational capabilities, with direction clearer, yet capabilities, creator base, and the commercial flywheel still have room to scale.

CD is earlier; despite opening externally, onboarding convenience and other aspects need work. It is still seeking product-market fit and a model, with modest investment and high daily internal usage.

Q: How does Maker’s R&D link with commercialization, and how do you balance compute cost vs. commercial upside?

A: Maker is not meant to make money standalone; it is a tool to feed quality games into TapTap’s ecosystem. It is not an independent product, but aims to help as many developers as possible ship products onto TapTap under a controllable economic model.

The key is whether ADN, Maker, and platform traffic together can create a healthy loop for developers within the ecosystem.

Q: What drove ADN’s HoH decline and the outlook?

A: The HoH softness is due to accounting presentation; the underlying scale is still growing. TapTap’s own placements took a higher share within ADN, which do not show up as third-party revenue. Management still expects growth this year.

Q: Progress on new games?

A: One project is on track; another remains early with no timetable. The earlier-stage title is smaller in scope, so once direction is set, progress should be faster.

Q: Will there be another buyback?

A: No new buyback committed; H2 allocation depends on performance. YTD buybacks are not small, and whether to continue will ‘depend on circumstances’.

Q: Commercialization progress for TapTap on PC?

A: No material PC commercialization progress or timetable disclosed. Over the next year, there is no clear plan for PC monetization, and the company is not focused on PC monetization for now.

Q: ‘Dream Town’ CN server recovered after two disruptions. What lessons were learned?

A: It comes down to the product itself; external factors matter but the core is the game. Long-term ops pose major challenges, and externalities can also affect the trajectory. Yet with ample publishing experience, management stressed the primacy of the product.

Regardless of pressure from players or rivals, the key is whether the product shows enough sincerity and delivers on user expectations. That remains the company’s main goal.

Q: The ‘Torchlight’ IP was brought in-house late last year and is in season 3. What’s next?

A: Continuing R&D investment; no new product plans disclosed. There is still significant room to improve the product from an R&D perspective and to keep developing around the IP, but no specifics on new titles, platforms, or timelines.

Q: User growth is up but CPM is down. How do you view product mix and advertiser appetite?

A: No direct response on CPM; MAU swings are mainly driven by the company’s user acquisition tactics. Top titles have limited impact on MAU; the real variable is which products can help acquire users and whether there are new growth tactics in external feeds.

Few games have added new users over the past two years, weighing on MAU. The outlook leans on TapTap-native games, aiming to help developers produce hits with better architecture and ecosystem support.

Q: How do you evaluate Maker ROI internally? Any changes to content curation?

A: Monthly subsidies will drop to ~10% of H1 levels or lower in H2. The core constraint is ‘how much to subsidize if fully open’ per month. The current model is sound enough to support many creators with this budget.

Subsidies might not be needed in the near future. The focus now shifts to scaling and rapid iteration to help creators make better games, cover more genres, and raise quality across categories.

Q: Over 6,000 new games were added in H1. Any breakout hits, and are ad budgets rising?

A: No hits by traditional standards; developer ad spend exists but is limited. Management values the creators’ progression from quality to monetization loops and sees ample room ahead.

Data: developer count did not expand aggressively in H1, while Maker developers, game count, and player base in China kept growing. Daily token consumption is ~100 bn and rising, with costs already low.

On Avg., each player plays 1.7 titles, and both time spent and retention are above the platform Avg.

Q: What is the plan for overseas? Tighten spend or keep experimenting?

A: Go overseas only after Maker, PC, and ADN are validated domestically. Overseas is an ecosystem move, not a single product. The aim is to help developers monetize and take the whole product ecosystem abroad with domestic developers.

Once the domestic business steps up, overseas priority will rise. For now, the focus is at home: ship enough strong games domestically and then use them as the overseas starting point.

Q: With token costs shouldered, is DAU/MAU growth for Maker on track? Will UGC dilute TapTap’s monetization and brand tone?

A: Monetization will only get bigger, not smaller; H1 user growth was deliberately constrained. Users naturally play multiple games, so the pie grows rather than shifting from heavy to casual.

User growth was held back in H1 because token costs only fell to a viable level by late H1. H2 will focus on creator growth, while brand tone will be protected through curation and algorithm optimization, a challenge shared across media industries.

Q: Will future game projects be AI-native?

A: No deliberate push for AI-native; focus on what was impossible before and is feasible now. The goal is not to make ‘AI games’ per se. The gameplay need not fail without AI; it is about enabling ideas that were previously impractical.

AI elements may grow and become highlights, but projects will not be back-solved from ‘AI-native’ positioning.

Q: Outlook for H2 and the summer season?

A: Summer is similar to last year; H2 likely sees no major change. No quantified guidance was given. Revenue may be slightly better, but new growth curves will land later, so the focus remains on new biz and growth drivers.

Q: ‘Dream Town’ is performing well overseas. Any learnings, and will Maker go abroad?

A: Overseas is not a separate play — all projects are built as global products. The company has a good track record in overseas publishing and focuses more on whether products can compete globally and adapt to local IP and cultures.

Q: What is the current pace and status of ‘Dream Town’ overseas ops?

A: Operations are normal overseas but will face similar issues as in China. Overseas launched later, but the essence of the challenges is the same. The only solution is to keep delivering on user expectations through ongoing version updates.

Q: Will there be systematic synergy between CD and TapTap? What about CD’s users and commercialization?

A: CD is still mainly used internally, monetized via token sales. External users are growing slowly, and sales and consumption experience are being refined. Management sees significant growth headroom.

CD complements Maker: Maker is a pure online, cloud computer-style tool, while local environments remain critical for many creators. Synergy aims to make building, publishing, and data monitoring on TapTap easier, while staying open so publishing is not limited to TapTap.

Q: Will the supply surge disrupt community rules and algorithms? Any H2 adjustments?

A: Multiple adjustments were made in H1 and are largely done; current strategy can handle a 5–10x increase. Management will keep tracking satisfaction and feedback while iterating recommendations. If supply rises by tens of times, further iterations will be needed.

Q: ADN outlook and long-term plan? Still prudent?

A: Remain prudent; ADN aims to give game developers a cleaner monetization option. Two goals: create ad-monetization value for TapTap’s internal ecosystem and address issues seen in existing networks by starting with games to deliver a cleaner option with higher developer revenue.

China’s ad network market is large but messy, and the company is still exploring. Long term, ADN should at least help Maker and IAA-type developers on TapTap, with current focus on product quality and TapTap synergy.

Q: Are AI model advances more a challenge or an opportunity for game companies?

A: More opportunity, but with bigger variables — the path is no longer fixed. Previously, continued investment after a launch could sustain an IP as teams scaled, creating barriers for smaller studios through IP, capital, and team accumulation.

With AI, small teams or individuals can achieve similar outcomes; film is the clearest example. For a company of this size, management sees more opportunity than challenge, with richer paths and more possibilities.

Q: With three new in-development projects, has the game strategy turned more optimistic?

A: Constructive but far from certain; more projects, smaller early spend per title. Current projects are still built the ‘old way’, but that will not be how games are made in 1–2 years. The approach is to keep learning and watching — essentially ‘changing engines mid-flight’.

As with traditional 3D animation facing the ‘single-creator high-quality’ shift, games will face a similar transition. The company remains cautious on new spend, though that may not mean fewer projects.

Q: If production methods shift in 2–3 years, must you accelerate? Any launch timetable?

A: Accelerate by changing methods, not by adding spend; no timetable to disclose. The aim is higher speed, output, and quality with the same resources. The company will not expand teams or budgets just due to pressure.

The change is too large to lock a title that launches in 2–3 years right now. Any disclosable timetables will be announced promptly.

Q: Without IAPs, how does ad-only monetization compete with WeChat and Douyin? How to solve payments?

A: Lack of IAPs pushes some developers to WeChat Mini Games; payments are policy-constrained and will not be quick. Even without IP or UA budgets, good games can still get free traffic on TapTap, which meaningfully differs from WeChat and Douyin.

But the absence of IAPs does push some developers away. The company will keep exploring payment solutions, though domestic policy constraints mean it will take time; if it works domestically, overseas offers big room as well.

Q: Of the three in-development projects, what is the undisclosed one?

A: Only that it is prepared by another team and remains early; no information disclosed. Management added there are more than three internal projects, most at very early stages. The number of games is not the key, given the large future variables.

<End of text>

Risk disclosure and statements:Dolphin Research disclaimer and general disclosure