Alibaba

ByteDance

CATL

Huawei

Meituan

Tencent

Xiaomi

Most Followed

KrASIA Connection

100 AI Creators

China's Auto Sector

Inside China's Startups

Venture Roundup

Our Events

View All

Collections

News

Features

Insights

Ecosystem

Press Releases

DeepSeek

Li Auto

Luckin Coffee

Mixue Bingcheng

Pop Mart

Xpeng

On the Radar

Events

Latest

About

Consumer

Robotics

Fintech

Collaborate

At this year’s Auto Guangzhou, <a href="/company/nio">Nio</a> founder and CEO William Li said Firefly has partnered with <a href="/company/ant-group">Ant Group’s</a> Zhima Credit and automotive platform SuperEV to launch a credit-based car subscription service. The program is designed to give users a low-barrier way to experience Firefly and decide whether the sub-brand fits their needs.
<h3>Why car subscription, and why now?</h3>
For years, car buying in China centered on ownership, with high down payments, long loans, and extensive paperwork. The rise of new energy vehicles is shifting this model. Faster product cycles, more advanced features, and less predictable resale values have made traditional buy-and-replace patterns less efficient. Consumers are switching cars more often, and flexibility is becoming more important.
According to <a href="https://36kr.com/p/3562402465725570">36Kr</a>, SuperEV founder Sun Zefeng describes car subscription as a third mode of vehicle access, distinct from purchasing and renting. Users pay periodically to use a vehicle, while insurance, taxes, and maintenance are bundled into the fee. Compared with buying, subscription reduces long-term commitments, and compared with renting, it offers lower costs and a stronger sense of attachment. Flexibility, cross-brand choice, and integrated operations form the model’s core advantages.
Global momentum supports this shift. Boston Consulting Group projects that by 2030, Europe’s car subscription market will represent 15% of new car sales, exceeding USD 30 billion. Research and Markets forecasts global subscription revenue rising from USD 8.09 billion in 2024 to USD 30.29 billion in 2029, a compound annual growth rate of around 30%. In China, Grand View Research estimates the market will reach USD 2.08 billion by 2030.
The model aligns strongly with younger consumers who prefer service-based access over ownership, and it fits automakers that now rely more on recurring revenue from software, batteries, and membership services. “Car subscription frees you from being tied to a single car, while giving you the excitement of an entire vehicle library,” Sun told <a href="https://36kr.com/p/3562402465725570">36Kr</a>.
The seven-day Firefly subscription is positioned as China’s first credit-based car subscription. Users with a Zhima score of 800 or higher can access the service without deposits or complex documentation. Through SuperEV, they can drive a Firefly vehicle with one tap and access the Nio ecosystem for seven days, including app features, Nio House access, and driving perks. The plan supports daily commuting and weekend trips and is intended as a meaningful trial before purchase. Afterward, users can switch to a monthly subscription, buy the vehicle, or try a different model.
<h3>Three-way collaboration</h3>
Each partner contributes distinct capabilities. Zhima Credit provides the trust infrastructure for credit-based access. Firefly provides the vehicle and brand experience. SuperEV delivers subscription infrastructure and user operations. Together, they are targeting consumers who value quality, trust, and flexibility.
SuperEV helps both Zhima 800 and Firefly reach broader user circles. For Firefly, this is its first attempt to pair its vehicles with a try-before-you-buy credit model, reducing the threshold for experiencing a premium compact <a href="/industry/ev">electric vehicle</a>. Zhima 800, which serves users with credit scores of 800 and above, extends access to Nio House and other perks and lowers barriers to engaging with premium EVs.
Firefly appeals to urban consumers and young drivers with its compact design, safety features, premium interior, and smart cabin. These users are often families purchasing an additional vehicle or first-time buyers seeking refinement, which aligns naturally with a subscription-first approach. The broader Nio ecosystem, including Nio House and Nio Life, supports the experience with integrated services and community offerings.
SuperEV has built a subscription system that spans short-term trials through long-term use. Its structure covers trial subscriptions, flexible subscriptions for short-term needs under a year, and long-term plans for commitments of more than 24 months.
The Firefly program is designed to reduce the barrier to premium EV access. It provides an immersive but low-commitment trial that improves on traditional rental formats. SuperEV’s operational infrastructure connects credit screening to real-world use, reducing friction for users.
In effect, the three companies have created a targeted service system for high-credit users, aligning benefits and experiences to strengthen loyalty and conversion. “We are jointly serving the same group of quality-seeking consumers,” Sun said. “Through the strongest possible experience, we satisfy their desire to try new things.”
<h3>Looking ahead</h3>
The long-term aim is to support a broader transition from trial to regular subscription use:
<ul>
<li aria-level="1">First, the partners make credit and subscription tangible through high-quality products and service standards.</li>
<li aria-level="1">Second, pairing a premium EV with a high-credit user base helps validate the model’s feasibility in the higher-end market.</li>
<li aria-level="1">Third, the collaboration establishes benchmarks for credit-based mobility services that could extend to more vehicle classes.</li>
</ul>
Future expansion may include SUVs, MPVs, sports cars, flying vehicles, and yachts, enabling a model in which one subscription unlocks access to multiple vehicle types across commuting, travel, and leisure. Subscription packages could integrate additional perks such as mileage-based charging or priority battery swaps, forming a premium mobility membership system.
As acceptance grows, tiered credit thresholds and varied subscription plans could expand the model to broader demographic groups. With smart city infrastructure developing, high-credit users may gain streamlined parking, prioritized charging, and other integrated mobility benefits.
For now, the partnership marks an early move toward a consumption model built on trust, flexibility and service. It points to a future where cars function less as owned assets and more as modular, on-demand spaces, and where credit effectively becomes a gateway to personalized mobility. By aligning user experience with product access, the companies seek to accelerate the shift from vehicle ownership to subscription and help shape the next phase of China’s automotive market.

Nio teams up with Ant’s Zhima and SuperEV on credit-based car subscription

Ant Group

East Asia

China

Transportation

Electric Vehicles

Mobility

Inno100’s global innovation flagship store opened in Shenzhen on November 22, marking the start of its partnership with Kickstarter, the world’s leading creative crowdfunding platform. The store is Kickstarter’s first officially authorized retail location in China.
At the opening ceremony, an Inno100 representative said the space is meant to function both as an early-stage incubator for new products and as a channel for early brand exposure, helping reduce market education costs.
Based on on-site observations by 36Kr, the flagship store has already attracted several professional hardware makers, including <a href="/company/anker">Anker Innovations</a>, Insta360, HoverAir, xTool, and Sharge.
The store is organized around several zones including a <a href="/industry/robotics">robotics</a> interaction area, a customization workstation, and a creativity showroom:
<ul>
<li aria-level="1">The robotics area features humanoid and quadruped robots as well as exoskeleton devices.</li>
<li aria-level="1">The customization workstation integrates laser engraving, 3D printing, and UV printing to provide instant personalization for items purchased in-store.</li>
<li aria-level="1">The creativity showroom highlights products designed for artificial intelligence-assisted work, education, and home use.</li>
</ul>
All products on display are available for purchase.
One standout product is Anker’s <a href="https://kr-asia.com/eufy-and-eufymake-mark-ankers-second-act-in-lawn-mowers-and-uv-printing">EufyMake UV printer</a>. “This is the only place where you can see and physically interact with an actual EufyMake UV printer,” said Peng Yiheng, Kickstarter’s chief strategic representative in China.
According to Inno100, the store recorded a 70–80% footfall rate during its ten-day trial run, with the average visitor staying more than 30 minutes. In retail, both metrics are key indicators of a strong offline experience and, the company noted, currently exceed industry averages.
“Tech products need storytelling and hands-on experience,” Peng said. “And the biggest advantage of a physical store is that it offers enough time and space for this kind of deep interaction.”
In this sense, Inno100 acts as both filter and amplifier for hardware innovation. It curates distinctive early-stage hardware products and uses immersive experiences and customization services to help these brands build early user awareness and brand equity.
KrASIA Connection features translated and adapted content that was originally published by 36Kr. This <a href="https://36kr.com/p/3566873750977673">article</a> was written by Ou Xue for 36Kr.

Kickstarter makes China debut with Inno100’s flagship store in Shenzhen

Retail

Consumer Economy

Haomo.ai, the driving assistance technology company backed by Great Wall Motor (GWM), has abruptly halted operations after instructing all employees to stop reporting to work starting November 24, according to internal notices shared by staff with Chinese media.
In a message circulated on November 22, the company told employees to cease work and await further instructions. The notice did not specify compensation arrangements or indicate whether operations would resume.
Employees told outlets including <a href="https://www.jiemian.com/article/13673074.html">Jiemian News</a> that attendance tracking on DingTalk was disabled and internal group chats were muted shortly after the announcement. Workers in Beijing also reported that salaries have been delayed for two months, with no communication from senior leadership about settlement plans or next steps.
GWM, Haomo.ai’s controlling shareholder, has not commented publicly. The shutdown caps a year of escalating instability marked by layoffs, frozen projects, and the departure of multiple executives.
Haomo.ai reportedly <a href="https://tech.ifeng.com/c/8oYm9JpMaQc">cut 30–50% of staff in late 2024</a> across both corporate and technical functions. The chairman and several vice presidents overseeing technology and product also left over the past year. By mid-2025, leadership churn had already raised concerns about the company’s ability to deliver on its roadmap.
Those concerns widened as Haomo.ai repeatedly <a href="https://www.epochtimes.com/gb/25/11/24/n14642344.htm">delayed its urban NOA (navigate-on-autopilot) program</a>. The system had initially been slated for rollout across multiple GWM models in 2022 and 2023, but delivery targets slipped repeatedly. Earlier this year, insiders told media that even limited features scheduled for August deployment on two Hyundai models were at risk.
A key challenge was Haomo.ai’s adherence to a technology stack built on high-definition maps and rules-based algorithms. As competitors shifted toward end-to-end perception models and map-lite or mapless architectures, analysts and employees said Haomo.ai struggled to keep pace.
As the company fell behind, GWM shifted advanced driving programs to Momenta in 2024. DJI provided solutions for midtier and entry-level models under the Haval and Ora brands. Haomo.ai’s remaining passenger vehicle business narrowed to the two Hyundai programs, while its previously promoted last-mile delivery robot line contracted sharply, with no new models under development and operations focused on clearing inventory.
Several of the company’s <a href="https://news.china.com/socialgd/10000169/20251125/49020753.html">expansion projects were also suspended</a>, including its planned domain controller factory in Zhejiang and a robotics plant in Chengdu.
Across multiple fundraising rounds, Haomo.ai raised several billion RMB and reached a valuation above USD 1 billion at its peak. But by late 2024 and into 2025, salary delays, shrinking activity, and headcount cuts pointed to mounting financial strain. Its push for a Hong Kong IPO was also reported to have stalled, despite earlier public statements that preparations continued.
Signs of deeper structural pressure had <a href="https://kr-asia.com/haomo-ai-hits-a-crossroads-amid-leadership-shakeup-and-tech-setbacks">emerged months earlier</a>, including weak commercialization across its two core business lines and a sharply reduced sales target for delivery robots. GWM’s rapid expansion of its internal smart driving R&amp;D, and its investment in Pony.ai’s end-to-end foundation model, further reduced Haomo.ai’s strategic relevance within the parent company’s ecosystem.
Once a unicorn backed by one of China’s largest automakers, Haomo.ai now faces an uncertain future. For remaining employees, the sudden shutdown offers little clarity and few signs that operations will resume.

Haomo.ai halts operations as employees told not to report to work

Autonomous Vehicles

At the DDC 2025 conference, D-Robotics announced a slate of upgrades to its development infrastructure and introduced the RDK S600, a compute platform built for <a href="/industry/embodied-intelligence">embodied intelligence</a>. The S600 is expected to launch in the first quarter of 2026.
The company also unveiled a development platform that integrates three core capabilities:
<ul>
<li aria-level="1">First, it includes a data closed-loop system that links models, simulation, and hardware to support autonomous system iteration.</li>
<li aria-level="1">Second, it provides an embodied intelligence training ground that offers full-chain task support across varied scenarios.</li>
<li aria-level="1">Third, it introduces agentic development tools that enable developers to build and deploy robotic applications using natural language.</li>
</ul>
As <a href="/industry/robotics">robotics</a> move into more complex environments, neither hardware nor software alone can meet performance demands. Increasing task complexity and more advanced models have intensified the gap between on-device compute limitations and real-time requirements. Cloud resources can offset those limits, which makes edge-cloud collaboration a natural progression. Rising competition is also putting pressure on development speed. Hardware and software co-design and integrated architectures improve performance while shortening the cycle from R&amp;D to deployment.
To address both compute constraints and development efficiency, D-Robotics has built a full-chain system spanning device and cloud. On the device side, the company continues to refine its BPU (Bernoulli processing unit) architecture and has developed two product lines for different compute tiers. On the cloud side, its platform provides hundreds of ready-to-deploy robotics algorithms to help teams build competitive solutions more efficiently.
D-Robotics also showcased the S600, which delivers 560 TOPS (tera operations per second) of compute (at INT8 precision) and supports edge deployment of large models. When running models such as Pi-Zero and Qwen2.5-VL-7B, its performance reportedly exceeds mainstream platforms by more than 2.2 times, meeting the decision-making demands of real-world environments.
In interviews with 36Kr and other outlets, CEO Wang Cong said the company does not intend to deliver turnkey end products. Instead, it focuses on identifying common technical needs across scenarios and refining them into standardized modules. Stereo vision, for example, is used in categories spanning robot vacuums, lawn mowers, drones, and quadruped robots. By targeting such bottlenecks, D-Robotics aims to support a wide range of product requirements. In embodied intelligence, its motion control algorithms are optimized for adaptation and generalization.
“For R&amp;D needs, we want to turn common processes such as data collection, labeling, generation, simulation, and testing into standardized toolboxes,” Wang said. “Like PowerPoint, we don’t interfere with creativity. We provide robust foundational tools that support highly abstract, multi-scenario R&amp;D workflows.”
D-Robotics is prioritizing three deployment areas: mass produced robotic products, emerging robotics applications, and future general-purpose embodied robots.
According to 36Kr, the company has partnered with more than 60 players across the robotics supply chain. These include supporting Narwal on the Xiaoyao 002 robot vacuum and bringing AI-powered stereo perception to the category; enabling Insta360’s launch of what is billed as the world’s first panoramic drone, the Antigravity A1; and working with Vbot on a quadruped companion robot.
Over the past year, D-Robotics said product shipments rose roughly 180% year-on-year, while its customer base grew about 200%.
KrASIA Connection features translated and adapted content that was originally published by 36Kr. This <a href="https://36kr.com/p/3566850799762563">article</a> was written by Huang Nan for 36Kr.

D-Robotics unveils upgrades to tighten the loop between cloud and edge robotics

Deep Tech

Embodied Intelligence

Bambu Lab is close to completing its latest funding round, 36Kr reported, citing multiple investor and industry sources. <a href="/company/tencent">Tencent</a> is said to be participating, and the company’s valuation could reach USD 100 billion upon closing.
There were signs earlier. On July 30, Bambu Lab and Tencent Cloud announced a broad partnership in which Bambu’s 3D model platform, MakerWorld, would integrate Tencent’s Hunyuan 3D generative model. That move aligns with information from an unnamed investor who told 36Kr that the two firms would explore software-related collaboration once Tencent invested. When asked directly, however, both companies denied the information.
Notably, Tencent is also a shareholder in Creality, a 3D printer maker preparing for a Hong Kong IPO.
Bambu Lab has become one of the private market’s brightest standouts. According to Soochow Securities, the company was projected to generate up to RMB 6 billion (USD 840 million) in revenue last year, with close to RMB 2 billion (USD 280 million) in net profit. Media reports have also indicated that Bambu Lab has shipped more than 1.2 million units, claiming a 29% global share.
Its last round closed in 2023 with investors including Mingshi Investment Management, Temasek, True Light Capital, and 5Y Capital. Investment amounts were not disclosed.
Bambu Lab’s rise reflects the rarity of its position: it found a fast-growing market and brought products roughly two generations ahead of competitors.
Precedence Research estimates the global 3D printing market will reach USD 24.61 billion in 2024, growing at a compound annual rate of 18.5% through 2034.
The industry is in an expansion cycle. Before Bambu Lab arrived, most desktop printers were open-source, hands-on devices requiring manual leveling and maintenance. Bambu Lab introduced automatic leveling, a fully enclosed structure, and a signature multicolor system, solving two longstanding pain points: speed and complexity. Its printers are several times faster than traditional machines and require far less user intervention.
“Bambu Lab greatly expanded multicolor printing. From the moment it launched its first product to when competitors finally offered their own high-speed, multicolor systems, there was a full two-year gap, and Bambu used that time to scale,” an industry insider told 36Kr.
In this person’s view, the company’s founding team has executed almost flawlessly. “Sometimes we envy Tao Ye. Building a 3D printing company for him seems like taking an open-book exam, he already knows the answers.”
Bambu Lab’s success has triggered an industrywide arms race. As peers follow, consumer printers are becoming more turnkey, with higher print success rates and more polished software. Printers are now turning up in everyday households, for school projects, gifts, and small items, rather than just in maker communities.
Bambu Lab also dominates the enterprise segment. The company said more than 70% of 3D printing farms globally run on its P1 series. Many hardware firms visited by 36Kr use Bambu printers internally for prototyping and small-batch parts.
Rising competition, however, is inevitable. Most major manufacturers have launched high-speed models in the past two years, and many are rushing to replicate Bambu’s multicolor approach.
whereRelatedArticle
The most unnerving move comes from <a href="/company/dji">DJI</a>, founder Tao’s former employer, which recently invested in Elegoo. DJI confirmed the investment, saying it reflects optimism about consumer-grade 3D printing and aligns with its technology strategy.
Elegoo is already strong in manufacturing, with its own factory and production lines. On November 10, it launched the China-facing version of Nexprint, its open-source model aggregation platform integrating libraries like Thingiverse and Printables. Analysts view Nexprint as a direct challenge to Bambu’s MakerWorld.
KrASIA Connection features translated and adapted content that was originally published by 36Kr. This <a href="https://36kr.com/p/3566526394481543">article</a> was written by Leslie Zhang for 36Kr.

Bambu Lab nears USD 100 billion valuation as Tencent eyes new investment

Hardware

Creative

… id="caption-attachment-158423" class="wp-caption-text">Through MindOn’s approach, bipedal robots could even become playmates for activities such as a game of frisbee. Graphic source: MindOn Robotics.</figcaption></figure>
Discussing the company’s financing, Zhu said MindOn’s technical differentiation is the main reason investors committed quickly. He noted that these investors are already heavily involved in embodied intelligence but chose MindOn because its technology complements their broader portfolios.
Zhu predicts that as training efficiency improves, the timeline for humanoid robots entering households will shorten to three to five years. Before that, he expects bipedal versions to enter retail or fast casual dining environments, such as unmanned fast food outlets, within one to two years. These environments feature fixed tasks and controlled conditions that support rapid validation and commercial deployment.
Asked about MindOn’s moat, Zhu said:
<blockquote>“When everyone believed in teleoperation, we recognized its fundamental flaw and found another path. Our real moat is the ability to take an immature idea and turn it into reality, step by step.”</blockquote>
The following transcript has been edited and consolidated for brevity and clarity.
<h3>36Kr: Why do you believe teleoperation has a fundamental flaw?</h3>
Zhu Qingxu (ZQ): At its core, teleoperation forces a person to use the brain’s “slow system” to control a robot. The operator must observe, think, and then act. That process is inherently slow, jerky, and full of unnecessary pauses.
Training robots like this is akin to learning from a teacher who moves unnaturally. It locks the robot’s performance ceiling. All those robot videos you see that require sped-up playback? That is the root cause.
For tasks requiring tactile sensitivity, such as twisting a bottle cap, teleoperation is even worse. Without true force feedback, operators cannot tell whether the robot’s hand is gripping correctly, which lowers efficiency.
<h3>36Kr: If teleoperation is flawed, why did it become so widely adopted?</h3>
ZQ: The initial goal was simple: to get robots to manipulate real objects and collect real-world data. Teleoperation was the first feasible solution.
<h3>36Kr: How does your system work, and what are its advantages?</h3>
ZQ: It balances data quality with data scale. Motion capture records natural, instinctive full-body movements. UMI allows operators to manipulate objects at scale, capturing detailed hand-object interactions. Together, they generate far higher quality data than video-only approaches and in far larger quantities than teleoperation. Essentially, it captures the subconscious actions robots should be learning.
<h3>36Kr: Once the data is collected, how does your architecture work?</h3>
ZQ: It mirrors how intelligence forms. The “cerebellum” learns basic human actions such as walking, running, crouching, grasping, and pulling. Using motion capture data, we train these in simulation. Once built, the library becomes universal.
The “cortex” handles perception, instructions, and planning, calling the cerebellum’s skills to execute actions. They evolve together. The richer the cerebellum’s skills, the more the cortex can do, and the smarter the cortex becomes, the more precisely it can deploy those skills.
<h3>36Kr: You mentioned humanoid robots entering unmanned stores in one to two years. How would that work?</h3>
ZQ: Tasks in places such as unmanned fast food outlets are limited and enumerable. We can capture every action in the motion capture lab. Because our data quality is high, robots learn these primitive actions efficiently. It takes only two to three days to learn all the roles. Then we collect UMI data in the real store to support generalization. Teleoperation cannot match this efficiency.
<h3>36Kr: Motion capture requires a studio filled with cameras. You are not going to build one inside a kitchen, so how do you handle scenario-specific teaching?</h3>
ZQ: You do not need to. Actions can be captured fully in the lab. Household or kitchen tasks break down into finite movements such as grasp, place, lift, and shake. These can all be captured with motion capture equipment. Then we gather UMI and environment data on-site.
<h3>36Kr: What is the biggest challenge in generalizing from controlled environments to real households?</h3>
ZQ: Generalization is the biggest challenge. We need to address three types: object generalization, positional generalization, and scene generalization. This requires substantial and diverse scene data. We believe in scaling laws, but only when data quality and quantity are sufficient.
<h3>36Kr: Why will the embodied intelligence system that enters households likely take the humanoid form?</h3>
ZQ: Humanoid robots have challenges such as stability and control, but their advantages outweigh the downsides. Homes are designed for people. Many layouts and tasks are unfriendly to wheels, including stairs, uneven levels, and thick carpets. Tasks that involve changing heights also favor the humanoid form. For non-human forms, you would need to enumerate and solve shape-specific problems continuously. Training relies on human motion data, and such datasets do not exist for non-human forms.
<h3>36Kr: How will users teach robots new tasks once they enter homes?</h3>
ZQ: Ideally, robots will come equipped with most commonly employed skills. For new tasks, we may later release a simple demonstration tool that allows users to teach a task once through hands-on demonstration, after which the robot can learn through observation and minimal practice.
<h3>36Kr: Some say algorithms are weak products and that hardware companies will eventually build their own algorithms. Where is MindOn’s moat?</h3>
ZQ: Technology has no permanent moat, only temporary leads. Our moat is judgment and iteration. When everyone believed in teleoperation, we recognized its fundamental flaw and found another path. Our real moat is the ability to take an immature idea and turn it into reality, step by step.
<h3>36Kr: What is your outlook for the sector?</h3>
ZQ: The sector will go through a shakeout. The companies that survive will be the ones “working out,” not “putting on makeup.” For us, we aim to deepen our technical foundation and focus on long-term capability. When the industry matures, we want to remain standing.
KrASIA Connection features translated and adapted content that was originally published by 36Kr. This <a href="https://36kr.com/p/3559983402023814">article</a> was written by Fu Chong for 36Kr.

Is teleoperation holding robotics development back? MindOn Robotics makes its case

… https://console.kr-asia.com/wp-content/uploads/2025/11/Hoyoverse-Varsapura-Mr-Shadow-720x405.jpg 720w, https://console.kr-asia.com/wp-content/uploads/2025/11/Hoyoverse-Varsapura-Mr-Shadow-1024x576.jpg 1024w, https://console.kr-asia.com/wp-content/uploads/2025/11/Hoyoverse-Varsapura-Mr-Shadow-768x432.jpg 768w, https://console.kr-asia.com/wp-content/uploads/2025/11/Hoyoverse-Varsapura-Mr-Shadow-1536x864.jpg 1536w, https://console.kr-asia.com/wp-content/uploads/2025/11/Hoyoverse-Varsapura-Mr-Shadow-2048x1152.jpg 2048w, https://console.kr-asia.com/wp-content/uploads/2025/11/Hoyoverse-Varsapura-Mr-Shadow-480x270.jpg 480w" sizes="(max-width: 2560px) 100vw, 2560px" /><figcaption id="caption-attachment-158448" class="wp-caption-text">Mr Shadow, shown in his first appearance in the demo in an encounter with Hollowone as she undergoes her SEAL interview. Image source: Hoyoverse.</figcaption></figure>
Combat hews closely to Hoyoverse’s trademark style: fast, fluid encounters built around character swapping and chained abilities against a variety of shadowy foes. One notable motif is the recurring symbolism of the umbrella, structured by the game as a multitoll used as both weapon and narrative marker. In a cutscene between Hollowone and Mr Shadow in the demo, the former discloses that the person she’s looking for within SEAL carries a snake-scale umbrella.
<figure id="attachment_158449" aria-describedby="caption-attachment-158449" style="width: 1920px" class="wp-caption alignnone"><img loading="lazy" decoding="async" class="wp-image-158449 size-full" src="https://console.kr-asia.com/wp-content/uploads/2025/11/Hoyoverse-Varsapura-Umbrella.jpg" alt="Image of Hollowone using an umbrella to glide through the air." width="1920" height="1080" srcset="https://console.kr-asia.com/wp-content/uploads/2025/11/Hoyoverse-Varsapura-Umbrella.jpg 1920w, https://console.kr-asia.com/wp-content/uploads/2025/11/Hoyoverse-Varsapura-Umbrella-720x405.jpg 720w, https://console.kr-asia.com/wp-content/uploads/2025/11/Hoyoverse-Varsapura-Umbrella-1024x576.jpg 1024w, https://console.kr-asia.com/wp-content/uploads/2025/11/Hoyoverse-Varsapura-Umbrella-768x432.jpg 768w, https://console.kr-asia.com/wp-content/uploads/2025/11/Hoyoverse-Varsapura-Umbrella-1536x864.jpg 1536w, https://console.kr-asia.com/wp-content/uploads/2025/11/Hoyoverse-Varsapura-Umbrella-480x270.jpg 480w" sizes="(max-width: 1920px) 100vw, 1920px" /><figcaption id="caption-attachment-158449" class="wp-caption-text">Hollowone uses an umbrella as both weapon and tool, enabling her to unleash a series of magical abilities and glide through environments. Image source: Hoyoverse.</figcaption></figure>
The game’s central threat, mindrot, ties the story to themes of mental health. Mindrot appears as a tar-like substance that leaks from objects and manifests into monsters. SEAL is introduced as the body responsible for containing these incidents and preventing further spread, as advanced forms of mindrot give rise to fully fledged shadow creatures. A brief explanation in the demo describes humanity’s collective consciousness as the “rain of thought,” and suggests that when this system is improperly drained, rot forms, becoming mindrot.
<figure id="attachment_158452" aria-describedby="caption-attachment-158452" style="width: 1920px" class="wp-caption alignnone"><img loading="lazy" decoding="async" class="size-full wp-image-158452" src="https://console.kr-asia.com/wp-content/uploads/2025/11/Hoyoverse-Varsapura-Shadow-Monster.jpg" alt="Image of a shadow monster that forms as a result of mindrot." width="1920" height="1080" srcset="https://console.kr-asia.com/wp-content/uploads/2025/11/Hoyoverse-Varsapura-Shadow-Monster.jpg 1920w, https://console.kr-asia.com/wp-content/uploads/2025/11/Hoyoverse-Varsapura-Shadow-Monster-720x405.jpg 720w, https://console.kr-asia.com/wp-content/uploads/2025/11/Hoyoverse-Varsapura-Shadow-Monster-1024x576.jpg 1024w, https://console.kr-asia.com/wp-content/uploads/2025/11/Hoyoverse-Varsapura-Shadow-Monster-768x432.jpg 768w, https://console.kr-asia.com/wp-content/uploads/2025/11/Hoyoverse-Varsapura-Shadow-Monster-1536x864.jpg 1536w, https://console.kr-asia.com/wp-content/uploads/2025/11/Hoyoverse-Varsapura-Shadow-Monster-480x270.jpg 480w" sizes="(max-width: 1920px) 100vw, 1920px" /><figcaption id="caption-attachment-158452" class="wp-caption-text">A shadow monster that forms as a result of mindrot. Image source: Hoyoverse.</figcaption></figure>
Varsapura leans further into the paranormal with scenes showing distortions of reality. In one sequence, Hollowone enters an alternate dimension as part of her SEAL evaluation, where a set of camera-like devices collectively opens a gateway. Operatives appear capable of manipulating the environment inside, a skill that becomes urgent when mindrot suddenly begins seeping into the space and Hollowone and Sayuki attempt to escape.
<figure id="attachment_158454" aria-describedby="caption-attachment-158454" style="width: 480px" class="wp-caption aligncenter"><img loading="lazy" decoding="async" class="wp-image-158454 size-full" src="https://console.kr-asia.com/wp-content/uploads/2025/11/Hoyoverse-Varsapura-Escape.webp" alt="A sequence showing Hollowone and Sayuki rushing to escape the dimension where Hollowone’s final SEAL assessment is taking place before it becomes overwhelmed by mindrot." width="480" height="270" /><figcaption id="caption-attachment-158454" class="wp-caption-text">Hollowone and Sayuki rush to escape the dimension where Hollowone’s final SEAL assessment is taking place, before the space becomes overwhelmed by mindrot. Graphic source: Hoyoverse.</figcaption></figure>
The title adds to Hoyoverse’s lineup of announced projects, which includes Petit Planet and Honkai: Nexus Anima. Whether familiar mechanics, at least as presented so far, will support or complicate the company’s expansion remains unclear. Hoyoverse’s earlier hits, including Genshin Impact and entries in the Honkai series, have long relied on recurring events and activities to maintain engagement. That formula has been broadly effective, but its time-consuming nature may test players’ willingness to adopt new titles with similar systems at the same time.

What we know so far about Varsapura, Hoyoverse’s next game

Gaming

… (max-width:768px){.embed-container{height:540px!important}}</style>
<div class="embed-container"><iframe src="https://flo.uri.sh/visualisation/26372158/embed" frameborder="0" allowfullscreen> frameborder="0" scrolling="no" sandbox="allow-same-origin allow-forms allow-scripts allow-downloads allow-popups allow-popups-to-escape-sandbox allow-top-navigation-by-user-activation"></iframe></div>
Meanwhile, the number of operating cinemas reached a record 13,400 nationwide, with 108 million screenings. Yet the number of films surpassing RMB 100 million (USD 14 million) is at its lowest in years. As more screens open and fewer hits emerge, theater owners face shrinking margins and tougher operating conditions.
A more concerning trend is the loss of younger viewers. According to TopCDB, in 2025 the share of viewers aged 25 and under dropped to a new low. Few films cater directly to younger audiences, and among those that surpassed RMB 500 million (USD 70 million), only The Legend of Hei 2 and Detective Chinatown 1900 had youth audiences making up more than 20%.
Shortform videos and dramas are drawing attention away from cinemas. TopCDB reported that the short drama market surged in the past two years, reaching RMB 48.46 billion (USD 6.8 billion) last year, more than triple its size from two years earlier.
When Zootopia premiered in 2016, a third-year middle schooler could enjoy it as a tween. Nine years later, that same viewer may now be an office worker with a stable income. Whether Zootopia 2 can bring younger audiences back to theaters may be the question exhibitors care about most.
KrASIA Connection features translated and adapted content that was originally published by 36Kr. This <a href="https://36kr.com/p/3559591276821639">article</a> was written by Wang Yuchan for 36Kr.

Zootopia 2 lands in China as Hollywood’s biggest test of the year

Entertainment

Video

Content

… pricing differential is based on industry estimates and is not independently verified.
This year, Russia and several Central Asian countries introduced restrictive policies. In April, Russia required Chinese vehicles transiting through Central Asia to pay all applicable taxes and fees upon entering the country. This removed the cost advantage of parallel-imported Chinese vehicles. As a result, Li Auto’s parallel export volume has dropped to a few hundred units per month.
China’s Ministry of Commerce and three other ministries also tightened rules governing used car exports. Starting January 1, vehicles registered for fewer than 180 days will not qualify for export licenses. This change effectively closes Li Auto’s parallel export channel.
While Li Auto hesitated, companies such as <a href="/company/xpeng">Xpeng</a>, <a href="/company/leapmotor">Leapmotor</a>, and Nio made overseas expansion a central part of their strategies. Within Li Auto, the topic surfaced repeatedly, yet the company did not settle on a direction.
Li Auto formed a 30-person team in 2021 focused on overseas marketing and strategic analysis, but by early 2022 the team was reassigned to domestic operations. Over the next two years, Li Auto continued to question when and how to approach overseas markets. CEO Li publicly stated in 2023 that the company would not enter global markets before 2025. In February 2024, Li Auto announced plans for a direct-sales model overseas, before shifting again in May to a hybrid model combining direct sales and dealer partnerships.
This year, Li Auto clarified its strategy. It will focus on the Middle East, Central Asia, and Europe. The company has taken concrete steps, including establishing R&amp;D centers in Germany and the US and opening retail locations in Uzbekistan and Kazakhstan. All new models launching in 2026 and beyond will comply with overseas regulatory requirements.
<h3>Doubling down on AI and chips</h3>
Like other ambitious automakers, Li Auto is deepening its investment in AI. During the strategy meeting, the company dedicated an entire day to AI and its future direction. Part of the discussion centered on AI and Li Auto’s second-generation chip. Lu Qi, head of the now-shuttered YC China, joined remotely.
According to sources, Li Auto has committed to increasing its investment in AI, especially in inference compute capacity. One insider said Lu argued, “In the future, the scarce resource will not be training compute, it will be inference compute.” According to the latest data, Li Auto currently has ten exaflops of training compute capacity and three exaflops of inference compute, with inference compute set to grow.
Sources also said Li Auto is rapidly iterating on its self-developed in-vehicle chips. The second-generation chip is planned for release in two years. Unlike <a href="https://kr-asia.com/nio-and-xpengs-in-house-chips-enter-tapeout-phase-paving-way-for-deployment">Xpeng’s Turing chip or Nio’s Shenji chip</a>, Li Auto’s second-generation chip will place greater emphasis on inference performance.
Li Auto’s first-generation chip was successfully taped out early this year and has returned from fabrication. It is now undergoing in-vehicle testing and is expected to be deployed in flagship models next year.
According to CTO Xie Yan, Li Auto has designed a new data flow architecture for its AI chips. In this design, model execution is driven largely by data instead of instructions, allowing higher parallelism and making it well-suited for large neural networks. “Compared with the most powerful chips on the market today, ours can deliver twice the performance for large language model workloads and triple the performance for vision models such as convolutional neural networks,” Xie claimed.
The meeting also concluded that Li Auto’s exploration of AI cannot remain limited to vehicle integration. It must invest substantially in robotics and terminal devices. Li Auto’s early approach focused on adapting AI technologies to individual business lines to find real-world use cases. At the 2023 autumn strategy meeting, Meituan’s Wang argued that combining large models with Li Auto’s in-car assistant would not unlock true AI potential.
Li Auto now believes it needs to go broader. At the 2024 autumn meeting, management discussed the necessity and feasibility of developing its own smart glasses. In June, the company established two new departments focused on “spatial robotics” and “wearable robotics.” To date, however, Li Auto has not released AI-driven hardware products outside the automotive category.
Li Auto’s overall pace remains in transition. But based on the reflections and recalibrations at its strategy meeting, the company is moving toward a faster and more agile product cycle and a more determined push into AI and global markets as it works to rebuild competitiveness for the next phase of industry competition.
KrASIA Connection features translated and adapted content that was originally published by 36Kr. This <a href="https://36kr.com/p/3559733792488584">article</a> was written by Xiao Man for 36Kr.

Li Auto enters crisis mode as sales falter and rivals pull ahead