Alibaba

ByteDance

CATL

Huawei

Meituan

Tencent

Xiaomi

Most Followed

Market News

Listings (IPO)

Earnings Reports

Venture Roundup

Big Tech

Automotives

Retail and Commerce

Collections

News

Features

Insights

Press Releases

Anta

DeepSeek

Geely

Li Auto

Mixue Bingcheng

Pop Mart

Going Global

Latest

Collaborate

Earnings

Auto

Retail & Commerce

… navigation, obstacle avoidance, manipulation, stair climbing, and other behaviors with whole-body control. Cosa sits at the center, responsible for task comprehension, planning, and orchestration.
This division of labor reduces development friction. Engineers can avoid rebuilding hardware or low-level motion systems from scratch and instead build on LimX’s validated platforms and integrated cognitive-motor stack, using open interfaces and development tools to focus on application layer innovation within specific verticals.
More broadly, this approach represents a shift in LimX’s position within the robotics ecosystem. Rather than competing across every layer, the company is defining clear technical and market boundaries.
As robotic bodies and cognitive systems become increasingly standardized, the company’s challenge moves toward deploying these systems in environments that demand adaptability and close human interaction.
Guided by the principle of serving people rather than production processes, LimX is prioritizing scenarios that require direct interaction and contextual flexibility. Initial deployments are focused on business-facing service environments such as reception and guided tours, with plans to expand gradually into consumer household services.
Internationally, the company plans to pair China’s supply chain advantages in rapid iteration and cost control with localized ecosystems of developers, customers, and distribution partners. The aim is to embed embodied intelligence more deeply into overseas markets.
KrASIA Connection features translated and adapted content that was originally published by 36Kr. This <a href="https://36kr.com/p/3664227663651718">article</a> was written by Huang Nan for 36Kr.

LimX Dynamics raises Series B funding to push embodied robots into real-world use

Venture Capital

Robotics

Deep Tech

Embodied Intelligence

China

… knit sets sold more than 1,000 units on TikTok Shop. Seeing the platform’s potential, Guo formed a small team to build a direct-to-consumer presence.
By studying content trends, the team learned that Western shoppers favored bright, high-saturation resort colors. Ruimin expanded its dress line from four colors to 16 and added elastic bands at the bust for comfort.
The redesigned dresses, now under the Finjani label, were promoted through TikTok advertising and influencer partnerships. Within five months, a single viral dress sold nearly 120,000 units, and annual store revenue surpassed RMB 23 million (USD 3.2 million).
Today, garments shipped from Ruimin’s Dongguan factory to Europe and North America carry the Finjani tag, visible across live streams and short videos.
“A person needs a name,” Guo said. “So do the clothes we make.”
<h3>When visibility becomes influence</h3>
For Creality, already known among 3D printing enthusiasts, the challenge was reaching a broader audience.
When Shi took over TikTok Shop operations in March 2025, he was working alone and had two weeks to launch a new product. He partnered with a US-based influencer using the handle @is3dp, a creator with fewer than 15,000 followers but deep technical expertise.
Because Creality’s printers often cost hundreds of dollars, buyers needed reassurance. Shi turned the live stream into a virtual showroom, staying awake in Shenzhen until as late as 4 a.m. to answer real-time questions on materials and compatibility. The transparent format helped ease hesitation.
One April night, the live stream generated USD 40,000 in orders.
Over time, Shi expanded collaborations with US creators, refined targeting, and experimented with formats. Creality’s TikTok Shop GMV (gross merchandise value) reached USD 400,000, while monthly consumables sales exceeded 10,000 units.
<div style="width: 100%; display: flex; justify-content: center; align-items: center; padding: 20px 49px;">
<blockquote class="tiktok-embed" style="border-left: none; margin: 0; padding: 0; width: 100%; max-width: 605px;" data-video-id="7495045357819235615"><script async src="https://www.tiktok.com/embed.js"></script></blockquote>
</div>
For established brands, TikTok Shop offers a stage to shape identity, build relevance, and occasionally spark cultural moments.
The global rise of Pop Mart’s Labubu illustrates that potential. Hail a ride in the US today, and there is a good chance the driver has a Labubu figurine hanging from the dashboard.
The surge can be traced back to November 2024, when Vanity Fair posted a TikTok video of K-pop star Lisa showing off her collection. Interest spiked. When Pop Mart launched the Labubu “Big into Energy” series on TikTok Shop in April 2025, nearly 100,000 viewers tuned in to the live stream, and all 20,000 figurines sold out within seconds.
Labubu-tagged videos now exceed 3.5 million on TikTok, with more than one billion cumulative views. On TikTok Shop’s US page, Pop Mart’s total sales surpass 2.5 million units.
From Guo’s first Finjani tag in a Dongguan factory to Shi’s late-night data monitoring and Labubu’s viral rise, TikTok Shop has played a role at every stage. It gives unbranded products an identity, turns niche items into mainstream hits, and helps established brands reach new cultural relevance.
TikTok Shop’s near USD 100 billion GMV in 2025 underscores how content-driven commerce has reshaped global retail.
For decades, Chinese manufacturers operated behind the scenes of global supply chains, distant from end consumers. TikTok Shop offers a direct line outward, allowing them to sense demand, iterate, and build brands.
When businesses such as OQ Hair and Finjani, once anonymous suppliers, can read global consumer signals in near real time and respond quickly, perhaps, it points to a deeper shift underway.

How TikTok Shop gives China’s factories a direct line to the world

TikTok

East Asia

TikTok Shop

Retail

Consumer Economy

… directly within Kimi, a capability that depends heavily on model performance.
In early December, Moonshot president Zhang Yutong said the company needed to find its own narrative and identify what it truly excels at.
“When competing with much larger companies that have more resources, we deliberately control our business boundaries,” Zhang said. “We focus on the large-model layer, the logic layer, the agent layer, and productivity-oriented, complex task workflows such as in-depth research, presentations, data analysis, and website development.”
Across the sector, model developers are converging on similar agent-driven scenarios, particularly in coding and office productivity. These use cases offer relatively clear paths to commercialization but also depend on continued gains in model capability.
For Moonshot, the stated ambition is to build a “platform agent.” Achieving that will require sustaining top-tier model performance while also developing consumer products with a distinct identity and consistent design sensibility.
KrASIA Connection features translated and adapted content that was originally published by 36Kr. This <a href="https://36kr.com/p/3664711869850499">article</a> was written by Deng Yongyi for 36Kr.

Moonshot AI sees overseas revenue surge as Kimi K2.5 gains traction abroad

Artificial Intelligence

Foundation and Large Models

India's exports to China have shot up in recent months, mostly thanks to marine products and electronics, as bilateral ties thaw and New Delhi's diversification strategy to offset the impact of steep US tariffs takes effect.
Nevertheless, experts see China's ballooning trade surplus with India as a matter of concern.
In the first nine months of the current fiscal year, which began last April, India’s merchandise exports to China rose 36.7% from the same period a year earlier to USD 14.25 billion, according to data from India’s Ministry of Commerce and Industry. In December alone, these shipments increased 68% to USD 2.05 billion.
"[It] is a welcome growth," commerce secretary Rajesh Agrawal told reporters earlier this month, adding that the "overall recalibration of supply chains across the world is helping" the South Asian nation boost its exports to the communist giant.
The jump in exports to China comes with US-Indian ties fraying. President Donald Trump in August imposed steep 50% tariffs on Indian goods, including a 25% penalty for India's purchases of Russian oil. India has a significant merchandise trade surplus with the US, by far its largest export destination.
As for the marine products India is shipping to China, it sent USD 1.056 billion worth in the April-November period, 20.4% more than in the same period a year earlier, according to ministry data.
A separate December 2025 report by the New Delhi-based think tank Global Trade Research Initiative said the surge in India's exports to China is driven by a narrow set of products. GTRI said naphtha is the biggest contributor, with exports in the April-October period up 172% from a year earlier to USD 1.4 billion, reflecting strong Chinese demand for petrochemical feedstocks.
Electronics also showed "unusually sharp spikes" with printed circuit board exports jumping over 2,000% to USD 418 million in the first seven months of the ongoing financial year, the report said. Mobile phone component exports rose 82% to USD 362 million, "an unusual trend given India's large imports of these items from China," it said.
As India seeks to diversify its export markets, it has stepped up efforts to strike trade deals with several countries in a bid to minimize the impact of the US tariffs. Last month, it concluded what it called a "free trade agreement" with New Zealand, close on the heels of signing a comprehensive economic partnership agreement with Oman during Prime Minister Narendra Modi's December visit to the Gulf nation and months after inking a trade deal with the UK.
In addition, India and the European Union are reportedly close to sealing a major trade pact, and New Delhi and Washington continue to negotiate a deal of their own.
S C Ralhan, president of the Federation of Indian Export Organisations (FIEO), said the country's top export destinations demonstrate "a well-diversified and resilient" export footprint.
"This diversification is particularly critical at a time when global trade routes are being reshaped due to geopolitical conflicts, sanctions, shipping disruptions and strategic realignments," he said in a recent statement issued by FIEO.
Regarding the rise of exports to China, Namrata Hasija, a research fellow at the New Delhi-based Centre for China Analysis and Strategy, said they are "not broad-based" and are heavily driven by specific commodities, an indication that the growth is "more reactive to immediate Chinese demand rather than a long-term, structural realignment."
"The issue is also that China is OK with buying these items from India now," she told Nikkei Asia. But "what if the state policy changes and they too tighten the screws" on shipments from India?
N R Bhanumurthy, director of the Madras School of Economics in Chennai, said that right now exports to China "may be a little more competitive" for Indian industry because of the Trump tariffs. Another reason for the surge in exports could be that "maybe China is looking at the Indian market for [procuring] some of the things they want to withdraw from … especially the production of low-value products," he told Nikkei.
However, Bhanumurthy added that India "running an increasingly wider trade deficit with China is definitely a [matter of] concern" and added that it needs to enhance its business relationship with the neighboring country in a way that addresses the trade gap.
In the April-December period, China's exports to India were worth USD 95.95 billion, up from USD 84.57 billion a year earlier. India's key imports from China include electronics, machinery, chemicals, and plastics.
Over the same period, India’s trade deficit with China widened to USD 81.7 billion from USD 74.15 billion a year earlier.
"India's recent export gains to China are narrow, volatile and heavily dependent on shifts in Chinese demand, rather than on durable market access or a diversified export base," Ajay Srivastava, GTRI's founder, said in the report. "Without a sustained strategy to expand competitive manufacturing, reduce import dependence in key sectors, and strengthen trade monitoring, short-term export spikes will do little to alter the fundamentally imbalanced nature of India-China trade."
This article first appeared on <a href="https://asia.nikkei.com/economy/trade/china-offers-indian-exporters-helping-hand-amid-trump-tariffs">Nikkei Asia</a>. It has been republished here as part of 36Kr’s ongoing partnership with <a href="https://www.0722vps.aieehr01.top/nikkei-teams-up-with-36kr-in-asia-tech-news-coverage/">Nikkei</a>.

China offers Indian exporters helping hand amid Trump tariffs

India

… ambitions:
<blockquote>“Looking ahead to the next decade, the company will actively transform… to create a new era of perception and interaction between cars and robots.”</blockquote>
Sharpa’s founding suggests Li and his co-founders were not content to remain component suppliers. Their backgrounds help explain that shift. Li studied robotics at Tsinghua University and earned a PhD in motion control from the University of Illinois at Urbana-Champaign. Sun holds a PhD in mechanical engineering from Stanford University. Xiang, also a Tsinghua alumnus, earned dual master’s degrees in electrical and mechanical engineering from Stanford and previously worked as an iPhone systems integration engineer.
From robotics research and mechanical design to real-world deployment, the trio has spent years bridging academic work and industrial engineering.
They are not alone in crossing over. Teams from the autonomous driving sector have increasingly moved into embodied intelligence, spawning startups such as Tars, founded by former Huawei scientist Chen Yilun. The company reportedly raised RMB 1.7 billion (USD 238 million) in its angel funding round.
“The common thread between automotive LiDAR and embodied intelligence is spatial perception,” said one financial adviser in the field. That overlap, he said, is why many investors suspect a close link between Sharpa and Hesai.
Sharpa is headquartered in Singapore and operates an office in Shanghai, Hesai’s home base. The setup allows it to manage global partnerships while remaining close to China’s manufacturing and supply chain ecosystem.
But one question lingers: why incorporate in Singapore and operate with such limited visibility?
The answer may trace back to a turning point in Hesai’s international expansion. Before and after its US listing in 2023, Li said in interviews that Hesai positioned itself as a global company, with the US accounting for a significant share of its business, a profile that appealed to overseas investors.
That narrative shifted in January 2024, when the US Department of Defense added Hesai to its list of “Chinese military companies,” triggering a selloff in its shares. Hesai filed a lawsuit challenging the designation but lost the case in the US District Court for the District of Columbia in July 2025. The issue has remained unresolved since.
KrASIA Connection features translated and adapted content that was originally published by 36Kr. This <a href="https://36kr.com/p/3647460169092745">article</a> was written by Fu Chong for 36Kr.

Singapore-based Sharpa surfaces as robotics venture linked to Hesai founders

Singapore

Southeast Asia

Shanghai-based chipmaker Iluvatar CoreX has unveiled a multi-year GPU architecture roadmap that lays out when it expects to surpass <a href="/company/nvidia">Nvidia’s</a> platforms, a rare level of specificity in the semiconductor sector.
Rather than leaning on conventional benchmark comparisons, Iluvatar has tied its ambitions to a public timeline, inviting direct comparison with Nvidia’s generational cadence.
According to the roadmap, Iluvatar said its current Tianshu architecture has already met the first target. Citing data reported by <a href="https://36kr.com/p/3656157912817797">36Kr</a>, the company said that in workloads related to <a href="/company/deepseek">DeepSeek’s</a> V3 large model, Tianshu delivered about 20% higher average performance than Nvidia’s Hopper platform, achieving more than 90% utilization of computing power. The company described these results as measured performance in real-world workloads rather than peak theoretical figures.
The roadmap extends this trajectory through successive designs. Following Tianshu, the Tianxuan architecture is intended to compete with Nvidia’s upcoming Blackwell generation, while Tianji is positioned to surpass it in 2026. Tianquan, the fourth architecture in sequence, is projected to outperform Nvidia’s Rubin platform in 2027. Beyond that, Iluvatar said it plans to move toward what it called a “breakthrough” architectural design, according to <a href="https://www.scmp.com/tech/big-tech/article/3341368/iluvatar-corex-targets-nvidias-rubin-gpu-road-map-amid-china-chip-push">SCMP</a>.
All four architectures are named after Chinese terms associated with the Big Dipper constellation. Iluvatar said it will release multiple products under each generation over the coming years.
Alongside its roadmap, the company highlighted commercial milestones to reinforce its claims. By June 2025, it reported delivering more than 52,000 general-purpose GPUs (GPGPUs) to over 290 customers across industries. Company disclosures indicated that its 2024 shipments <a href="https://0722vps.aieehr01.top/iluvatar-corex-becomes-first-chinese-gpgpu-firm-to-gain-hkex-listing-approval">ranked third</a> among domestic GPGPU suppliers in China. Iluvatar has also expanded into edge computing, introducing four products under its Tongyang (TY) series with performance ranging 100–300 TOPS (tera operations per second). The company said its entry-level TY1000 outperformed Nvidia’s AGX Orin in several testing scenarios.
The announcement follows Iluvatar’s Hong Kong IPO on January 8. Its shares recently closed at HKD 179.2 (USD 23), just below the debut price of HKD 190.2 (USD 24.4), giving it a market capitalization of about HKD 45.5 billion (USD 5.8 billion).
According to its prospectus, Iluvatar’s revenue rose from RMB 189 million (USD 26.5 million) in 2022 to RMB 540 million (USD 75.6 million) in 2024, though net losses also widened over the same period. The figures reflect the challenge of balancing expansion and profitability faced by many chipmakers.
The timing of the roadmap coincides with ongoing uncertainty over Nvidia’s ability to sell its most advanced GPUs in China, a backdrop that has encouraged domestic suppliers to position themselves as alternatives. After years of emphasizing local breakthroughs, mass production, and deployment, Iluvatar is now anchoring its ambitions to a public schedule, raising the stakes on execution.

Iluvatar CoreX lays out GPU roadmap to surpass Nvidia by 2027

Nvidia

Semiconductors

Jingdong Property, referred to hereafter as JD Property, has filed a fresh application for an IPO in Hong Kong, reviving a listing plan that stalled in 2023 while awaiting regulatory approval, according to <a href="https://www.reuters.com/world/asia-pacific/jdcoms-jingdong-property-files-hong-kong-ipo-2026-01-27/">Reuters</a>.
The logistics and industrial real estate manager did not disclose the size or timing of the proposed offering. If completed, IPO proceeds would be used to expand its overseas infrastructure asset network, deepen its presence in key Chinese cities, upgrade services, and fund general corporate purposes, according to its filing.
JD Property is majority owned by <a href="/company/jd">JD.com</a>, which holds about 75% of the company and would remain its controlling shareholder following the proposed listing.
The renewed IPO attempt comes as JD Property reports improved operating performance. Revenue for the first nine months of 2025 rose 21% year-on-year to about RMB 30 billion (USD 4.2 billion), while net losses narrowed sharply from the same period a year earlier.
As of September 30, 2025, JD Property managed around 270 infrastructure projects in China and overseas. Assets outside China accounted for about 12.8% of total assets under management, according to its prospectus.
The filing identifies overseas expansion as a stated use of IPO proceeds, although it does not name specific markets or projects. Public disclosures over recent years show the company has added logistics and industrial assets across the Asia Pacific, Middle East, and Europe.
In the Middle East, JD Property has invested in logistics facilities in Saudi Arabia and the UAE, including projects in Riyadh and Dubai, as well as partnerships with local authorities. In Europe, it expanded its UK footprint in December 2025 with the acquisition of logistics assets near Leicester, following an earlier investment in Milton Keynes in 2022.
The company has said these projects are intended to support demand for modern, high-standard warehousing and logistics infrastructure across multiple sectors.
Singapore is also part of JD Property’s overseas portfolio. In August, <a href="https://www.reuters.com/business/retail-consumer/china-tech-giant-jdcom-unit-two-other-firms-plan-1-billion-singapore-reit-2025-08-27/">Reuters</a> reported that the company was working with EZA Hill Property Management and Partners Group on a potential Singapore-listed industrial REIT (real estate investment trust).
The proposed trust was reported to be valued at more than USD 1 billion and could be listed as early as this year, subject to market conditions and final asset composition. JD Property has not publicly confirmed a listing timeline.
Since 2023, JD Property and EZA Hill have acquired nine industrial properties in Singapore. The most recent transaction closed last month, when the partners bought a four-property logistics portfolio for SGD 306 million (USD 242.5 million). The assets were aligned with a disposal by CapitaLand Ascendas REIT.
The renewed IPO effort follows other capital markets activity within the JD.com group. Last month, Jingdong Industrials, also referred to as JD Industrials, completed its Hong Kong listing, raising about HKD 2.8 billion (USD 358.9 million).
JD.com has previously listed logistics-related assets through real estate investment trusts in mainland China, according to public disclosures.
JD Property said details of the Hong Kong offering, including pricing and timing, would remain subject to regulatory review and market conditions.

JD Property revives Hong Kong IPO plan as overseas portfolio grows

JD.com

… natural-language interaction could reshape hardware terminals into new traffic gateways.
It remains a non-consensus frontier, but one that StepFun has already begun to explore through customer collaboration and hardware experimentation.
KrASIA Connection features translated and adapted content that was originally published by 36Kr. This <a href="https://36kr.com/p/3656118724010374">article</a> was written by Xiao Xi for 36Kr.

As AI consolidates, what makes StepFun worth a RMB 5 billion raise?

Chinese researchers and companies were front and center at NeurIPS in San Diego, one of the world's top three international conferences on <a href="/industry/ai">artificial intelligence</a>, underscoring the deep ties between research communities in the US and China, along with the political risks those connections now face.
A graduate student from the University of Toronto researching ways to use AI in biology visited <a href="/company/bytedance">ByteDance's</a> booth at the recent event, saying it is "one of a few companies that works on using AI for accelerating science discovery."
ByteDance was a top-tier sponsor and had a recruiting booth at the event, which has become a place for AI companies to scout promising talent.
While ByteDance is best known to consumers as the company behind TikTok, its "AI for Science" program engages in a broad range of basic research, including developing foundation models for biological and materials sciences. It has offices in San Jose, California; and Seattle, Washington; as well as Beijing, and recruiters at the event spoke in both English and Chinese.
The 2025 NeurIPS conference had roughly 25,000 attendees, more than six times the 2015 figure. The number of papers submitted jumped 38% on the year to 21,575, with 5,290 accepted. Chinese institutions were among the biggest sources of papers.
The top affiliation among authors of papers accepted by NeurIPS in 2025 was Tsinghua University, according to data from AI World, which is affiliated with a European think tank. While Google ranked second, Peking University and the University of Chinese Academy of Sciences took the next two spots.
One of the top papers was led by a researcher from the team at <a href="/company/alibaba">Alibaba Group Holding</a> developing the Qwen family of large language models.
US companies and institutions rely heavily on talent from China. According to a study by the Paulson Institute, 38% of top AI researchers at US organizations in 2022 had graduated from Chinese universities. Many moved to the US for advanced education and were then hired by American companies. <a href="/company/openai">OpenAI</a>, Google, and Meta have many Chinese employees working on AI development.
It is "honestly quite surprising" how well the US has done in retaining this talent, said Matt Sheehan, a senior fellow at the Carnegie Endowment for International Peace.
Sheehan followed the careers of 100 Chinese researchers affiliated with US organizations who had papers accepted at NeurIPS in 2019. As of 2025, 87 were still in the US, while ten had returned to China and three had moved elsewhere. This speaks to the US's "staying power as a magnet for top-tier researchers," Sheehan said.
But as tensions between the US and China over AI escalate, Chinese students and researchers are in an increasingly unstable position. US President Donald Trump's administration has tightened visa issuance for immigrants and international students, and begun revoking visas of Chinese students with ties to China's government.
A Chinese doctoral student attending the University of Texas at Austin expressed concern. "O-1 visa holders should be OK, but I'm not sure what Trump" will do, the student said, referring to visas for those with extraordinary ability.
A Chinese student from the University of Florida, who is researching advanced sensors with military applications, is debating whether to seek permanent residency and get a job at a US defense technology company. "I feel the atmosphere toward China has become tense," the student said.
Distrust of the US government is leading to uncertainty about working at US companies. "I can work on the frontier of AI development back in China, too," said a Chinese student at Princeton considering taking a job at ByteDance.
Companies such as Microsoft have AI research facilities in Beijing. But "I heard a rumor that you cannot use the same cloud services as in the US," the student said. "If so, it makes more sense to go to Chinese companies right away."
Chinese players see the situation as an opportunity to bring back talented Chinese graduates of US universities. And ByteDance is not alone. At NeurIPS, <a href="/company/ant-group">Ant Group</a>, Taobao, short-form video platform <a href="/company/kuaishou">Kuaishou Technology</a>, and others all had corporate booths, looking to recruit returnees.
"Five or ten years ago, if you wanted to do the best AI research, you had to come to the United States," Sheehan said. "There were some opportunities in China, but there were very few opportunities to work at the forefront of AI in China. In 2025, there's lots of opportunities to work at the forefront of AI in China."
The University of Texas doctoral student said: "From what I see among the AI researchers around me, 70% stay in the US, while 30% return to China."
Returnees from the US are already leading AI development efforts in China. Yang Zhilin, CEO of "AI tiger" Moonshot AI, earned his doctorate at Carnegie Mellon University before returning to China to found the company.
More recently, Yao Shunyu, a top researcher at OpenAI, moved to <a href="/company/tencent">Tencent Holdings</a>, where he was appointed chief AI scientist.
"We cannot know what percentage of people are doing economic espionage," Sheehan said. "But when you just look at the sheer size of the contribution [from Chinese researchers], it's very difficult to imagine that anything that is taken away or stolen from the US could match the size of the contribution. We're talking about one-quarter to one-third of the US AI research being done by these Chinese researchers."
The government should make it easier for Chinese nationals with doctorates to work in the US, he said.
This article first appeared on <a href="https://asia.nikkei.com/business/technology/artificial-intelligence/chinese-ai-researchers-in-us-torn-between-promise-politics">Nikkei Asia</a>. It has been republished here as part of 36Kr’s ongoing partnership with <a href="https://www.0722vps.aieehr01.top/nikkei-teams-up-with-36kr-in-asia-tech-news-coverage/">Nikkei</a>.

Chinese AI researchers in US torn between promise, politics