Alibaba

ByteDance

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

Anker

DeepSeek

Li Auto

MiniMax

Pop Mart

Xpeng

On the Radar

Events

Latest

About

Consumer

Robotics

Fintech

Collaborate

On September 22, the markets were jolted by news that <a href="/company/nvidia">Nvidia</a> and <a href="/company/openai">OpenAI</a> had signed a letter of intent to form a partnership. According to reports, OpenAI will deploy at least ten gigawatts of data center capacity on Nvidia’s computing systems to train and run its next-generation models. To support that effort, Nvidia has pledged to invest up to USD 100 billion in OpenAI in phases.
The first round of funding is expected in the second half of 2026, after OpenAI completes its first gigawatt-scale data center. This facility will be built on Nvidia’s upcoming Vera Rubin platform. The deal raises key questions: how should the investment be interpreted, and what impact will it have on both Nvidia and OpenAI?
<h3>A capital game of chips for equity</h3>
This USD 100 billion investment is not a straightforward cash infusion. It resembles a structured exchange of computing power for equity. Based on disclosures, Nvidia’s phased investment will finance OpenAI’s data center expansion, with a combined capacity of about 10 GW. Each tranche will be tied to construction milestones, ensuring the funds go specifically to infrastructure.
In return, OpenAI is expected to channel the capital back into Nvidia’s ecosystem by purchasing four to five million GPUs and related systems built on the Vera Rubin platform.
The result is a closed capital loop: Nvidia invests in OpenAI, OpenAI buys Nvidia’s chips, and the money cycles back. It is less about one-time spending than about restructuring asset ownership.
For OpenAI, the arrangement secures access to Nvidia’s accelerators, ensures large-scale computing capacity, and deepens integration with Nvidia’s software ecosystem. For Nvidia, it provides years of substantial chip sales.
During an earnings call in August, Nvidia CEO Jensen Huang said that building a 1 GW data center costs USD 50–60 billion, with roughly USD 35 billion going to Nvidia’s hardware. By that calculation, OpenAI’s planned buildout would cost USD 500–600 billion, of which about USD 350 billion would flow to Nvidia.
Nvidia’s USD 100 billion investment is also expected to secure a 25% stake in OpenAI, based on the company’s reported valuation of USD 500 billion. That would give Nvidia both a guaranteed hardware revenue stream and exposure to OpenAI’s software income and longer-term developments in artificial general intelligence.
The deal also signals to the industry that Nvidia’s dominance in <a href="/industry/ai">artificial intelligence</a> infrastructure remains firm. It suggests that competition is expanding beyond chip sales to include funding, technology platforms, talent, and alliances.
<h3>Microsoft’s exclusive partnership under threat</h3>
Microsoft has been OpenAI’s closest ally since 2019, investing more than USD 13 billion in the company. In return, OpenAI made Microsoft Azure its exclusive cloud provider, channeling its computing needs into Azure’s data centers. This arrangement helped anchor Azure’s growth strategy on OpenAI’s development of large-scale models.
The partnership also gave Microsoft the right to integrate OpenAI’s technologies into many of its core products, including Office, Bing, and GitHub Copilot. That integration reshaped Microsoft’s software into adaptive, AI-driven services and strengthened its position in enterprise applications.
But the exclusivity of the pact has weakened. OpenAI’s cooperation with Oracle was an early sign, and Nvidia’s investment could accelerate the shift.
By funding OpenAI, Nvidia reduces the company’s dependence on Azure and strengthens its negotiating position. If OpenAI is no longer tied to Microsoft for compute, Microsoft’s influence over its roadmap diminishes. This could open the door for OpenAI to work with other tech companies, including Google and Amazon, altering the balance of competition in applied AI.
If that happens, Microsoft’s advantage in AI applications may narrow.
<h3>Nvidia’s revaluation moment</h3>
For Nvidia, the OpenAI deal could reshape how investors value the company.
Until now, Nvidia has been viewed primarily as a hardware company, subject to cyclical demand and large capital commitments. Software companies like OpenAI, by contrast, are valued more highly for their scalability, recurring revenue, and lower marginal costs once infrastructure is in place.
By taking a significant stake in OpenAI, Nvidia positions itself on both sides of AI’s evolution. It secures near-term chip revenue while gaining exposure to longer-term, higher-margin software growth.
That dual role could shift Nvidia’s valuation from being seen as a hardware supplier to an integrated AI ecosystem leader spanning hardware, software, and services.
There is precedent for such a shift. Apple’s valuation once rose and fell in line with iPhone sales cycles. As Apple expanded its software and services, investors began treating it as an ecosystem company, giving it steadier multiples and shielding it from hardware downturns.
Nvidia may now be entering a similar phase. If so, this deal could mark the start of a new valuation framework, driven not only by chips but also software, partnerships, and developments in artificial general intelligence.
KrASIA Connection features translated and adapted content that was originally published by 36Kr. This <a href="https://36kr.com/p/3480159679896448">article</a> was written by Ding Mao for 36Kr.

Nvidia to invest in OpenAI, tying GPUs to equity in landmark deal

Nvidia

OpenAI

Artificial Intelligence

Deep Tech

AI Infrastructure

… https://console.kr-asia.com/wp-content/uploads/2025/03/DJI-Osmo-Pocket-3-1024x1024.jpg 1024w, https://console.kr-asia.com/wp-content/uploads/2025/03/DJI-Osmo-Pocket-3-150x150.jpg 150w, https://console.kr-asia.com/wp-content/uploads/2025/03/DJI-Osmo-Pocket-3-768x768.jpg 768w, https://console.kr-asia.com/wp-content/uploads/2025/03/DJI-Osmo-Pocket-3-600x600.jpg 600w" sizes="(max-width: 480px) 100vw, 480px" /><figcaption id="caption-attachment-152155" class="wp-caption-text">Photo of the DJI Osmo Pocket 3. Photo source: DJI.</figcaption></figure>
It is a familiar playbook: strong brand equity, mastery of supply chains, and R&amp;D strength. The same strategy will likely appear soon in the panoramic drone market.
Multiple sources confirm DJI began developing a panoramic drone in 2022, refining the design through several iterations. The product has now entered mass production and is expected to debut around the start of the new year.
Leaked photos suggest a lightweight design with propeller guards similar to the Avata series, a panoramic lens on a rotating gimbal, best-in-class video resolution and frame rate, ten-bit D-Log M, and likely a one-inch sensor. It may feature DJI’s O4 transmission system for high-definition video at up to ten kilometers, plus stereo vision or even LiDAR (light detection and ranging) for obstacle avoidance.
Citing insider knowledge, 36Kr said DJI’s strategy is to maximize specs while keeping prices lower than any comparable product.
So why, if its product strategy is bold and unhesitating, does DJI stay muted in publicity?
<h3>Staying above the noise</h3>
A supply chain insider close to DJI says there are two reasons.
First, DJI’s style is to talk less and do more.
Second, it does not play along even as others chase clout by dragging its name. The view is that responding would only validate rivals far behind, creating a false impression of competitive equality.
For years, challengers across industries have followed this formula: benchmark against the leader. Every Android smartphone launch seemed to both salute and provoke Apple. By adding more cameras or sharper displays, rivals claimed to “beat” Apple. But Apple rarely responded, knowing engagement would only legitimize the comparison.
The same playbook appears everywhere. Pepsi defined itself against Coca-Cola as the “young” choice. New hair dryer brands rose by touting “Dyson-like performance at lower prices.”
But unlike colas or hair dryers, drones are orders of magnitude more complex. Panoramic drones especially cannot be faked with simple assembly or outsourced software.
Most panoramic shooting still relies on multi-lens stitching, plagued by seams, color mismatches, and lag. On the ground, consumers may tolerate it as “good enough.” In the air, under fast motion, these flaws cause distortion, crashes, and signal loss.
A panoramic drone is not just a panoramic camera strapped to propellers. Safety, stability, and image quality demand rigorous engineering.
Consider DJI’s stabilization systems. Strong winds, high altitudes, and rugged terrain routinely defeat traditional stabilization. DJI’s fusion of mechanical, optical, and electronic stabilization has become the gold standard, enabling everything from smooth hovering to coordinated drone shows.
Its propulsion systems undergo equally extreme testing: agricultural drones tested on vast farms for endurance, and heavy-duty models delivering 15 kilograms of supplies to Mount Everest in 12 minutes despite low oxygen, freezing temperatures, and fierce winds.
Then there is imaging, DJI’s strongest suit. Beyond lens assembly, it designs optics, tunes sensors, and engineers color science. Its custom CMOS improves low light and noise control, while its proprietary D-Log M gives professionals richer post-production options.
These capabilities come from years of teamwork among optical engineers, imaging algorithm specialists, and color scientists. And they are at a level far beyond what rivals can replicate with off-the-shelf components.
<figure id="attachment_156828" aria-describedby="caption-attachment-156828" style="width: 1920px" class="wp-caption alignnone"><img loading="lazy" decoding="async" class="wp-image-156828 size-full" src="https://console.kr-asia.com/wp-content/uploads/2025/08/Antigravity-A1-Cover.jpg" alt="" width="1920" height="1280" srcset="https://console.kr-asia.com/wp-content/uploads/2025/08/Antigravity-A1-Cover.jpg 1920w, https://console.kr-asia.com/wp-content/uploads/2025/08/Antigravity-A1-Cover-720x480.jpg 720w, https://console.kr-asia.com/wp-content/uploads/2025/08/Antigravity-A1-Cover-1024x683.jpg 1024w, https://console.kr-asia.com/wp-content/uploads/2025/08/Antigravity-A1-Cover-768x512.jpg 768w, https://console.kr-asia.com/wp-content/uploads/2025/08/Antigravity-A1-Cover-1536x1024.jpg 1536w" sizes="(max-width: 1920px) 100vw, 1920px" /><figcaption id="caption-attachment-156828" class="wp-caption-text">DJI’s closest rival in the drone segment is Insta360, also based in Shenzhen, China. The company recently <a href="https://kr-asia.com/shenzhen-breeds-a-new-rivalry-as-dji-and-insta360-clash-across-markets">entered the market</a> with its new Antigravity drone brand. Image source: Antigravity.</figcaption></figure>
<h3>Playing the long game</h3>
In many ways, DJI reflects the depth of China’s industrial ecosystem, and its journey is a parable about focus and discipline.
Its founding story is one of visionaries building a massive market from nothing, when few believed in consumer drones. Its evolution highlights China’s supply chain and R&amp;D advantage.
Despite foreign crackdowns and domestic competition, DJI’s moat has only widened through years of patient engineering, from flight control to drones, gimbals, and panoramic cameras.
Today, in an era of capital chasing hype and rivals leaning on flashy marketing, DJI stands apart like a calm chess player. It keeps its eyes fixed on the board—products and technology—while building an edge too deep to challenge. Each new product quietly sets the industry standard.
Some learn from DJI’s culture of innovation. Some chase it, discovering both their strengths and limits. Each sees in DJI the reflection of their own ambitions and anxieties.
But DJI itself is not playing for one move or one market. It is playing to define the game.
KrASIA Connection features translated and adapted content that was originally published by 36Kr. This <a href="https://36kr.com/p/3462149854041735">article</a> was written by Xiao Xi for 36Kr.

DJI eyes panoramic drones, extending its quiet dominance into a new frontier

East Asia

China

Drones

… results.
If Chinese AI companies can clearly show business value, winning clients becomes much easier. We’ve benefited from China’s industrial AI push. Our system shows value directly, so client acquisition hasn’t been difficult.
<h3>36Kr: Which industries or companies show the best results with your OS?</h3>
DZ: It’s universal, a general operating platform. We already have clients in metal smelting, steel, environmental management, renewable energy, and electronics manufacturing. The range is broad.
<h2>Competition builds confidence</h2>
<h3>36Kr: Do you think your business has a moat?</h3>
DZ: First, we understand industries deeply. Second, we understand model development just as deeply. Third, our technology and products are forward-looking. And finally, we execute well. It’s the combination of knowledge, design, and practical ability that sets us apart.
<h3>36Kr: Many companies claim to have industry know-how. What does that really mean?</h3>
DZ: At Huawei, I worked on more than 100 projects related to AI and industry. Even with traditional AI, we learned how companies operate, how executives think, what their data looks like, and how value chains function. That insight came from years of hands-on work.
So today, a few points matter:
<ul>
<li aria-level="1">Understanding how industries run and how decision-makers think.</li>
<li aria-level="1">Understanding AI’s current limits, and where it will be in six months, a year, or two. If you only design for today’s AI, you’ll soon have to rebuild.</li>
<li aria-level="1">Understanding costs, compute needs, business setups, and data sensitivity. It’s always a complex system.</li>
</ul>
<h3>36Kr: How do you forecast AI’s capabilities one or two years ahead?</h3>
DZ: We don’t need precise forecasts, just direction.
First, I don’t think large models will break the data barrier soon. Most natural data has already been consumed. Progress is more like standing on your own shoulders. So general capabilities will improve linearly, not exponentially.
Second, in specific scenarios, models will improve quickly in reasoning, self-correction, and reinforcement learning. That’s where real breakthroughs will be.
<h3>36Kr: How do these judgments shape AlignBase’s work?</h3>
DZ: It means we can’t rely on AI improvements alone. We need to solve problems through business workflows. That’s why our models are not just industry-specific, but DAG (directed acyclic graph)-based compositions and logic graphs. That’s the foundation of our full-factor modeling.
<h3>36Kr: In your two years of entrepreneurship, do you regret any technical decisions?</h3>
DZ: At 01.AI, the team was excellent. In mid-2024, we explored deep reasoning. The models demonstrated strong reasoning ability, but each inference took more than ten minutes, which we judged too slow for complex tasks.
Looking back, we could have invested more in chain-of-thought and reinforcement learning.
<h3>36Kr: What lesson does that hold for your current startup?</h3>
DZ: Tie tech closely to real and future scenarios. That’s the most direct measure of value. And judge more from technical principles, not just a handful of metrics.
<h3>36Kr: Any recent technical progress you find critical?</h3>
DZ: Not yet. I’m waiting for models that learn from feedback in real scenarios without relying on short-term memory. That’s possible, and we’re working on it too.
<h3>36Kr: Are you worried about competition?</h3>
DZ: Not at all. We’ve built knowledge, understanding, and technology over nearly a decade. Other teams may match us in some areas, but as a whole, we’re rare.
The market is also huge. China has more than 500,000 major industrial companies. Most have the ability and willingness to pursue intelligent transformation.
We’re the first in China to fully automate workflow modeling. I hope more companies join us. That will help traditional enterprises upgrade faster.
Competition, overall, is positive. More successful case studies build client confidence and make acquisition easier. If the market were small, one or two failures could kill it. But here, it’s large enough for many players to thrive.
KrASIA Connection features translated and adapted content that was originally published by 36Kr. This <a href="https://36kr.com/p/3458998949467527">article</a> was written by Zhou Xinyu for 36Kr.

Cutting months to a day: AlignBase taps AI to automate enterprise workflows

Enterprise Services

… competition, and of the company’s globalization, will be judged not only by technical breakthroughs but also by how well they translate into durable protections for consumers, institutions, and the financial systems that serve them.
This article was published in partnership with <a href="https://en.finvgroup.com/">FinVolution Group</a>.

A decade in the making: FinVolution’s competition spotlights visual deepfake detection

Chinese customs authorities seized 1.83 million counterfeit Labubu dolls from January to mid-August, as a country that has faced international criticism over counterfeits of foreign brands now scrambles to protect its newfound global hit.
On a recent day at a wholesale market in Zhejiang, a man touted counterfeit Labubu toys, promising prices as low as RMB 2.5 (USD 0.35) each if buying ten or more. Authentic Labubu dolls usually retail for about RMB 100 (USD 14).
"These are fake, so if I'm caught selling them, I'll get arrested," the man said.
Labubu first appeared in a picture book by Hong Kong-born illustrator Kasing Lung. In 2019, Chinese toy maker <a href="/company/pop-mart">Pop Mart</a> began selling Labubu stuffed toys and figurines under its The Monsters series.
The character became a worldwide sensation after Lisa, a member of K-pop group Blackpink, shared photographs of Labubu goods on Instagram. Soon, young people around the world were hanging the dolls from their bags, and other celebrities joined in.
The dolls are also wildly popular in China this year. Sales of "The Monsters" series in the January-June period jumped to RMB 4.8 billion yuan (USD 672 million), nearly eight times higher than a year earlier.
As Labubu's popularity has grown, so has the appearance of counterfeits. The high prices fetched by some Labubus on Chinese resale apps have fueled the rise of fakes.
Some Chinese consumers are outraged by the spread of fakes. Others say they are fine since the doll is only going to get dirty anyway hanging from their bag.
Authorities are cracking down.
"The trick to authenticating is to look at the teeth," a senior customs official said at a July press conference. "A real Labubu has nine teeth."
Counterfeit dolls typically have a different number of teeth because they are made using different molds, but fakes are now appearing with nine teeth, as the cat-and-mouse game continues.
From January to mid-August, Chinese authorities uncovered 237 cases of Labubu counterfeiting nationwide, seizing 1.83 million units intended for export to 61 countries and regions.
Pop Mart has begun working with customs authorities in the UK, France, and other jurisdictions to combat counterfeits and said 7.91 million fakes have been intercepted abroad.
In addition to attaching QR codes to products to verify authenticity, the company is increasing the production capacity at partner factories to boost the supply of genuine dolls. The increased supply has led to lower prices in the resale market, Chinese media report.
China's factories have long been criticized as a source of counterfeit and imitation goods. In 2018, during US President Donald Trump's first term, his administration imposed punitive tariffs on Chinese products, accusing the country of violating the IP rights of US companies. This helped spark a trade war between the world's two largest economies.
In a US-China trade agreement that took effect in 2020, Beijing promised to crack down harder on IP infringement. But in a report released in April of this year, the Office of the United States Trade Representative noted that serious concerns remain about China's commitment to IP protection.
China is now working to improve its legal and enforcement systems as it aims to become an IP powerhouse. Under the country's revised copyright law from 2021, violators are required to pay significant compensation.
The number of IP disputes has grown to around 500,000 per year, an increase from before 2018, according to the Supreme People's Court.
"Counterfeiting is no longer something that China can see as someone else's problem," a content industry insider said. China's soft power has grown thanks to global hits such as Ne Zha 2, the highest-grossing animated film of all time, and video game Black Myth: Wukong.
"In advanced technologies like <a href="/industry/ai">artificial intelligence</a>, we are seeing China set world-leading precedents," said Takashi Nomura, chief representative of the Shanghai office of law firm Nishimura &amp; Asahi.
In 2024, a court in Guangzhou ruled that a generative AI service provider was liable for copyright infringement. The AI was found to have created images resembling characters from Japan's Ultraman series, infringing on its creators' copyrights. The court ordered the AI service provider to pay RMB 10,000 (USD 1,400) in damages.
This article first appeared on <a href="https://asia.nikkei.com/business/media-entertainment/china-battles-labubu-fakes-in-role-reversal-on-ip-protection">Nikkei Asia</a>. It has been republished here as part of 36Kr’s ongoing partnership with <a href="https://www.kr-asia.com/nikkei-teams-up-with-36kr-in-asia-tech-news-coverage/">Nikkei</a>.

China battles Labubu fakes in role reversal on IP protection

Consumer Economy

… feeding polluted answers could erode trust in AI models themselves.
<h3>A game of trust</h3>
<blockquote>“Is GEO just a scam? Does it really have technical value?”</blockquote>
That is the question Lao Mo, founder of EchoSurge, often hears. In 2024, he and several veterans from Google and ByteDance started the company to build GEO tools.
SEO’s reputation for spam has spilled into GEO, which is often associated with questionable practices. Yet in Silicon Valley, interest is growing. Profound, a GEO startup, raised USD 35 million in a Series B funding round in August, reaching a nine-figure valuation in under a year. Sequoia Capital led that investment. Others, such as Scrunch, are also attracting attention.
Profound emphasizes tools over content mills. It scrapes user-AI chat logs, analyzes high-frequency prompts, and reverse-engineers the content models favor.
For companies, this reframes visibility in AI. The task is less about “buying traffic” and more about treating AI as a savvy assistant.
“In the past, posting your message was like putting it on the village noticeboard. Now the internet has given you a secretary,” Lao said. “You need to know how to talk to the secretary.”
That demands higher standards. “Old tricks like content stuffing are basically cheating,” he added. “Once models get serious about commercialization, spammers will be wiped out.”
Research already shows that AI-generated text performs worse than human-written content in both Google search and ChatGPT citations. The more AI-generated material a page contains, the lower it tends to rank.
Dayu’s Yuan has noticed the shift. While SEO once required little specialization, GEO now demands analysts who can study model behavior. His firm has begun hiring college graduates with stronger technical skills instead of generalists.
For brands, building visibility starts with basics: maintaining a professional website, publishing manuals and updates, and engaging with credible media.
If you don’t tell your story, someone else might, and that could be a competitor.
For now, GEO offers a window of opportunity, particularly for niche businesses that previously ignored online presence. Yuan cited one offline restaurant that quickly improved visibility after creating a website and posting quality content.
But the window is narrowing. Perplexity has announced that sponsored content will be clearly labeled, while ChatGPT and Doubao are testing product links. The gray zone is shrinking.
For everyday users, the safest approach is caution. Treat AI as an informed but imperfect guide: useful for reference, but not the final word. Choose reliable sources yourself, paste them into the model, and let it answer based only on that context.
Or as Lao put it:
<blockquote>“Taking the high road is actually the straightest path. Large models are double-edged swords. If your content is junk, it will eventually come back to harm you.”</blockquote>
In the end, this is not a contest of traffic. It is a contest of trust, and trust cannot be faked forever.
KrASIA Connection features translated and adapted content that was originally published by 36Kr. This <a href="https://36kr.com/p/3453363864819332">article</a> was written by Deng Yongyi for 36Kr.

GEO or gimmick? The murky business of buying visibility in AI search

<blockquote>China Recap is a weekly roundup tracking Chinese companies expanding abroad, covering market entries, funding rounds, product launches, and global partnerships.</blockquote>
China’s corporate globalization strategy is evolving fast. Industry giants are rewriting the global playbook, while a new generation of companies charts fresh paths overseas.
China Recap tracks both—focusing on strategic expansion, brand building, and localized operations—to help readers make sense of shifting trends and understand how Chinese firms are reshaping their global approach.
This edition tracks how Chinese <a href="/industry/ev">electric vehicle</a> makers are balancing overseas growth with domestic partnerships, while managing risks that test the industry’s resilience.
<h3>BYD cuts Qin Plus sedan prices in China</h3>
<a href="/company/byd">BYD</a> is offering RMB 10,000 (USD 1,370) discounts on select Qin Plus sedans through December, extending earlier cuts across models such as the Tang DM-i and Seal 07 DM-i. The Qin Plus remains one of BYD’s bestselling models, but the move follows four consecutive months of weaker sales in China amid official calls to curb price wars. —<a href="https://www.reuters.com/world/chinese-carmaker-byd-cuts-prices-some-models-until-year-end-2025-09-25/">Reuters</a>
<h3>Chery soars as Hong Kong IPO launches</h3>
<a href="/company/chery">Chery</a> shares jumped 11.2% in their Hong Kong debut, raising HKD 8.8 billion (USD 1.1 billion) in the IPO. The Anhui-based automaker canceled its listing ceremony due to typhoon disruptions and said it will exit Russia, its largest overseas market, by 2027. —<a href="https://asia.nikkei.com/business/markets/equities/chinese-state-carmaker-chery-jumps-11-in-hong-kong-debut">Nikkei Asia</a>
<h3>BYD ready if Nvidia chips are banned</h3>
BYD has a backup plan if access to <a href="/company/nvidia">Nvidia’s</a> automotive chips is cut, executive vice president Stella Li said, citing the company’s in-house tech and past resilience during chip shortages. While no restrictions exist, China’s stance on Nvidia remains uncertain. BYD currently uses Nvidia’s Orin system in its EVs. —<a href="https://www.cnbc.com/2025/09/23/chinas-byd-says-it-has-a-backup-plan-if-cut-off-from-nvidia-chips.html">CNBC</a>
<h3>Huawei, Dongfeng deepen EV alliance</h3>
<a href="/company/huawei">Huawei</a> and Dongfeng Motor are deepening their EV partnership following a leadership meeting in Shenzhen between Ren Zhengfei and Yang Qing. The firms will co-develop smart car tech, with all Voyah EVs this year to feature Huawei systems. A new joint lab aims to drive further innovation. —<a href="https://www.scmp.com/tech/big-tech/article/3326556/huaweis-ren-zhengfei-dongfeng-motors-yang-qing-discuss-deeper-collaboration-evs">SCMP</a>
whereRelatedArticle
<h3>Berkshire Hathaway exits BYD after 17 years</h3>
Warren Buffett’s Berkshire Hathaway sold its remaining shares in BYD, ending a 17-year investment that began with a 10% stake in 2008. Berkshire started reducing its holding in 2022 and cut it below 5% by mid-2024. BYD shares fell in Hong Kong and Shenzhen after the disclosure, as the company thanked Buffett and the late Charlie Munger. —<a href="https://asia.nikkei.com/business/markets/equities/buffett-s-berkshire-exits-chinese-ev-maker-byd">Nikkei Asia</a>
<h3>Li Auto, SEVB form EV battery JV</h3>
<a href="/company/li-auto">Li Auto</a> and Sunwoda’s SEVB arm have formed a joint venture in Shandong to produce EV batteries. The new entity will manufacture Li Auto’s self-developed battery designs. Unlike its partnership with <a href="/company/catl">Contemporary Amperex Technology</a> (CATL), Li Auto leads design, materials, and processes to retain brand control and manage costs.
<h3>Magna Steyr to build Xpeng SUVs in Austria</h3>
<a href="/company/xpeng">Xpeng Motors</a> has partnered with Austrian contract assembler Magna Steyr to produce its G6 and G9 SUVs in Graz. It is the first time Magna has produced complete vehicles for a Chinese automaker. Xpeng said the deal will create a European production base to bring its smart cars closer to local customers. —<a href="https://asia.nikkei.com/business/automobiles/electric-vehicles/china-s-xpeng-partners-with-magna-steyr-to-make-evs-in-europe2">Nikkei Asia</a>
<h3>Safety risk prompts Xiaomi SU7 recall</h3>
<a href="/company/xiaomi">Xiaomi</a> will recall about 110,000 SU7 cars built before August 30, after regulators flagged safety risks in the vehicle’s Level 2 highway navigation assist system. The recall, covering around 30% of all SU7 sales, will be addressed through over-the-air software updates. It is Xiaomi’s second recall and comes amid stricter oversight of driver assistance systems in China. —<a href="https://www.jiemian.com/article/13376114.html">Jiemian News</a>
That wraps up this edition of China Recap. If your company is expanding internationally, we’d love to hear about your latest milestones. <a href="mailto:changweiqian@global.36kr.com">Get in touch</a> to share your story.

China Recap | EV rush hour

China's Yadea Group Holdings, the world's largest manufacturer of electric two-wheelers, will enter the Japanese market in November, looking to challenge dominant brands like Honda Motor with a model around 30% cheaper.
"The Porta is compact and affordable," said Yasumasa Hasegawa, president of Hasegawa Mobility, Yadea's official Japan importer. "Its performance is suitable for everyday use."
The Yadea Porta is a low-output electric moped categorized as Class 1 in Japan, meaning it can be ridden with a standard driver's license. It can be charged using a household outlet, with a maximum range of 60 kilometers on a charge.
Its body features a retro design and color scheme to appeal to younger consumers. While some predicted a price in the JPY 100,000 (USD 680) range, it has been priced at JPY 217,800 (USD 1,480) after adjustments to comply with Japanese safety standards, including requirements for top speeds and turn signal positioning.
Yadea enters Japan while the government is tightening motorcycle exhaust regulations. The new regulations are on par with those in Europe and require future products to substantially reduce carbon monoxide and nitrogen oxide emissions.
As a result of the change, vehicle manufacturers will cease production of their current two-wheelers at the end of October. Though each company plans to release gasoline-powered models that comply with the regulations, only Yamaha Motor has indicated a clear release window, in the first half of 2026.
Honda and Yamaha sell Class 1 electric scooters, but their cheapest models cost around JPY 300,000 (USD 2,040), roughly JPY 100,000 more than their gas-powered models. Yadea is offering the Porta at the same price point as established manufacturers' gas-powered models.
Japan's domestic two-wheeler sales peaked in the early 1980s, topping three million units per year. Factors like the country's declining birth rate and aging population pushed sales down to about 320,000 units in 2024, one-tenth of the peak.
Major changes to the market like stricter exhaust regulations can present emerging players with opportunities. Yadea looks to expand Japan's nascent electric two-wheeler market.
Yadea was founded in China in 2001. China's two-wheeler market was once a stronghold of Japanese companies like Honda. But since the 2000s, there has been a shift to <a href="/industry/ev">electric vehicles</a> as a means of daily transportation, giving local makers an opportunity to leverage their low prices to control the market.
Yadea says its strength is in developing and manufacturing core components like batteries and motors in-house. It is known as the <a href="/company/byd">BYD</a> of two-wheelers, referring to the Chinese electric vehicle giant, and boasts a high share of China's under-RMB 5,000 (USD 700) electric two-wheeler market.
The company also is rushing to expand overseas, with sales networks in over 100 countries and regions.
"China's domestic market is sluggish due to deflation," said Hikaru Todoroki, principal at KPMG Consulting. "Two-wheelers are also in a state of overproduction, so companies probably want to sell their excess production overseas."
Looking at the global two-wheeler market, Japanese manufacturers Honda, Yamaha, Suzuki Motor, and Kawasaki Motors hold a combined share of over 40%, an overwhelming position built on their strengths in internal combustion engines.
But factors such as the push toward decarbonization are driving electrification. New two-wheeler sales totaled 59.9 million units worldwide in 2024, according to Tokyo-based Yano Research Institute, about 8.5% of which were electrified. In Yano's optimistic scenario, the electrification rate is projected to reach 18.8% by 2035.
New challengers are emerging as well, including India's Ola Electric Mobility, VinFast in Vietnam and Taiwan's Gogoro. Some models have ranges of more than 100 km, offering greater convenience.
"Unlike for four-wheeled vehicles, the market for internal combustion two-wheelers is still expanding," said Takaki Nakanishi of Nakanishi Research Institute. "Japanese manufacturers don't need to be at the forefront of electrification."
Issues such as high costs and underdeveloped charging infrastructure mean that electrified two-wheelers have not spread significantly outside of China and a few other regions.
But Japanese makers are not sitting idle. Honda plans to invest JPY 500 billion (USD 3.4 billion) in two-wheeler electrification by 2030. The company will launch 30 new models worldwide, targeting annual sales of four million units. The company aims for about one-fifth of its global sales to be electrified vehicles.
"It will take a bit more time before we can make a profit on electrics," Minoru Kato, the head of Honda's two-wheeled vehicle division, said at a briefing in Tokyo in January. For now, Honda plans to rely on internal combustion engine units while refining its electrification technology.
Yamaha also plans electrification for about 30% of the new two-wheeled vehicles it introduces over the three years from 2025. The company will electrify its existing models or introduce models from US startups in which it has invested.
This article first appeared on <a href="https://asia.nikkei.com/business/transportation/china-s-yadea-to-take-on-honda-in-japanese-two-wheeler-market">Nikkei Asia</a>. It has been republished here as part of 36Kr’s ongoing partnership with <a href="https://www.kr-asia.com/nikkei-teams-up-with-36kr-in-asia-tech-news-coverage/">Nikkei</a>.

China's Yadea to take on Honda in Japanese two-wheeler market

Transportation

Electric Vehicles

Mobility

Japan

Wondershare Technology, a leading creativity software company, has announced the launch of Recoverit V14. The latest version of its flagship data recovery solution offers broader device support, higher accuracy, and artificial intelligence-powered restoration capabilities. It provides professional-grade data recovery with a streamlined interface that requires minimal technical expertise.
Recoverit V14 marks a milestone in AI-powered data recovery and repair. It can recognize more than one million devices and covers over 10,000 recovery scenarios. Internal testing shows an industry-leading recovery success rate of 99.6%.
Key highlights of Recoverit V14 include:
<ul>
<li aria-level="1">Smarter recovery: AI-driven device recognition identifies hardware models and applies the most accurate recovery strategy. After retrieval, AI-powered restoration can be applied to enhance and repair corrupted images and videos for higher-quality results.</li>
<li aria-level="1">Expanded device and media support: Upgraded fragmentation restoration algorithms now support premium digital camera brands. The software is compatible with over one million devices from brands including Sony, Canon, Nikon, DJI, GoPro, Insta360, and more. It also supports encrypted portable drives from Western Digital, Seagate, and other manufacturers.</li>
<li aria-level="1">Comprehensive media and system compatibility: Camera recovery includes advanced support for professional video and image encoding and storage formats. SD card recovery offers improved compatibility across multiple speed classes, memory card specifications, and low-level file systems. Mac recovery provides dedicated support for ProRes, ProRes RAW, HEVC, HEIF, Apple ProRAW, and iCloud-native storage, with full compatibility for the upcoming macOS 26.</li>
</ul>
“Recoverit V14 sets a new benchmark for intelligent data recovery,” said Sam Zhang, vice president of Wondershare. “By combining AI-powered recognition and restoration with expanded device compatibility and system-level integration, we are enabling users to recover and enhance their most valuable files with greater ease and confidence than ever before.”
<h3>Availability</h3>
Recoverit V14 is available for Windows and macOS users, with flexible pricing options tailored to diverse needs. For more details, please visit the <a href="https://recoverit.wondershare.com/">official website</a>.
<h3>About Wondershare</h3>
Founded in 2003, Wondershare is a global software company known for its creativity and productivity tools. Driven by its mission “Creativity Simplified,” the company offers a range of products including Filmora, Virbo, and SelfyzAI for video editing; PDFelement for document management; and EdrawMax and EdrawMind for diagramming.
With a presence in more than 200 countries and regions, Wondershare provides intuitive software and creative resources to empower the next generation of creators, and broaden the possibilities of creativity worldwide.
This press release was published in partnership with <a href="http://wondershare.com/">Wondershare</a>.

Wondershare launches Recoverit V14 with AI-powered recovery, 99.6% success rate, and support for over a million devices