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

… to be opening up, according to their operators.
At its earnings briefing earlier this month, SGX revealed a slate of about 50 companies at "various stages of engagement and preparation" for IPOs, without naming them. A third of this batch are in tech, advanced manufacturing, and digital infrastructure; another third in the consumer and healthcare sectors and the rest include players in real estate, the exchange said.
"A year ago we said, at least 30 [IPOs], right? And I can tell you it's comfortably above that now," SGX's Syn told Nikkei Asia. "It's not just a list on an Excel sheet. It's people who paid money to professionals to start work [toward a listing], and that is what we call a pipeline."
SGX also had a major listing last month. Medical services provider <a title="" href="https://asia.nikkei.com/business/markets/ipo/singapore-hospital-group-makes-market-debut-as-population-ages" target="_blank" rel="noopener">Foundation Healthcare Holdings</a> raised SGD 242 million (USD 190.5 million) through its IPO.
The Thai bourse has disclosed a near-term pipeline of ten companies for IPOs, five of which have already been approved and five that are under review. The exchange expects the total number of new listings across its main and secondary boards this year to at least match the 18 reached last year.
"From our perspective, what we can control is to make sure that our regulations are competitive," SET's Asadej said. "Often, if the businesses are based in Thailand and revenues are based in Thailand, in the longer run, it makes more sense for them to list in Thailand."
This article first appeared on <a href="https://asia.nikkei.com/business/markets/trading-asia/thailand-singapore-exchanges-seek-tech-listings-as-ai-booms">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>.
Note: SGD, THB figures are converted to USD at rates of SGD 1.27 = USD 1 and THB 32.70 = USD 1 based on estimates as of August 25, 2026, unless otherwise stated. USD conversions are presented for ease of reference and may not fully match prevailing exchange rates.

Thailand, Singapore exchanges seek tech listings as AI booms

Investment

Singapore

Thailand

Southeast Asia

… existing ones. Unitree was 8,000 times oversubscribed by retail investors. That demand exists. Perpetuals just give international capital a way to participate in a mania that's already happening."
This article first appeared on <a href="https://asia.nikkei.com/business/markets/global-investors-use-perp-bets-to-chase-china-tech-stocks-like-unitree">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>.

Global investors use "perp" bets to chase China tech stocks like Unitree

China

… src="https://console.kr-asia.com/wp-content/uploads/2026/08/Galbot-Wang-He-World-Robot-Conference-2026-Presentation.jpg" alt="" width="1920" height="1080" srcset="https://console.kr-asia.com/wp-content/uploads/2026/08/Galbot-Wang-He-World-Robot-Conference-2026-Presentation.jpg 1920w, https://console.kr-asia.com/wp-content/uploads/2026/08/Galbot-Wang-He-World-Robot-Conference-2026-Presentation-720x405.jpg 720w, https://console.kr-asia.com/wp-content/uploads/2026/08/Galbot-Wang-He-World-Robot-Conference-2026-Presentation-1024x576.jpg 1024w, https://console.kr-asia.com/wp-content/uploads/2026/08/Galbot-Wang-He-World-Robot-Conference-2026-Presentation-768x432.jpg 768w, https://console.kr-asia.com/wp-content/uploads/2026/08/Galbot-Wang-He-World-Robot-Conference-2026-Presentation-1536x864.jpg 1536w, https://console.kr-asia.com/wp-content/uploads/2026/08/Galbot-Wang-He-World-Robot-Conference-2026-Presentation-480x270.jpg 480w" sizes="(max-width: 1920px) 100vw, 1920px" /><figcaption id="caption-attachment-163828" class="wp-caption-text">Photo source: Galbot.</figcaption></figure>
The second condition may prove just as important as the first.
Even when a model can perform much of a task, customers may still need algorithm engineers to collect robot motion data, label it, and debug the resulting behavior. That adaptation cost can make deployment difficult for smaller businesses and individual users.
For embodied intelligence to reach a mass market, teaching a robot also has to become substantially easier.
During his presentation, Wang introduced WAM-TTT, a test-time training approach intended to address that problem.
Under the approach, a user wears a first-person camera and records themselves doing a job. The model can use that unlabeled video to adapt to the task without requiring a fresh round of robot motion data collection, according to Galbot.
The foundation model supplies general knowledge about the physical world. A person demonstrates the specific work to be done.
If that process becomes reliable, adapting a robot could begin to look less like an engineering project and more like teaching a new worker.
At the end of his presentation, Wang outlined a longer-term vision in which robots could eventually approach smartphone-scale shipment volumes while retaining product values closer to automobiles and the continuous upgradeability of software.
A model improvement could then be distributed across a large installed base of robots.
The scale of that opportunity remains hypothetical, but the underlying problem is already clear.
Humanity spent decades generating the language, images, and other data that now train models for the digital world. Comparable training data for the physical world is far less abundant.
Robots will have to help generate it themselves by identifying products in pharmacies, responding to disruptions in factories, manipulating unfamiliar objects, and learning jobs from human demonstrations.
Each deployment can become another source of experience for the model.
The ChatGPT moment for embodied intelligence, by Wang’s definition, will arrive only when that accumulated intelligence can transfer across different robot bodies and when teaching a robot a new skill becomes simple enough for ordinary people to do.
KrASIA features translated and adapted content that was originally published by 36Kr. This <a href="https://36kr.com/p/3946454123822213">article</a> was written by Xiao Xi for 36Kr.

Is the frontier of humanoid robotics shifting from hardware to intelligence?

Robotics

Deep Tech

Embodied Intelligence

… 28, 2026, unless otherwise stated. USD conversions are presented for ease of reference and may not fully match prevailing exchange rates.

Can Lord of Mysteries become Kuaishou’s gaming breakout?

Kuaishou

East Asia

Gaming

Creative

… data-tmfx-note="1">Note: RMB figures are converted to USD at rates of RMB 6.74 = USD 1 based on estimates as of August 28, 2026, unless otherwise stated. USD conversions are presented for ease of reference and may not fully match prevailing exchange rates.

JD.com wants to build the infrastructure robots still lack

JD.com

… demonstrations. They are being built directly into production pipelines.
TapTap Maker, for example, has integrated multiple models into creative processes such as game-logic generation and coding, allowing creators to use different models according to their needs.
Tencent’s Codename Craft lets people without programming or art backgrounds describe ideas in natural language and generate playable game prototypes.
<figure id="attachment_163808" aria-describedby="caption-attachment-163808" style="width: 1920px" class="wp-caption alignnone"><img loading="lazy" decoding="async" class="size-full wp-image-163808" src="https://console.kr-asia.com/wp-content/uploads/2026/08/Tencent-Games-Codename-Craft-ChinaJoy-2026.jpg" alt="" width="1920" height="1280" srcset="https://console.kr-asia.com/wp-content/uploads/2026/08/Tencent-Games-Codename-Craft-ChinaJoy-2026.jpg 1920w, https://console.kr-asia.com/wp-content/uploads/2026/08/Tencent-Games-Codename-Craft-ChinaJoy-2026-720x480.jpg 720w, https://console.kr-asia.com/wp-content/uploads/2026/08/Tencent-Games-Codename-Craft-ChinaJoy-2026-1024x683.jpg 1024w, https://console.kr-asia.com/wp-content/uploads/2026/08/Tencent-Games-Codename-Craft-ChinaJoy-2026-768x512.jpg 768w, https://console.kr-asia.com/wp-content/uploads/2026/08/Tencent-Games-Codename-Craft-ChinaJoy-2026-1536x1024.jpg 1536w" sizes="(max-width: 1920px) 100vw, 1920px" /><figcaption id="caption-attachment-163808" class="wp-caption-text">Photo source: 36Kr.</figcaption></figure>
Tools like these are beginning to reshape development pipelines and, in turn, the workflows built around them.
Many creators 36Kr spoke with this year were no longer discussing AI as a standalone application for one part of production. Instead, they were considering how AI could be integrated across planning, art, development, testing, and publishing.
The second level is identity.
A growing number of game companies are beginning to redefine themselves and their products around AI.
At an event in May, Mihoyo co-founder Liu Wei said the company would invest up to RMB 100 billion (USD 14.8 billion) over the next three years to support AI projects.
Earlier, Anuttacon, a Silicon Valley company led by Cai Haoyu, had set its sights on developing a next-generation, game-specific large model capable of emotional understanding and autonomous interaction.
Veteran game company 37 Interactive Entertainment, meanwhile, has used investments to build exposure across several areas of physical AI, including world models, perception and interaction, execution control, AI embodiment, and end devices.
NetEase Fuxi has also assembled a broad set of enterprise-facing capabilities through products including its Youling agent platform, ACE high-performance distributed engine, and systems for collective intelligence and collaboration.
The third level is the industry’s understanding of AI itself.
AI has shifted from being treated primarily as a new technology to becoming a default consideration for many game companies. That change in mindset may be its deepest form of integration.
During the live stream discussion at ChinaJoy, Chen divided the industry’s concerns over the past few years into three stages.
Two years ago, people were anxious about feasibility. Last year, they worried about whether agents could actually execute tasks. Now, the questions increasingly concern what happens after something has been built: how to distribute it, make it compliant, and determine how it should be rated.
Zhao pointed to TapTap Maker’s numbers. The platform had 502 games in the first half of last year. This year, it has more than 5,000, including over 4,000 games produced with some form of AI involvement, according to Zhao.
That increase in supply also creates a challenge: ensuring there are enough players for all those games.
<h3>Game companies place different bets</h3>
Look beyond the individual AI tools on display at ChinaJoy, and it becomes clear that game companies are taking several different routes.
Tencent is pursuing an ecosystem strategy.
The company spent RMB 85.8 billion (USD 12.7 billion) on R&amp;D in 2025, while annual spending on Hunyuan and Yuanbao reached RMB 18 billion (USD 2.7 billion), according to 36Kr. That investment has continued into 2026.
For a large internet company, connecting the AI model layer with platforms, cloud infrastructure, and an existing ecosystem offers one way to build a full-stack position.
Mihoyo is taking a different route, centered on longer-term investment in proprietary R&amp;D.
Years of revenue from successful mobile games have given the company resources to support an AI investment plan of up to RMB 100 billion, according to Liu.
Its interest in AI also predates the current boom. In 2018, Mihoyo established an internal research division focused on frontier technologies and began developing Glossa, its proprietary game-specific large model.
The company later invested in MiniMax around the time the AI startup was founded in 2021. In 2023, Cai established Anuttacon in Silicon Valley. In 2026, Mihoyo announced its three-year, RMB 100 billion AI investment plan.
37 Interactive Entertainment has followed yet another path, using investments to build exposure across different parts of the AI and <a href="/industry/robotics">robotics</a> stack.
The strategies vary, but AI is becoming a more important competitive capability in gaming. It was not always the focus at ChinaJoy—not yet—but its growing presence across development tools, production workflows, game experiences, infrastructure, and corporate strategy points to what may lie ahead.
KrASIA features translated and adapted content that was originally published by 36Kr. This <a href="https://36kr.com/p/3946527216958851">article</a> was written by Liu Shiwu for 36Kr. 
Note: RMB figures are converted to USD at rates of RMB 6.74 = USD 1 based on estimates as of August 28, 2026, unless otherwise stated. USD conversions are presented for ease of reference and may not fully match prevailing exchange rates.

ChinaJoy 2026 shows how deeply AI is becoming embedded in gaming

Artificial Intelligence

… Newslink.
Note: RMB figures are converted to USD at rates of RMB 6.74 = USD 1 based on estimates as of August 28, 2026, unless otherwise stated. USD conversions are presented for ease of reference and may not fully match prevailing exchange rates.

Chery reports higher overseas and NEV revenue in first interim results

Chery

Transportation

Electric Vehicles

Mobility

Why do companies go public?
For money. Obviously.
That is particularly easy to see in artificial intelligence. Developing large models and securing the compute to train and run them is expensive, making repeated fundraising almost a feature of the business. In China, companies such as Z.ai and MiniMax have already turned to public markets as their capital needs grow.
Then there is <a href="/company/shein">Shein</a>.
The online fashion retailer <a href="https://kr-asia.com/shein-launches-hong-kong-public-offering-with-its-efficiency-model-in-focus?utm_source=chatgpt.com">launched its Hong Kong public offering</a> on August 24, seeking to raise up to USD 1.8 billion ahead of a planned September 1 listing.
But Shein does not appear short of cash.
As of March 31, it had USD 14.831 billion in cash resources. From 2023 to 2025, it generated about USD 6 billion in net cash from operating activities. Customers also pay upfront, while inventory turned over in just 36 days in 2025.
Another USD 1.8 billion would be useful, but hardly essential. So what is the IPO for?
Growth is part of the answer. Shein has been expanding beyond fashion, while technology, branding, and international expansion all require investment.
But another answer sits on the <a href="https://kr-asia.com/why-shein-may-need-an-ipo-despite-having-usd-14-8-billion-in-cash?utm_source=chatgpt.com">other side of its balance sheet</a>.
As of March 31, Shein’s convertible redeemable preferred shares were valued at USD 17.294 billion—more than its cash holdings. Those shares can convert into ordinary equity or, under certain conditions, be redeemed by investors.
They reflect more than a decade of private fundraising dating back to 2015. Shein’s pre-money valuation reached USD 98.2 billion in its 2022 Series D round before falling to USD 64 billion in the subsequent Series D+ round.
To cushion investors against that decline, holders of preferred shares issued before and during the Series D and D+ rounds receive annual cash returns. The rate rose from 8% to 12% on March 5 and will continue to accrue until Shein completes its listing.
There is also redemption risk. The redemption rights attached to Shein’s preferred shares were suspended after it submitted its IPO application, but can be restored under certain circumstances if the listing fails.
If that happened and preferred shareholders fully exercised those rights, Shein’s cash resources would not cover the USD 17.294 billion carrying amount of the preferred shares.
Completing the IPO changes that equation. The preferred shares would convert into equity, removing the redemption risk and resolving a substantial overhang accumulated during Shein’s years as a private company.
For a capital-hungry AI company, going public can mean raising money to finance what comes next. For Shein, the listing is also meant to resolve what came before.
The point is, public markets can serve a broader purpose. An IPO can raise capital, but it can also provide liquidity to shareholders, establish a traded valuation, simplify ownership structures, and give investors an exit from arrangements that become more complicated the longer a company remains private.
Shein shows why asking whether a company needs the money can, at times, miss the point.
Sometimes an IPO is about financing the future. Sometimes it is also about settling the past.
<h3>Related readings</h3>
<ul>
<li>Behind Shein’s USD 1.8 billion offering is a business built to move quickly while tying up relatively little cash.<a href="https://kr-asia.com/shein-launches-hong-kong-public-offering-with-its-efficiency-model-in-focus"> Its IPO prospectus offers a closer look at the operating efficiency underpinning that model.</a></li>
<li>The longer Shein waits to list, the more expensive some of its earlier financing becomes. <a href="https://kr-asia.com/why-shein-may-need-an-ipo-despite-having-usd-14-8-billion-in-cash">A look at the terms attached to its preferred shares reveals why a cash-rich company can still face pressure to go public.</a></li>
<li>Shein wants investors to look beneath the clothes. Its IPO case also rests on <a href="https://kr-asia.com/sheins-ipo-case-rests-on-the-technology-behind-its-fashion-business">the technology connecting consumer demand, suppliers, production, and inventory across its sprawling fashion business.</a></li>
</ul>

The Uptake | What an IPO is for

Shein

Fashion

Retail

Consumer Economy

Hong Kong

Newsletters

… advantages here since American labs have to comply with many additional restrictions on training data," Zuckerberg said in the post. "US policy must reduce this additional friction if we want American open source models to lead over time."
Case in point: DeepSeek currently ranks as the most-used AI model in the world, with a 27% market share, according to OpenRouter. <a href="/company/alibaba">Alibaba</a>, meanwhile, said its Qwen is the world's most downloaded open-source model family. The Chinese tech giant released the latest version, Qwen 3.8-27B, on August 17. All told, Alibaba has made more than 460 of its models open source and accumulated more than three billion global downloads.
Washington may be heeding Zuckerberg's warning.
Multiple US media outlets reported this month that the Trump administration has excluded open models from a voluntary review framework intended to evaluate the safety of advanced AI models.
It stands to reason that both Silicon Valley and Washington are increasingly nervous about China's lead in open-weight AI.
"A lot of manufacturers, both in Europe and in the US, are using open-weight models from China because they don't trust Anthropic or OpenAI. They don't want to send their critical data there," said Pierre Baque, founder and CEO of Neural Concept, an AI design software company that works with automakers, chipmakers, and other manufacturers to speed up their design workflow.
"The Chinese cannot spy on you through a model that you host," Baque said, adding that a lack of American open-weight frontier models was "really creating a huge opportunity for China."
Part of that opportunity could be geopolitical.
At the World AI Conference in Shanghai this July, Chinese President Xi Jinping <a title="" href="https://asia.nikkei.com/business/technology/artificial-intelligence/xi-dabbles-in-ai-diplomacy-at-shanghai-forum-with-thai-cambodian-pms" target="_blank" rel="noopener">reaffirmed China's commitment</a> to open-source AI. By positioning itself as an advocate of open AI models, China can create structural reliance among the countries and companies that adopt them, reliance it could later weaponize if it chooses to, said an associate at an international law firm who asked not to be named due to the firm's presence in China.
"Open-source AI is probably becoming China's new chokehold, alongside rare earths, and it's something the Chinese government could potentially weaponize," the lawyer told Nikkei Asia.
As open AI becomes the latest frontier in the US-China tech race, analysts expect Washington to restrict Chinese models despite Silicon Valley's pushback.
Because it is difficult to remove open-source or open-weight models from the internet, or prevent users from downloading them once they are available, Corso at University of Michigan said the most straightforward option for Washington is to make it illegal to use such models.
Export controls are another potential tool. The Trump administration said in July that Moonshot's Kimi K3 was trained on Nvidia chips in violation of US export restrictions.
"We believe the most likely US response is to tighten export controls to prevent AI chips from being diverted to China, and require all US AI players to put anti-distillation features in their advanced models," Jefferies analyst Edison Lee said.
Any action, however, is unlikely to come ahead of next month's meeting between Trump and Xi in Washington, analysts say.
Trump said in July that he will discuss AI with Xi when the latter visits the White House in September. The two leaders met in Beijing in May, but the readout from that meeting focused mostly on trade issues.
"There's a reluctance on the part of both governments to create the conditions that would lead to an unproductive meeting," Mehta at CSIS said, adding that there might be "work happening below the surface" in Washington targeting Chinese AI, but a lot of those actions will be "held in reserve, pending the outcome of the upcoming summit."
This article first appeared on <a href="https://asia.nikkei.com/business/technology/tech-asia/why-open-weight-ai-is-the-new-frontier-in-the-us-china-tech-race">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>.

Why open-weight AI is the new frontier in the US-China tech race

Foundation and Large Models