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 July 9, British tech brand Nothing officially launched its first smartphone in China, pricing it at USD 799, the same as the iPhone 16. Alongside the flagship Phone (3), the company also released its first over-ear headphones, the Headphone (1), priced at RMB 2,299 (USD 321.9).
In terms of performance, the Phone (3) isn’t exactly a standout. It runs on the Qualcomm Snapdragon 8s Gen 4 processor, a chip commonly found in midrange models like the Redmi Turbo 4 Pro and iQoo Z10 Turbo Pro. The device features a triple-camera setup with 50-megapixel sensors and a 5,150-milliampere-hour battery that supports 65-watt wired and 15-watt wireless charging. These specifications fall short of what is typical in most flagship smartphones today.
But performance was never the brand’s primary focus. As a startup, Nothing has leaned more heavily on design and brand identity. The Phone (3) introduces a notable visual update over its predecessors. Its signature “Glyph” light strips now form a monochrome LED subdisplay called the “Glyph Matrix,” composed of 489 individually controlled LEDs. This screen can display pixel-style alerts for calls, app notifications, and battery status, and it supports user-defined animations.
The device also debuts a new software suite called Essential Space, which includes features like “flip to record” and “essential search.”
<figure id="attachment_155559" aria-describedby="caption-attachment-155559" style="width: 640px" class="wp-caption aligncenter"><img loading="lazy" decoding="async" class="wp-image-155559" src="https://console.kr-asia.com/wp-content/uploads/2025/07/Nothing-Phone-3.jpg" alt="Image shows accessories included with the Nothing Phone (3)." width="640" height="480" srcset="https://console.kr-asia.com/wp-content/uploads/2025/07/Nothing-Phone-3.jpg 1920w, https://console.kr-asia.com/wp-content/uploads/2025/07/Nothing-Phone-3-720x540.jpg 720w, https://console.kr-asia.com/wp-content/uploads/2025/07/Nothing-Phone-3-1024x768.jpg 1024w, https://console.kr-asia.com/wp-content/uploads/2025/07/Nothing-Phone-3-768x576.jpg 768w, https://console.kr-asia.com/wp-content/uploads/2025/07/Nothing-Phone-3-1536x1152.jpg 1536w" sizes="(max-width: 640px) 100vw, 640px" /><figcaption id="caption-attachment-155559" class="wp-caption-text">Image shows accessories included with the Nothing Phone (3). Image and header image source: Nothing.</figcaption></figure>
The smartphone industry has long operated in a cycle of performance upgrades. With diminishing returns on hardware improvements, the current challenge is to serve niche audiences with precision.
The generation that was born the same year the iPhone launched is now reaching adulthood. Raised on digital devices, these users want more than technical specifications, seeking emotional connection and personal expression. Nothing’s global user base has an average age of just 26, about ten years younger than that of many competing brands. More than 60% of its users cite design as the top reason for purchase. That alignment gives Nothing an edge among younger buyers who are growing tired of uniformity across established brands.
With this audience in mind, Nothing has gradually broadened its product lineup. This month’s debut of the Headphone (1), an over-ear model, reflects that direction. Unlike TWS (true wireless stereo) earbuds, which are often seen as purely functional, over-ear headphones are viewed more as a style choice.
<figure id="attachment_155560" aria-describedby="caption-attachment-155560" style="width: 1920px" class="wp-caption alignnone"><img loading="lazy" decoding="async" class="size-full wp-image-155560" src="https://console.kr-asia.com/wp-content/uploads/2025/07/Nothing-Headphone-1.jpg" alt="Image of the Headphone (1) by Nothing." width="1920" height="1080" srcset="https://console.kr-asia.com/wp-content/uploads/2025/07/Nothing-Headphone-1.jpg 1920w, https://console.kr-asia.com/wp-content/uploads/2025/07/Nothing-Headphone-1-720x405.jpg 720w, https://console.kr-asia.com/wp-content/uploads/2025/07/Nothing-Headphone-1-1024x576.jpg 1024w, https://console.kr-asia.com/wp-content/uploads/2025/07/Nothing-Headphone-1-768x432.jpg 768w, https://console.kr-asia.com/wp-content/uploads/2025/07/Nothing-Headphone-1-1536x864.jpg 1536w, https://console.kr-asia.com/wp-content/uploads/2025/07/Nothing-Headphone-1-480x270.jpg 480w" sizes="(max-width: 1920px) 100vw, 1920px" /><figcaption id="caption-attachment-155560" class="wp-caption-text">Image of the Headphone (1) by Nothing. Image source: Nothing.</figcaption></figure>
Nothing’s global market share remains small, at just 0.2%. However, it is the only smartphone startup in the past decade to achieve any meaningful scale. While the industry remains dominated by major players, Nothing demonstrates that it is possible to build a business by serving niche segments.
The company is unlikely to become a dominant force in the market, and expectations are modest. But maintaining relevance with a focused user base would, in itself, be a notable accomplishment.
KrASIA Connection features translated and adapted content that was originally published by 36Kr. This <a href="https://36kr.com/p/3371522037469701">article</a> was written by Yuan Silai for 36Kr.

Nothing brings its design-focused phones to China, starting with the Phone (3)

Smartphones

Consumer Economy

China

Europe

… translated and adapted content that was originally published by 36Kr. This <a href="https://36kr.com/p/3360208796125189">article</a> was written by Zhou Xinyu for 36Kr.

China’s video AI startups are doing what LLM firms haven’t: Making money

Artificial Intelligence

Deep Tech

Foundation and Large Models

… that’s what we’re after.
<h2>A non-follower meant only to be a CEO</h2>
<h3>AN: You dropped out at 18. Was that a hard decision? How did you convince your parents?</h3>
KF: It wasn’t hard. I went to high school in the US and was already working on Web3 projects. Not going to college just felt natural. I knew it wouldn’t help me. I was absolutely certain I wanted to build something of my own.
I don’t think I ever sat down and officially told my parents. We’re a pretty independent family. Everyone’s focused on their own work. We don’t talk a lot.
Honestly, I still don’t know if they are fully aware of my decision. They just know I’m running a company. Whether I’m studying or not, I kind of left that vague.
<h3>AN: You started making money from Web3 projects while in high school. Were you drawn to the industry’s fast pace?</h3>
KF: Back then, I just wanted independence. Studying overseas gave me a strong sense of insecurity. I wasn’t sure my family could always support me the same way.
Also, school was just boring. I couldn’t sit through classes. If the teacher was talking about economics, I’d instantly fall asleep. Only after I got home would I feel alive. That’s when I started exploring Web3. It was fun and exciting.
<h3>AN: Did you always dislike school?</h3>
KF: Yes, it was painful. I was constantly getting into trouble, being called out in class, sent to stand in the hallway, and got my parents brought in for meetings. I think most of the teachers found me annoying.
<h3>AN: Did that ever make you question yourself?</h3>
KF: No. I thought the problem was with the system, not me. My grades were good. I just didn’t fit the format.
<h3>AN: Does your confidence come from your family or yourself?</h3>
KF: My parents didn’t praise me much. They had no specific expectations.
I’ve always trusted my own judgment. For instance, I told my parents I wanted to switch to a US high school because I felt I couldn’t grow in China. They didn’t agree at first, but I insisted.
<h3>AN: How did you discover your talent?</h3>
KF: Trial and error. At one point I thought I wanted to do development work and write code. But I quickly realized I couldn’t sit still. I also wasn’t that good at it.
I’m not the type to grind at something I don’t think I’ll ever excel at.
When I started doing growth work in Web3, something just clicked. I could instantly tell which influencers would work. I have a kind of instinct for it.
<h3>AN: You were already head of growth at Dora AI. What pushed you to start your own thing?</h3>
KF: Dora AI’s two founders were much older than me. I constantly had to persuade them to try things. We clashed a lot. It was frustrating. I thought, if I’m capable, why am I stuck here?
The CEO said I’m someone who can’t be a follower. That’s true. I can only be a CEO. I can’t be someone’s deputy.
<h3>AN: You said you’re impatient. Isn’t patience key to being a CEO?</h3>
KF: I’m working on it. I process things fast, and I get frustrated if others don’t keep up. Sometimes I think I’m being clear, but people feel I’m being too intense.
I’m not a cuddly boss. But I reflect daily whether I’ve improved a little, whether I’m stretching myself beyond my instincts.
<h3>AN: What else are you trying to improve?</h3>
KF: I generate too many ideas. I used to dump 100 a week on the team. Now I filter and only share 50.
<h3>AN: How close is Head AI to your ideal vision?</h3>
KF: Still far off. We’re starting with one pain point in today’s marketing landscape, like how Salesforce started with customer relationship management. My end goal is to build a system that replaces the entire marketing team’s brain and limbs.
whereRelatedArticle
<h3>AN: If you weren’t in marketing, what would you do?</h3>
KF: I’d run political campaigns.
Politicians are basically products. You sell them to the public to win votes. It’s just marketing at scale. But that’s just a fun thought. Right now, I’m all in on Head AI.
<h3>AN: What’s your weekday and weekend routine?</h3>
KF: Weekdays are intense. I barely check WeChat. I work from 11 a.m. to 12 or 1 a.m.
On weekends, I unplug completely and don’t even want to talk. I take ballet and modern dance classes. I go to plays. Lately I’ve been especially into theater. Losing myself in someone else’s story feels incredible. It makes me happy.
<h3>AN: Do you date?</h3>
KF: I used to have crushes in middle school, but not anymore. I started liking people too early, and now it feels pointless. Work is more interesting.
<h3>AN: Name three of your favorite books.</h3>
KF: “Heart” by Kazuo Inamori taught me the importance of staying on the right path. “The Ride of a Lifetime” (by Bob Iger) helped shape how I think as a CEO.
I used to love reading Eileen Chang and detective novels, especially works by Keigo Higashino and “Murder on the Orient Express.” But since starting the company, I haven’t had much time to read.
<a href="/collections/100-ai-creators">100 AI Creators</a> is a collaborative project between AI Now! and KrASIA, highlighting trailblazers in AI. Know an AI talent we should feature? <a href="mailto:ainow@kr-asia.com">Reach out to us.</a>

With Head AI, this Gen Z dropout aims to replace marketing teams entirely

Enterprise Services

… https://console.kr-asia.com/wp-content/uploads/2025/07/LLVision-Leion-Hey2-VivaTech-1024x683.jpg 1024w, https://console.kr-asia.com/wp-content/uploads/2025/07/LLVision-Leion-Hey2-VivaTech-768x512.jpg 768w" sizes="(max-width: 480px) 100vw, 480px" /><figcaption id="caption-attachment-155500" class="wp-caption-text">Attendees at the recent VivaTech event in Paris had a chance to try out the Leion Hey2 smart glasses in person. Photo source: 36Kr.</figcaption></figure>
<h3>Designed for the real world</h3>
The Leion Hey2 marks a significant leap from its predecessor, offering a more refined design, lighter build, and brighter display. It also reflects LLVision’s years of accumulated technical experience.
LLVision has operated in the AR and AI space for over a decade. Since 2014, it has been developing geometric array waveguides, which entered mass production in 2018. The company has long supplied AR solutions to clients in sectors such as manufacturing, energy, aerospace, and logistics, and it has led China’s enterprise AR shipments for several years.
In 2022, LLVision entered the consumer market, adapting enterprise technology for everyday users. Its first product, the Leion Hey, sold more than 30,000 units. On average, users spent 150 minutes per day using the device, suggesting it served more than a niche need. It became especially useful to travelers and people with hearing impairments who needed a visual aid for spoken language.
The Leion Hey2 took three years to develop. Wu said the company deliberately focused on perfecting a single use case: translation.
“Translation might seem like a simple feature, but it’s incredibly demanding in different real-world scenarios,” he said. “We’d rather make one product that scores 95 out of 100 than several that score 70.”
Translation involves more than converting language. Timing and conversational flow are critical to preserving the natural rhythm of speech.
Smart glasses like the Leion Hey2 are redefining the possibilities. In a Tokyo restaurant, users can read subtitles while their server speaks. In a noisy Seoul subway, the glasses can zero in on a specific voice. At a tech expo in Munich, an engineer can hold a live conversation across languages, complete with real-time technical references.
And despite its expanded capabilities, the Leion Hey2 is priced over 60% lower than the high-end version of its predecessor, thanks to supply chain advances in China.
Improved domestic production, especially for components like system-on-chip (SoC) circuits and flexible boards, has cut AR manufacturing costs by more than 30%, boosting consistency and reliability.
Riding this momentum, LLVision is expanding into North America, Europe, and Southeast Asia.
Over the past two years, the AR device sector has entered a new phase of competition. Established players are consolidating their leads while new entrants are pushing forward with innovations in design, software, and core features.
According to iResearch, AR glasses shipments in China are projected to reach 286,000 units in 2024. Advances in waveguide optics and large AI models are steering the industry toward lighter, more functional, and socially integrated products.
A report by Huajin Securities estimates that shipments of smart audio glasses could exceed five million units in 2025 and reach 9.5 million by 2030. These devices, which prioritize audio but may incorporate AR features, could soon compete directly with today’s true wireless stereo earbuds.
Subtitle-enabled smart glasses are just one slice of this broader market, but a standout one. With continued adoption, their “iPhone moment” may not be far off.
KrASIA Connection features translated and adapted content that was originally published by 36Kr. This <a href="https://36kr.com/p/3353338990949257">article</a> was written by Xiao Xi for 36Kr.

Break the language barrier: Leion Hey2 brings live translation to eyewear

Augmented and Virtual Reality

Hardware

… overseas.
If those 1,000 brands succeed, JD.com succeeds. Amazon doesn’t have them. Local retailers don’t have them either. If we just sell what Amazon sells, we’d have no advantage. But this will take another five years. We have to work with each brand to ensure compliance, obtain certifications, and so on. It’s a long road.
<h3>13. From 38 employees to 900,000 full-timers in 12 years</h3>
RL: In 2003, including myself, JD.com had 38 people. We made about RMB 120,000–200,000 (USD 16,800–28,000) that year. I told our staff: however many employees we have in the future, we’ll adopt the same number of orphans. As of the first quarter this year, we have 720,000 employees.
In Q2, with the addition of delivery staff, we surpassed 900,000. It’s the largest single-quarter headcount increase since our founding, with over 150,000 net new hires.
<h3>14. All P5-level employees saw their salaries doubled</h3>
RL: I don’t believe in squeezing our people. I believe in giving raises. All employees at the P5 level and above have had their salaries doubled.
<h3>15. Full-time couriers in Beijing earn over RMB 13,000 (USD 1,800) per month before taxes</h3>
RL: From the day I launched JD Logistics in 2007, we’ve paid full social benefits for every courier. In Beijing, our full-time delivery staff now earn more than RMB 13,000 (USD 1,820) per month, and we pay benefits based on that amount.
<h3>16. Only five layers of management separate Liu from frontline workers</h3>
RL: I’ve set a hard rule whereby as long as JD.com has fewer than two million employees, we’re only allowed five layers of management. Most companies with just a few thousand people already have six or seven layers. We only have five. That’s what a flat organization means.
<h3>17. JD.com dominates China’s home appliance market</h3>
RL: JD.com is the undisputed leader in home appliances. Gome and Suning combined aren’t even 20% of our scale. Yet our profit margin in this segment is just three to four percentage points, compared to the national average of over 10%.
Appliance brands make twice as much as we do. Back in the day, Gome and Suning made three to four times as much as the brands.
<h3>18. JD Retail operates at 10% of total cost</h3>
RL: Our core advantage is efficiency. JD Retail’s total cost of operation is just 10%, on par with Costco, Sam’s Club, Aldi, Amazon, and us. Traditional supermarkets run at 20%. Gome and Suning are at 15%. This cost discipline is what keeps us alive.
<h3>19. Keeping to a 59-day payment cycle</h3>
RL: Our inventory turns over in 30 days. For anyone who’s worked in retail, that number means everything. We believe we can bring that down to 20–25 days. Each day saved generates RMB 10 billion (USD 1.4 billion) in cash flow.
Our payment cycle is 59 days. By comparison, most major retailers take 120. That’s why we don’t squeeze suppliers, and we don’t need to.
<h3>20. Net profit is 1% of GMV, and Liu is proud of that</h3>
RL: As of last year, JD.com’s GMV (gross merchandise volume) was over RMB 4 trillion (USD 560 billion). Net revenue was RMB 1.16 trillion (USD 162.4 billion). Net profit was just over RMB 40 billion (USD 5.6 billion).
We paid RMB 116.1 billion (USD 16.3 billion) in salaries. We paid RMB 18 billion (USD 2.5 billion) in social benefits last year alone, and more than RMB 100 billion (USD 14 billion) in total over the last 18 years. That could have been profit. It could have been my personal wealth. But that’s not how we operate.
Yes, we need to make money. But profit has never been our top priority. And I’ll say that anywhere.
KrASIA Connection features translated and adapted content that was originally published by 36Kr. This <a href="https://36kr.com/p/3341555301677576">article</a> was written by Ren Cairu for 36Kr.

Richard Liu and 20 things he wants you to know about JD.com’s future

JD.com

East Asia

On-demand Services

Travel

E-commerce

Logistics

Food Delivery

… region.
"The reason why the peg works so well is also due to the clarity of the signal that the Hong Kong authority is sending to the market," according to Kelvin Lam, senior China economist at Pantheon Macroeconomics. "If you widen the band, the objectives may not be that clear."
The broader economic challenges Hong Kong faces could give the authorities reason to keep interest rates lower for longer.
Unlike in 2009 when, a weaker US dollar prompted similar action by the Hong Kong authority to pump liquidity into the market, fueling concerns over asset bubbles, Hong Kong's economy could now get a boost from lower rates, helping to reinvigorate its sluggish housing market and sectors sensitive to rates, such as banking.
"There is no reason not to peg to a weak currency at a time when the underlying economy is facing deflation," added Wang of BNP Paribas.
This article first appeared on <a href="https://asia.nikkei.com/Business/Markets/Trading-Asia/Hong-Kong-s-borrowing-costs-plunge-to-nearly-zero-as-it-works-to-maintain-dollar-peg">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>.

Hong Kong's borrowing costs plunge to nearly zero as it works to maintain dollar peg

Hong Kong

… MiniMax Agent, the company’s general-purpose AI system.
While Hailuo focuses on video, MiniMax Agent handles long-horizon, complex tasks. It can deconstruct goals, create expert-level plans, and execute multi-step operations. It’s widely used internally for tasks like coding, building websites, and making presentations.
It understands longform content across text, audio, image, and video. With its model context protocol (MCP) extensions, it supports animations, advertising content, and slide decks.
One demo showed it building an interactive web page about the Louvre Museum, complete with media and engagement features.
Beyond being a productivity tool, MiniMax Agent reflects the depth of MiniMax’s foundation model and multimodal system.
And why should companies building foundation models also build agents? Because they have control over model architecture and training loops. They can create feedback systems that refine performance and reduce costs.
Put simply, model innovation lifts the ceiling. Agents help define the next frontier. The two are interdependent.
MiniMax launched in early 2022, nearly a year before ChatGPT became a household name. Once public attention spiked, many began asking: who is MiniMax?
The answer: a company that has consistently followed its own path.
In 2023, while most of China’s AI market was fixated on dense models, MiniMax allocated more than 80% of its compute to MoE. In early 2024, it launched China’s first MoE-based foundation model.
In January, it open-sourced its lightning attention protocol, altering core transformer architecture in pursuit of new performance gains.
By June, its M1 model had reached second place in global open-source rankings, backed by architectural and algorithmic advances across the board.
Launched alongside it, the general-purpose MiniMax Agent addresses key challenges in AI agent deployment, including multimodality, long-horizon reasoning, task decomposition, and tool coordination.
MiniMax is a company that never seems to tire, that constantly reinvents itself, and that is always chasing the next performance ceiling.
At its core, this drive reflects a belief in the foundation model race’s endgame: that a great business starts with a simple idea, pursued through focused, uncompromising technical innovation.
KrASIA Connection features translated and adapted content that was originally published by 36Kr. This <a href="https://36kr.com/p/3360386652817158">article</a> was written by Xiao Xi for 36Kr.

MiniMax finds its “aha moment,” and a way to challenge the AI status quo

… challenges. Xu noted that traditional automakers often complete their production-to-dealership cycle within 60 days. Once vehicles reach a 4S dealership, the dealer pays the automaker, allowing cash to flow even if the cars are still sitting unsold.
In contrast, direct-to-consumer brands like <a href="/company/xiaomi">Xiaomi</a> may be able to turn over inventory quickly. But companies with weak sales performance face greater difficulty.
“For companies that still aren’t profitable, strictly enforcing 60-day terms could severely compromise their cash flow,” Xu said.
According to a 2018 <a href="https://www.wsj.com/articles/china-has-487-electric-car-makers-and-local-governments-are-clamoring-for-more-1531992601">Wall Street Journal</a> report, China once had over 487 electric vehicle manufacturers. Today, that number has shrunk to around 40.
What began as a price war has devolved into broader uncertainty. The 60-day policy may be intended to restore order, but whether it will lead to a more sustainable ecosystem remains uncertain.
KrASIA Connection features translated and adapted content that was originally published by 36Kr. This <a href="https://36kr.com/p/3368561494525705">article</a> was written by Geng Chenfei for 36Kr.

For China’s auto suppliers, a 60-day pledge may still mean waiting months for payment

Transportation

Electric Vehicles

Mobility

… Embodied intelligence spans a wide range of tasks. A robot might master pick-and-place but still be unable to lift a child or help someone stand up.
Each product needs its own metric. To be called intelligent, a robot should reach at least Level 4: autonomous operation, not just functioning as a tool.
Unlike the rapid leap seen with ChatGPT, embodied intelligence is progressing slowly and incrementally.
<h3>36Kr: So the “ChatGPT moment” for embodied AI is still far off?</h3>
WH: Yes. ChatGPT works for general Q&amp;A. Embodied AI faces hurdles in hardware, sensors, and data collection. Many components are still immature. We’re five to ten years away.
Humans rely on more than vision, language, and action. We use hearing, smell, touch, and temperature. Vision-language-action (VLA) models are just the beginning. To achieve human-level embodied intelligence, we’ll need to integrate more modalities.
Right now, the most viable path is perfecting mobile pick-and-place in predictable environments like supermarkets and factory lines. If we can scale that, it will be a huge milestone akin to building a fully automated dark factory.
<h3>36Kr: Are others in the industry aiming for that same milestone, or pursuing different breakthroughs?</h3>
WH: Honestly, not many are doing the hard work. Some just sell hardware or platforms without ensuring end users can solve problems.
Even among those focused on models, most are still doing academic work. Very few are building practical products for real deployment. That lack of responsibility and realism is why the field feels fragmented.
<h3>36Kr: What needs to be done to deploy mobile pick-and-place robots in services like pharmacies and convenience stores?</h3>
WH: Every new deployment requires data prep. Whether it’s synthetic data, small-scale real-world data, or even reinforcement learning in a specific setting, some kind of groundwork is essential for building a product that performs consistently.
We’re not trying to solve every problem at once. We’re starting with grocery and retail shelving and making sure our robots can generalize well in those specific scenarios. Only after that will we scale to more varied environments. It’s a tougher path than most people realize.
<h3>36Kr: Beyond pick-and-place, what’s Galbot’s next frontier?</h3>
WH: Our team of researchers is working on legged robots, dexterous hands, and more. Dexterous manipulation is a personal focus—I’ve won several awards in this area. It's the ultimate challenge in robotic control.
Our strategy is to stay ahead of the curve. Galbot’s mission is to create general-purpose robots that can serve every industry and household.
KrASIA Connection features translated and adapted content that was originally published by 36Kr. This <a href="https://36kr.com/p/3350684885707399">article</a> was written by Wang Fangyu for 36Kr.

Galbot’s humanoid robots do the boring stuff, and that’s the point

Embodied Intelligence