Author: Alan, Amelia | Biteye Content Team; Denise | XHunt Operations Team

In the summer of 2026, the information flow on social platforms is refreshing in milliseconds. One second, a major language model releases an update; the next second, tens of thousands of "in-depth interpretations" are flooding in.

An independent developer told us that the first thing he does every day upon waking is no longer scrolling through his timeline, but quickly scanning a few familiar avatars to see what new tricks they have Vibe Coding the night before.

"I only trust those who have practical experience," he said.

This seemingly paranoid trust points to a truth that most people overlook:

In today's rapid advancement of large model technology, general information itself is depreciating at the speed of light.

Traditional tech media accounts that relied on quickly relaying news, translating overseas announcements, and simply piecing together articles are gradually losing users' patience. The truly scarce resource is no longer "who said what first," but "who can tell me whether this is reliable and how I should use it."

To uncover the real operational logic of this hidden circle, we conducted an in-depth analysis of nearly 400 top AI KOLs in both Chinese and English ecosystems based on exclusive data and capability models from the social analysis tool @xhunt_ai, examining tens of thousands of tweet samples.

We found that the opinion leaders in the AI era are undergoing a profound transformation from "information intermediaries" to "productivity enablers."

1. Core Finding: From Distributing Opinions to Distributing Productivity

In the context of traditional internet, an individual with a brilliant idea needs to mobilize a complex human chain to bring that idea to fruition: backend, frontend, UI, product managers… The lengthy collaboration process can drain most of the enthusiasm. Today, AI tools have drastically compressed this production chain. Codex, Claude Code, Cursor, and Lovable have transformed programming barriers into logical and structural capabilities; Seedance, GPT Image, 可灵, and Nano Banana have directly eliminated the complex barriers to image and video production.

However, this has triggered a counterintuitive industry phenomenon: when anyone can use AI to churn out lengthy articles in bulk, high-quality content becomes "cheap" and readily available, while trust becomes more scarce than ever.

The core value of AI KOLs lies not in their ability to make AI produce a piece of fluff faster than ordinary people. Instead, it is their ability to be the first to visualize the ambiguous power of AI usage through human-machine collaboration into results that others can see, run, and directly reuse. This is no longer about distributing opinions, but about distributing productive capabilities.

For example, when a new model claiming to "defeat Claude Opus 4.7" is released, users are already tired of the same old press releases. They are eager to know from trusted KOLs: "Will it hallucinate in real code development? Is this product, which looks incredibly cool in the official polished video, really a productivity tool that ordinary people can use right out of the box?"

The attention indicator has clearly reversed: from "what happened," upgraded to "is it important," and then to "how to use it."

In a noisy environment, AI KOLs play the role of practical pioneers and trust anchors.

2. Who Plays This Role: Veteran Technologists and New Generation Blue Ocean

A common industry bias is that "most AI KOLs are marketing accounts that quickly gained traction after the explosion of ChatGPT at the end of 2022." However, XHunt's generational statistics on account registration times debunk this claim: the generational structure of AI KOLs shows an inverted pyramid distribution.

• Dominance of seasoned practitioners: Among English accounts, early users registered between 2007 and 2015 account for as much as 62.9%; this proportion is also 58% in the Chinese list. This means that the majority of top accounts active in the core AI circle are practitioners and entrepreneurs who have accumulated experience through the PC, mobile internet, and Web3 cycles. With the arrival of the large model wave, they have keenly completed the migration of productivity tools.

• Growth of the new generation in the Chinese region: Notably, during the ChatGPT explosion period from 2022 to 2023, the proportion of newly emerged AI-native accounts in the Chinese region reached 13.0%, higher than the 9.7% in the English region. This indicates that the Chinese ecosystem has provided significant traffic incentives for practical content; as long as the tools are proficient and the tutorials solid, new accounts can establish competitive advantages through continuous posting.

In contrast, the registration times of Web3 KOLs often show a clear spindle shape, with a large number of new accounts created during the DeFi Summer, NFT explosion, and Meme craze, coinciding with market heat.

3. Symbiotic Evolution of AI KOLs and OPC

The evolution of AI is transforming the concept of One Person Company (OPC) from a superhuman concept into a clearly actionable reality. The core essence of OPC is that users can sensitively call upon various vertical AI Agents, liberating themselves from solitary struggles and the burden of all the hard work, infinitely amplifying their ideas, and using AI to complete independent product construction, commercial distribution, and precise marketing.

In this transformation, "application distribution" AI KOLs have firmly established their core ecological position through their composite advantages:

• Understanding technical boundaries: Many of them come from major AI companies or are seasoned developers, possessing foundational technical knowledge that allows them to understand the real limitations of tools better than pure marketers.

• Understanding market pain points: As long-term content creators facing their audience, they possess strong productization and marketing awareness, understanding real needs better than pure researchers.

It is this dual Buff of "technology + internet sense" that enables them to publicly build (Build in Public) and transform abstract technology into usable scenarios, thereby accumulating a continuous stream of user trust.

The trending Vibe Coding phenomenon has pushed this personal IP tension to the extreme: when a leading AI KOL recommends a development framework, they no longer write a few bland lines of recommendation but directly demonstrate on X how they can quickly launch a complete, interactive AI application in 15 minutes, collaborating with the model using just a natural language command, all while maintaining a relaxed atmosphere.

Traditional KOLs harvest traffic by distributing opinions, while AI KOLs consolidate the ecosystem by distributing productive capabilities.

4. Data Portrait: Ecological Divides Between Eastern and Western KOLs

To explore the real operational logic of the AI KOL ecosystem, this report extracted 100 tweet samples from the Top 300 AI KOLs in the English region and the Top 100 in the Chinese region over the past three months, conducting in-depth calculations and comparisons of their tweet content and various data.

We found that there are significant differences in attention structure and content production models between Chinese and English AI KOLs. The following will reveal the true face of AI KOLs across seven core dimensions: traffic volume, discussion fields, account creation time, and personal profiles.

1️⃣ Attention Map: English Region Focuses on Sources, Chinese Region Focuses on Practice

• Traffic volume distribution: The total number of followers in the English list exceeds 350 million, with an average of 1.17 million and a median of 110,669. The Chinese ecosystem leans towards refined vertical fields, with an average follower count of about 77,000 and a median of 43,006.

• Posting activity comparison: In the past 90 days, the 100 accounts in the Chinese list produced nearly 30,000 tweets, with a median posting volume as high as 210. In contrast, the 300 English accounts produced only 37,000 tweets in total, with a median of just 38. Top English accounts often remain in low-frequency posting states, while Chinese accounts form a high-frequency application diffusion network.

• Follower tier structure: The English list presents a pyramid structure, with accounts having 50,000 to 200,000 followers accounting for the highest proportion at 41.8%, while accounts with over 1 million followers account for 7.4%. The Chinese list, on the other hand, concentrates on the long-tail application layer, with accounts having 10,000 to 50,000 followers accounting for 53.0%, and those with over 200,000 followers only accounting for 4.0%.

• KOL followers: Although the average peer follow count in the English list (510.7) is higher than that in the Chinese list (320.2), when adjusted for the base number of AI KOLs in both regions (approximately 1,000 and 5,000 respectively), the penetration rate of top Chinese KOLs reaches 32%, far exceeding the 10% in the English region. This indicates that the Chinese AI KOL circle is a highly dense community with extremely tight connections.

• Activity map: As high as 70% of Chinese KOLs share industry dynamics and practical insights frequently every day. In the English region, low-frequency active accounts account for 39.8%, while stable active accounts account for 26.4%. The English region leans towards an industry source network, while the Chinese region leans towards a practical network.

Summary: English AI KOLs are part of an industry source network that masters first-hand technology and major strategic releases; Chinese AI KOLs are part of a super diffusion and practical network that translates, evaluates, and tutorializes cutting-edge technology for mainstream workflows.

2️⃣ Mental Preferences: English Region Prefers Macro, Chinese Region Prefers Practicality

Peeling away broad labels, we can clearly see the focus of the Chinese and English ecosystems through word frequency and label extraction from the overall discussion content:

Whether in the English or Chinese region, foundational models, AI agents, AI commercialization, and AI programming are common axes of consensus, but the paths they extend outward are entirely different:

• The English region emphasizes underlying technology and macro perspectives: English KOLs have far greater coverage in AI commercialization (44.7%), foundational models (39.6%), AI safety (13.8%), AI chips (12.6%), and embodied intelligence (5%) compared to the Chinese region. They devote a lot of energy to discussing AGI safety alignment, supply-demand patterns of computing power, open-source vs. closed-source debates, and embodied intelligence.

• The Chinese region emphasizes application landing and practical orientation: Chinese KOLs exhibit strong practicality. AI programming reaches 72.1%, nearly double that of the English region. AI agents account for 51.5% vs. 39% in the English region. In visual generation, the data at 20.6% is still close to double that of the English region. Tool evaluations at 11.8% are even more exaggerated, approaching nine times that of the English region. Tutorials and prompts are also significantly higher than in the English region, indicating that Chinese bloggers are better at breaking down complex technologies into specific operational guides like coding and agent building.

3️⃣ Capability Radar: English Focuses on Technical Insights, Chinese Focuses on Full-Stack Applications

To reduce misjudgment of broad labels, we utilized XHunt's KOL capability scoring model to conduct a comprehensive analysis of the content quality published by AI KOL accounts across multiple scoring dimensions:

• The English list occupies the industry source and underlying logic: The highest score in the English list is for multimodal at 88.3, followed by foundational models and prompts. Their core capabilities lie in insights into model architectures, large-scale engineering tuning experience, and predictions of cutting-edge trends. In AI safety and chip fields, the English list has a natural first-mover advantage.

• The Chinese list focuses on full-stack application practices: In the Chinese samples, the average relevance of AI programming capabilities reaches 88.9, while AI agents reach 87.1. A large number of creators with natural language development capabilities are active on Chinese Twitter, adept at monetizing AI in private domains or light entrepreneurial models.

4️⃣ Large Model Mention Rate: Workflow Map of Voting with Feet

The large model mention rate (i.e., the rate at which any tweet from an account within three months hits keywords) not only represents the discussion heat of large models within communities but also serves as a "vote with feet" on KOLs' reliance on various models in actual workflows and their evaluative tendencies:

As shown in the figure, Claude and GPT form the kings of bilingual models. In the Chinese region, the mention rate of Claude reaches 95.7%, still the top choice for independent developers and Vibe Coders; notably, with the continued heat of AI programming scenarios, Codex has recently surged in popularity, maintaining a third place with an 80.9% mention rate, further confirming the fervent pursuit of practical workflows by Chinese geeks.

Additionally, domestic large models DeepSeek (68.1%) and Kimi (58.5%) also demonstrate strong local penetration. In contrast, in the English region, GPT (76.2%) and Claude (75.2%) are evenly matched, focusing more on multimodal evolution and the overall narrative of the industry rather than discussions of single toolchains.

5️⃣ MBTI Content Style: Account Expression Facets

Using a dedicated style inference algorithm, XHunt classified the public personality facets of Chinese and English KOL accounts based on their profiles, long tweet structures, interactive debate logic, and topic preferences into MBTI types:

As shown in the figure, regardless of the Chinese or English region, accounts with a voice tend to belong to the NT (Rationalist) camp. During a period of rapid technological iteration, content that possesses logical analysis and productivity guidance is evidently more favored. The English list is dominated by ENTJ (38.4%) and ENTP (25.8%), leaning towards framework construction and macro analysis; the Chinese list is led by ENTP (41.2%), reflecting the Chinese region's enthusiasm for exploring diverse uses of new tools.

6️⃣ Identity Structure: English Prefers Frontier, Chinese Prefers Practice

By clustering and cross-verifying the profile descriptions and historical tweet narratives of the two sample groups, XHunt has drawn a map of Chinese and English AI KOLs:

Core identity structure:

1. Over 65% of KOLs in the English region are founders (31.4%), executives (34%), or scientists of large models, and their content output itself is a form of strategic distribution.

2. The top accounts in the Chinese region are tool/evaluation experts (69.1%) and product engineers (57.4%). Overall, the English ecosystem leans more towards a source release network, while the Chinese ecosystem focuses more on a productivity practice network.

Overall:

The English AI KOL network resembles a source technology and paradigm release network led by scientists and technology leaders at the forefront of Silicon Valley; the Chinese AI KOL network represents a comprehensive productivity tool and survival practice network led by full-stack independent geeks and application pioneers.

7️⃣ Tweet Validity Evolution: From Wild Growth to High-Quality Development

Combining the trends of the past eight months, the traffic distribution logic in the AI field has shifted from wild growth to high-quality development, with continuous improvements in exposure and tweet volume, presenting three core characteristics:

• Attention dilution in March: Influenced by industry hotspots like Openclaw, the total number of tweets in March surged to 12.4K, with total views reaching 310M, but the average views per tweet dropped to a low (25.0K). The massive amount of homogenized news led to severe information overload and decreased dissemination efficiency.

• Peak of dissemination efficiency in May: The total number of tweets in May fell back to 9.0K, but both total views (335M) and average views per tweet (37.4K) reached historical highs. In-depth practical and evaluative content is leveraging fewer posts to drive greater traffic.

• Views growth rate outpacing tweet production: By the end of May, the growth rate of the views index (+88%) significantly exceeded the growth rate of tweet quantity (+62%). This indicates that the AI traffic dividend is following the 80/20 rule, rapidly concentrating on tweets that output high-quality content at a high premium.

5. Authority Matrix: Global AI KOL Influence Electronic Business Cards

Based on the attention graph network, follower quality, and content quality performance of Chinese and English AI KOLs, we have tailored electronic business cards for top AI KOLs based on their preferred AI tools.

Here are the most representative Top 20 AI KOL business cards from both the Chinese and English regions:

English AI KOL Top 20

1. Andrej Karpathy @karpathy | AI KOL Followers: 1,444 | Followers: 2,358,391 Currently employed at Anthropic's pre-training team, founder of Eureka Labs, former member of the OpenAI founding team, and former head of Tesla AI. A top evangelist who breaks down large model training, AI coding, and agents into digestible pieces, making engineers want to jump out of bed at midnight to open their IDEs.

2. Sam Altman @sama | AI KOL Followers: 1,406 | Followers: 4,741,565 CEO of OpenAI, the absolute master of GPT and Codex, a cosmic figure whose posts automatically highlight key points for half the AI circle.

3. Greg Brockman @gdb | AI KOL Followers: 1,142 | Followers: 968,930 President and co-founder of OpenAI, the person who daily shares hardcore updates on products, research, developer ecosystems, and infrastructure like an official construction log.

4. Ilya Sutskever @ilyasut | AI KOL Followers: 1,069 | Followers: 664,828 Co-founder of SSI and former chief scientist at OpenAI, one of the most watched researchers in the large model era, whose casual remarks can be interpreted for half a day by the entire circle.

5. Jeff Dean @JeffDean | AI KOL Followers: 1,058 | Followers: 436,747 Chief scientist at Google DeepMind/Google Research, head of Gemini, the most hardcore navigator of Google's AI technology roadmap.

6. Elon Musk @elonmusk | AI KOL Followers: 1,057 | Followers: 239,771,643 The super amplifier behind SpaceXAI, Tesla, and SpaceX, demanding everything from models, robots, computing power, and platforms, sometimes causing more explosions with a single tweet than a press conference.

7. OpenAI @OpenAI | AI KOL Followers: 1,050 | Followers: 4,798,535 Official OpenAI account, parent company of ChatGPT and Sora.

8. Demis Hassabis @demishassabis | AI KOL Followers: 1,002 | Followers: 864,498 CEO of Google DeepMind, a super promoter of AlphaFold, scientific intelligence, and AGI narratives, and a Nobel laureate who is making AI truly transition from chat boxes to scientific discovery.

9. roon @tszzl | AI KOL Followers: 968 | Followers: 326,221 Observes the capability boundaries and safety issues of cutting-edge models, often providing sharp and insightful commentary, preferring to be straightforward.

10. Patrick Collison @patrickc | AI KOL Followers: 961 | Followers: 811,554 CEO of Stripe and co-founder of Arc Institute, a macro perspective player who integrates AI into research organizations, infrastructure, and entrepreneurial frameworks.

11. Logan Kilpatrick @OfficialLoganK | AI KOL Followers: 957 | Followers: 304,209 Important disseminator of Google AI Studio/Gemini API ecosystem, pulling developer tools back from magic to practical products, code, and money.

12. Yann LeCun @ylecun | AI KOL Followers: 947 | Followers: 1,144,073 Co-founder and executive chairman of AMI Labs, former chief scientist at Meta AI, one of the three giants of deep learning, a Turing Award winner, and a long-time resident on the debate stage regarding AI routes, open source, world models, and the essence of intelligence.

13. Mira Murati @miramurati | AI KOL Followers: 931 | Followers: 498,732 Founder of Thinking Machines and former CTO of OpenAI, a female leader at the forefront of computing power competition and technology landing, leading the commercialization pains and transformations of cutting-edge models from the lab to mainstream workflows.

14. Garry Tan @garrytan | AI KOL Followers: 892 | Followers: 779,149 A hardcore mentor from Silicon Valley's top incubator, not one to follow trends and shout slogans, he prefers to publicly demonstrate how he uses prompt engineering and personal AI systems to create a truly runnable architecture.

15. Anthropic @AnthropicAI | AI KOL Followers: 884 | Followers: 1,216,610 Official account of Anthropic, developer of the Claude model, focusing on AI safety.

16. Dwarkesh Patel @dwarkesh_sp | AI KOL Followers: 879 | Followers: 221,274 Host of the top tech podcast Dwarkesh Podcast, known for conducting in-depth, high-quality long interviews with core AI scientists and hardcore scholars, regarded as one of the best conversationalists in the global tech community.

17. Alexandr Wang @alexandr_wang | AI KOL Followers: 878 | Followers: 444,108 Founder of Scale AI and head of Meta AI, occasionally releasing macro trends regarding underlying data labeling and traditional AI directions.

18. Andrew Ng @AndrewYNg | AI KOL Followers: 870 | Followers: 1,499,288 Professor at Stanford University, former leader of Google Brain and Baidu AI teams, a ballast and long-term evangelist for AI education and application landing.

19. Aravind Srinivas @AravSrinivas | AI KOL Followers: 836 | Followers: 483,015 CEO of Perplexity, a practical rewriter of search entry points for AI search and answer engines.

20. Jim Fan @DrJimFan | AI KOL Followers: 819 | Followers: 396,349 A core figure in NVIDIA's robotics direction, a top player in embodied intelligence and physical world models.

Chinese AI KOL Top 20

1. 宝玉 @dotey | AI KOL Followers: 559 | Followers: 214,553 A hardcore translation super node in the Chinese-speaking world, not engaging in mystical narratives, relying purely on deep dissection of cutting-edge papers, top interview recaps, and high-quality prompts that are ready to use, firmly establishing the baseline for content quality.

2. Orange AI @oran_ge | AI KOL Followers: 483 | Followers: 170,533 An entrepreneur in the AI field who is both rigorous and sharp in business warfare, always able to extract the most fundamental business philosophy between the evolution of agent architecture and organizational change.

3. 歸藏(guizang.ai) @op7418 | AI KOL Followers: 468 | Followers: 144,288 A "hexagonal warrior" in the independent developer circle, a hardcore player in visual generation and AI programming, helping countless people navigate the pitfalls of tool implementation with a wealth of practical tutorials and sharp evaluations.

4. Bear Liu @bearliu | AI KOL Followers: 453 | Followers: 115,339 A super design geek of the AI era, creatively playing with Vibe Coding, agents, and generative UI, tirelessly exploring how AI can disrupt traditional product development, guiding independent creators.

5. Baye @waylybaye | AI KOL Followers: 452 | Followers: 158,294 A distinctive independent developer benchmark, focusing on practical operations rather than concepts, frequently outputting various AI programming tools' physical confrontations and comparisons, tearing apart the marketing disguises of tools with plain language.

6. 向阳乔木 @vista8 | AI KOL Followers: 441 | Followers: 107,140 A technical hardliner in the Chinese-speaking world, adept at digesting the most hardcore multimodal cutting-edge papers and feeding them to practitioners, using high-density tutorials and technical insights to elevate developers' understanding.

7. Ding @dingyi | AI KOL Followers: 431 | Followers: 151,205 An observer with extremely keen technical and business acumen, pixel-level disassembling of AI programming tools and intermediary ecosystems, always capturing business opportunities that others miss in complex tables or marketing cases.

8. 铁锤人 @lxfater | AI KOL Followers: 569 | Followers: 101.2k Continuously using AI for entrepreneurship, content, and product experiments, having maintained high-star projects; not just shouting "AI changes the world," but already getting hands-on in the mud.

9. Tw93 @HiTw93 | AI KOL Followers: 423 | Followers: 141,827 An extremely low-profile yet high-output representative of independent developers, with a hardcore foundation in large model training and a track record of creating multiple high-value open-source tools, proving strength through code.

10. Yangyi @yangyi | AI KOL Followers: 415 | Followers: 122,284 A commercial hacker who combines technology and making money to the extreme, deeply exploring the monetization limits of AI programming and agents in private domains, while unexpectedly reminding you of the blind spots behind technology from a security researcher's perspective.

11. yetone @yetone | AI KOL Followers: 413 | Followers: 82,680 A hardcore faction in the AI application layer, with deep muscle memory in engineering practices of agent architecture, computer use, and programming tools, relying on high-quality tool replication and evaluations to attract hardcore followers.

12. Mr Panda @PandaTalk8 | AI KOL Followers: 410 | Followers: 74,602 A "super connector" between cutting-edge papers and practical monetization, skilled at translating the most academic agent papers into down-to-earth prompt techniques, clearly seeing the trends of AI commercialization and employment.

13. Dash @DashHuang | AI KOL Followers: 408 | Followers: 113,575 A hardcore cross-border perspective from a large factory founder, deeply exploring the extreme squeezing of AI programming tools in traditional R&D and game development scenarios, providing substantial references for "regular army" implementation.

14. Cell 细胞 @cellinlab | AI KOL Followers: 407 | Followers: 26,667 A fervent evangelist and practitioner of the "One Person Company" model, frequently testing various AI programming tools and visual generation workflows, paving the way for the commercial rise of super individuals.

15. YC (Yucheng) @yucheng | AI KOL Followers: 393 | Followers: 18,728 An entrepreneurial thinker focusing on organizational changes triggered by AI, not only delving into the technical practices of tools like Claude Code but also fascinated by using agent architecture to reshape company operational efficiency.

16. 图拉鼎 @tualatrix | AI KOL Followers: 393 | Followers: 108,450 An "AI evolution sample" of veteran independent developers in the Chinese-speaking world, frequently publicly sharing the real process of reconstructing and developing independent apps using Codex and Claude Code, making the programming capabilities of large models tangible.

17. ruanyf @ruanyf | AI KOL Followers: 384 | Followers: 198,977 A beacon for Chinese developers with a long-standing technical presence, continuously capturing the destruction and reshaping of the traditional software development industry by AI from a keen macro perspective, helping countless people achieve productivity leaps with solid tutorials.

18. 小互 @xiaohu | AI KOL Followers: 379 | Followers: 105,522 A "super intelligence station" and sharp evaluator in the AI tools circle, frequently scanning the latest programming tools and agents across the internet, always able to unearth the most practical skills from monotonous news.

19. 𝗖𝘆𝗱𝗶𝗮𝗿 @Cydiar404 | AI KOL Followers: 378 | Followers: 62,106 A practical product engineer, not discussing abstract grand visions, but focusing entirely on hardcore evaluations of large models like Claude and the life-and-death reviews of their own API projects in business warfare.

20. Frank Wang 玉伯 @lifesinger | AI KOL Followers: 378 | Followers: 36,454 A super individual and long-term entrepreneur in the AI era, breaking down grand agent architectures and product design concepts, openly sharing how he uses the "One Person Company" model to disrupt traditional software development.

6. The Inner Journey of Excellent AI KOLs: Building Trust Through Continuous Validation

Cell 细胞 @cellinlab | Chinese AI KOL Influence Ranking: 15 | Creator Matrix Community Founder

We are about to enter a bountiful era—people will enrich their lives through new channels of creation and self-expression, new paths of self-discovery and belonging, and new ways to engage in meaningful work. We need to redefine work as creation: for a long time, our work has been for survival. In the post-scarcity era, new forms of work mean creation, growth, self-expression, and giving life meaning.

Cuimao @CuiMao | Chinese AI KOL Influence Ranking: 34 | AI KOL

In the AI industry, a "KOL" is not just someone with traffic, but someone who truly participates in building. People recognize me not only because I have created many AI creative videos related to Anthropic but also because they see possibilities that suit them in this content. Some have started creating because of it, some have come to understand the tools, and some have regained their belief that the AI era is not a table for a few, but a new classroom where everyone can take a seat again.

I clearly feel that public influence in the AI era is no longer just about being seen, but about helping more people see their own position. It is not a life-and-death challenge of grabbing chairs like in "Squid Game," but a re-seating for a new semester. Positions will change, orders will change, but everyone still has the opportunity to find their coordinates.

So, if I were to summarize my attitude towards this era in one sentence: maintain passion, learn, create, and share.

Asa @appsail | Chinese AI KOL Influence Ranking: 52 | Partner at @appsail, tutti.so

Initially, I started sharing as part of Build in Public, not to become a KOL. Because I have long been on the front lines of AI going global, global payments, and X operations growth, every path I validate and every pit I step into, I record and share, gradually discovering that I have become an AI KOL.

These shares have also made me increasingly convinced that the era of attention economy has arrived, and everyone, every product, and every organization should actively manage their influence.

In my view, KOLs are more like connectors, linking information and cognition, products and users, and people from different cultural backgrounds. Although AI has made content production more efficient, real experiences, insights, independent judgments, and long-term commitment remain scarce.

It is based on this understanding that we created tutti.so, hoping to help more Chinese enterprises and creators build global influence, allowing good products and good stories to be seen by the world. In the future, what will truly be scarce is not traffic, but trust.

Jason Zhu @GoSailGlobal | Chinese AI KOL Influence Ranking: 71 | Founder of GoSail Lab, AgentSkillsHub

Through exploration, at 31, I received a big gift, jumping from the track into the wilderness, only to realize: KOL is not a persona for me, but a record of real exploration.

My understanding of KOL is twofold:

1. Authenticity is the bottom line: I do not write unverified second-hand content; the gains from flipping shoes, the losses from escape rooms, and the growth of followers from zero are all things I have experienced myself.

2. Leverage is the method: AI accelerates, I steer. Deeply engaging with OpenClaw and Claude Code, I built agentskillshub.top, putting my soul into the engineer's meticulousness. Those who share real experiences in open source are the most scarce products in the AI era.

Gorden Sun @Gorden_Sun | Chinese AI KOL Influence Ranking: 75 | AI KOL

I have been writing an AI news daily for over three years; by persisting in something simple and altruistic, you too can become a KOL. Writing a daily report has allowed me to accumulate the best AI practices across various scenarios, and I have shared almost everything without reservation. Altruism, sincerity, and writing something useful to the best of my ability are the principles I believe in. In an AI era where software products are easier to build, distribution and marketing are becoming increasingly important; everyone should try to share something. This is a compounding matter, and you will not incur any losses.

鱼总聊AI @AI_Jasonyu | Chinese AI KOL Influence Ranking: 84 | AI Going Global KOL

Creating the "Fish General Talks AI" IP initially was just about sharing tools I found useful and pitfalls I encountered. Later, after receiving more and more feedback, I realized that a piece of real practical content can really save others a lot of time and costs.

I have always believed that KOLs do not stand high above teaching others, but first go down to the field, personally trying out new tools and new opportunities, then clearly explaining the truly useful methods. Along the way, my positioning has become clearer: focusing on AI, going global, and products, only sharing content that I have researched, practiced, and that can solve real problems.

AI can improve efficiency, but it cannot replace human judgment and experience. Rather than pursuing rankings, I hope to be a trustworthy and practical source of information in the long term, helping ordinary people truly utilize AI and avoid detours.

得否 @wangdefou | Chinese AI KOL Influence Ranking: 101 | Founder of Defu Technology, Enterprise AI Application Consultant

I have always felt that KOLs are not "content creators," but people who continuously accumulate trust in public domains.

As a liberal arts student, I initially grew my followers on X slowly, taking three years to reach 5,000, and I encountered many pitfalls along the way. What truly helped me grow was not algorithmic mysticism, but authenticity, sincerity, and continuously providing useful content to others.

Now that AI is developing so rapidly, various tools can help us collect information, organize materials, and improve efficiency, but it is ultimately a tool that cannot replace a person's judgment, experience, and expression.

My positioning is simple: to explain AI tools, content creation, and personal commercialization from a liberal arts perspective in a more grounded way. Helping more ordinary people avoid detours while getting to know a group of people who are genuinely doing things is what I find most interesting about being a KOL.

Recently, I have been busy with offline business, and my Twitter operations have become a bit lax.

Star @starzq | Chinese AI KOL Influence Ranking: 262 | Founder of @day1globalpod

In the AI era, everyone is anxious: why can others use large models better and buy stocks ten times over? But I want to say that AI is a super long cycle, and laying a solid foundation at the beginning of the cycle is essential to enjoy more of the era's dividends. I hope my sharing can help everyone gain a deeper understanding of all aspects of the AI cycle and use AI without anxiety.

qinbafrank@qinbafrank | Chinese AI KOL Influence Ranking: 287 | AI Macro Blogger

In fact, I consider myself more of a blogger than a KOL, recording real thoughts and reasoning chains, primarily serving my research and investments, and secondly helping others identify truths and rational thinking amidst the flood of information.

The AI era has greatly accelerated global information integration, code verification, and trend disassembly, requiring us to grasp the core: AI can significantly improve efficiency, but human judgment, experience, thinking chains, logical chains, and analytical frameworks are even scarcer. Focusing on truth-seeking and practical insights, sharing verifiable frameworks and practical thoughts rather than simple conclusions, helps ordinary people truly utilize AI and avoid detours. I hope to be a trustworthy and rational thinking partner in the long term.

XinGPT@xingpt | Chinese AI KOL Influence Ranking: 359 | Former VC Fund Partner, AI KOL

The initial intention was to bring financial equity through AI: ordinary investors can also leverage AI to compete with professional investors. We have developed various tools for AI industry research, AI market tracking and reviews, and AI real-time alerts, and the practical effects are gradually improving. I can't imagine how to complete so much industry research without AI. In the future, we will gradually mature these tools and welcome exchanges with investment and AI experts.

Crypto_Painter@CryptoPainter | Chinese AI KOL Influence Ranking: 490 | AI KOL

Although I am not strictly an AI blogger, the improvements AI has brought me far exceed the past decade…

When you hire a human to do non-physical work for you, you are actually hiring their brain, a neural network made up of billions of neurons, of which 60% of the computing power is unrelated to work…

AI and agents can perfectly replace this role, completing better work with higher efficiency.

So I have been insisting on converting most non-physical work into AI execution, with data analysis done by dedicated agents, quantitative trading monitored by AI, and even tweet inspirations generated from AI-processed content…

My sharing about AI is not purely to follow traffic and attention; it is mainly because I had a baby this year and simply do not have the time and energy to do so many things, and the emergence of AI agents has saved me a lot of time!

This is a magical experience I could never have imagined before, and I sincerely recommend and encourage everyone to try delegating non-physical, simple, and repetitive tasks in their daily work and life to AI; the freedom this brings is true happiness.

Haotian@tmel0211 | Chinese AI KOL Influence Ranking: 570 | Amber Consultant

In fact, being a KOL is about using "output to force learning," accelerating my understanding of various industries and even the upstream and downstream of industries, thus reaping the investment dividends of cognitive realization.

In the past two years, I seized the rapid iteration and hype dividends of the crypto industry's technical narrative, becoming a hardcore tech blogger in everyone's mind. However, in the past six months, as the crypto market cooled, attention has shifted back to the main narrative of AI technology, and I have begun to transition from scratch, embarking on a new journey, thinking and tracking the entire AI technology mainline from a fresh industrial insight perspective, including semiconductors, robotics, chips, storage, etc.

This process has certainly been painful, as many upstream and downstream industries are completely unfamiliar to me. I found that in the process of delving into industries, the things I could output became fewer and fewer, with most of my time spent silently investing and experimenting, then exploring and researching the entire industry with my portfolio. The results were, of course, beyond my expectations, not only yielding unexpected investment profits but also changing my original perception of KOLs; it turns out that what matters is not whether to output or not, but how to refresh one's "input" in a timely manner. Only by ensuring that one always possesses a forward-thinking mindset and think tank can one have the true confidence to be a KOL.

DeFi Teddy@DeFiTeddy2020 | Chinese AI KOL Influence Ranking: 666 | Founder of Biteye/XHunt

In the AI era, becoming an excellent KOL requires understanding and leveraging AI to improve content generation efficiency, such as using AI to collect information and conduct preliminary analysis, and leveraging AI for topic selection.

However, AI is merely our co-pilot, only able to sit in the passenger seat, not control the overall situation. Truly valuable content requires KOLs to generate their unique perspectives and analyses based on AI's analysis, showcasing the KOL's own "soul."

Anita@Anitahityou | Chinese AI KOL Influence Ranking: 895 | Head of Senitent APAC, AI KOL

There are countless KOLs, but those who truly come from personal viewpoints are few and far between. To be a minority, people pay attention to you not to see uniform information, but because you Build In Public or because you can see the essence through phenomena. I am committed to being an ordinary person with independent thinking. Communication is for growth.

7. Paradigm Divide: The Big Showdown Between AI KOLs and Web3 KOLs

AI KOLs and the Web3/Crypto KOLs that have also gained popularity on Twitter in recent years coexist within the social ecosystem of X, but their influence operation axes and survival monetization logic are fundamentally different.

The essence of Web3 KOLs is a network for distributing opportunities and capital. Their influence stems from breaking information asymmetry, amplifying wealth effects, and mobilizing community emotions. Their core assets are "messages" and "call-to-action," with value realization paths typically involving project promotion, token distribution, and community building.

The essence of AI KOLs is a network for distributing productivity and capabilities. Their influence is anchored in verifiable and reproducible real abilities. What they provide is not a get-rich-quick scheme but efficiency tools and usage methods. Their core assets are "trust" and "methods."

In 2026, interesting marginal overlaps are occurring between the two circles: some sharp Web3 accounts are beginning to use AI agents to automatically monitor on-chain movements and batch account creation; some Web3 KOLs are also starting to transition into AI KOLs. But ultimately, the enduring influence of AI KOLs must be firmly anchored in "verifiable real capabilities."

8. In Conclusion: From the Information Age to the Trust Age

In the past twenty years, the core problem the internet solved was: how information spreads.

In the AI era, the fundamental problem it is addressing has become: how capabilities spread.

In this grand transformation, a new scarcity is emerging: trust.

Because AI can generate content, generate code, and generate summaries. But it can never generate real experiences, trial-and-error paths, validation processes, or long-term consistency.

The core of future influence will no longer be the volume of information, the number of followers, or the speed of dissemination, but whether you are continuously and publicly validating the real world and delivering results.

The true discourse power in the AI era does not belong to the strongest expressers, but to those who continuously build trust networks.

Research Statement:

1. All research conclusions, percentage distributions, and scoring results in this report are based on data generated up to June 2026 and do not represent the long-term, fixed influence rankings of each account. Statistical dates: March to May 2026.

2. The identity classifications and MBTI attributes mentioned in this report are inferred based on publicly available tweet behavior characteristics using AI model algorithms, intended to present a macro view of the group's content ecosystem, and do not represent the actual test results or professional identity certifications of individuals in the real world.

About XHunt

XHunt is an industry-leading, AI-driven social intelligence and KOL influence analysis tool. It is dedicated to providing the most authentic and authoritative KOL measurement metrics and account holographic portraits for the global Web3 and AI ecosystems.

• X/Twitter: @xhunt_ai

• Official Website: https://xhunt.ai/