Visionary Leaders
They measure the business empire from the commanding heights of the era.
Yang Zhilin
“General artificial intelligence is the biggest opportunity of this generation, and we must be the explorers of AGI.”
From a top student at Tsinghua University to the head of an AI unicorn, he created one of the most powerful AI models in China
Yang Zhilin
Asian Cover Figure | Yang Zhilin: From Tsinghua Scholar to AI Unicorn Commander
Asian "Cover Figure" Magazine · Vision
In 2023, as the large model wave swept global tech, a 26-year-old Chinese youth quietly stepped onto center stage. Yang Zhilin — Tsinghua University computer science graduate, Carnegie Mellon University PhD — founded Moonshot AI, launching the intelligent assistant Kimi, renowned for its ultra-long context processing capability. Within less than two years, the company's valuation surpassed several billion dollars, making it one of China's most watched AI unicorns.
Yang Zhilin's story is a microcosm of China's new generation of AI entrepreneurs — possessing top academic backgrounds, global vision, and unwavering belief that technology can change the world. Between academia and industry, he found his unique coordinate.
I. Academic Path: Elite Training from Tsinghua to CMU
Yang Zhilin's academic trajectory is exemplary. During his Tsinghua computer science studies, he demonstrated extraordinary talent in AI. As an undergraduate, he published multiple high-level papers in natural language processing, attracting international academic attention.
After his bachelor's, Yang Zhilin pursued his PhD at Carnegie Mellon University (CMU), studying under machine learning luminary Professor Tom Mitchell. CMU is a global AI research stronghold; here, Yang Zhilin deeply researched language model core architectures and training methods.
His representative research contributions include Transformer-XL and XLNet. Transformer-XL solved standard Transformer's context fragmentation in long-sequence modeling by introducing recurrence mechanisms, laying technical foundations for subsequent long-text processing. XLNet overcame BERT-type models' pre-training and fine-tuning inconsistency through permutation language modeling, achieving leading results on multiple NLP benchmarks. These papers, published at NeurIPS, ICML, and other top venues, have been cited thousands of times, profoundly impacting large model technology development.
II. The Entrepreneurial Choice: From Academic Star to Business Leader
After his PhD, Yang Zhilin faced a critical life choice. With his academic record, he could easily continue at a top university or research institution. But he chose another path — entrepreneurship.
In early 2023, the global AI fever ignited by ChatGPT showed Yang Zhilin the window for technology moving from lab to mainstream. He keenly judged: large language models were about to enter the era of mass application, and the Chinese market urgently needed excellent localized solutions.
In April 2023, Yang Zhilin founded Moonshot AI. The company name carries deep meaning — "Moonshot" represents relentless pursuit of ambitious goals. He assembled a core team of technical talent from Tsinghua, CMU, and other top institutions, combining academic depth with engineering capability.
III. Kimi: The Ultra-Long Context Breakthrough
In October 2023, Moonshot AI launched the intelligent assistant Kimi. Unlike the many large model products on the market, Kimi's core differentiation was ultra-long context processing — supporting a 2-million-character lossless context window, meaning users could input entire books, complete codebases, or massive document sets for Kimi to understand, analyze, and summarize.
This technical breakthrough directly stemmed from Yang Zhilin's academic research. Transformer-XL's exploration of long-sequence modeling provided key technical insights for large context windows. In engineering implementation, the Moonshot team innovatively adopted Grouped Query Attention (GQA) and other optimizations, greatly expanding context length while maintaining inference speed.
Kimi quickly went viral upon launch. Within months, registered users exceeded ten million, with daily actives at the million level. On social media, Kimi was widely used for paper reading assistance, legal document analysis, long-text summarization, code debugging, and more. Its arrival gave ordinary users their first experience of large models' powerful long-text understanding capability.
IV. Technical Strength and Funding Performance
Moonshot AI's technical strength gained broad industry recognition. The company's self-developed MoE (Mixture of Experts) architecture large model performed excellently in multiple evaluations, particularly leading in long-text understanding and reasoning.
In capital markets, Moonshot also performed brilliantly. Within months of founding, the company received investment from Sequoia China, ZhenFund, and other prominent firms. In 2024, the company completed several hundred million dollars in funding, with post-money valuation exceeding $3 billion, becoming one of the fastest-growing AI startups globally. Investors universally favored the Yang Zhilin team's technical strength and commercialization capability.
V. Vision: Making AGI Serve Everyone
Yang Zhilin's ultimate goal is achieving Artificial General Intelligence (AGI). In multiple public forums, he stated AGI is not distant science fiction but a gradually achievable technical goal within foreseeable future. He believes the key to AGI lies not in simple parameter accumulation but in deep understanding of intelligence's essence and correct technology route selection.
On the commercialization path, Yang Zhilin insists on "technology-driven, product is king." He believes large model companies' ultimate core competitiveness lies in whether they can provide users with genuinely valuable product experiences, not just benchmark rankings. Moonshot AI will continue investing in fundamental research while accelerating product iteration, ensuring AI technology truly benefits broader user groups.
In Yang Zhilin, one sees the typical characteristics of China's new-generation AI entrepreneurs: deep academic foundation, broad international vision, firm belief in technology changing the world, and pragmatic ability to transform frontier research into commercial products. From Tsinghua scholar to AI unicorn commander, his ascent continues. In this era of AI redefining everything, Yang Zhilin and his Moonshot AI are writing a new chapter for Chinese AI entrepreneurs.
Follow
Scan to follow for more stories
Cover ArchiveArchive
View archive file →Every cover publication is registered and kept in the ACF Cover Archive. The code is permanent.
Not Everyone Makes the Cover
Asia Cover Figure · Influencing Those Who Influence