Dreams Can Come True

From Humble Beginnings to AI Visionary

Liang Wenfeng's story is a rare tale that reminds us that dreams aren't just for dreamers—they're for doers. Born in 1985 in Zhanjiang, Guangdong, Liang grew up in a modest household. His father, a primary school teacher, instilled in him the value of education and hard work. But little did anyone know that this boy with a knack for numbers would one day become a revolutionary force in artificial intelligence.

Liang's journey began with his exceptional talent in mathematics. He eventually earned a master's in electronic engineering and AI from Zhejiang University. But this was just the beginning. Liang wasn’t just good at math; he had a vision to change the world through AI.

The Birth of a Dream: High-Flyer Quant

In 2015, Liang took his first big leap by founding High-Flyer Quant, a hedge fund that used AI-driven algorithms for trading. Within four years, the company managed assets worth a staggering $8–10 billion. But here's the twist: Liang wasn’t satisfied. While the world saw him as a successful entrepreneur, his heart remained in fundamental AI research. He wanted to create something bigger that could genuinely change the game.

A Bold Move: Stockpiling Chips and Building DeepSeek

In 2021, when the U.S. announced export restrictions on chips, most people panicked. But not Liang. He saw the crisis coming and made a bold move—he stockpiled 10,000 NVIDIA A100 GPU chips. While the rest of the world struggled with an AI hardware shortage, Liang's courage and foresight put him two steps ahead.

In 2023, he spun off High-Flyer's research division and founded DeepSeek. His goal? To transform China from a nation of imitation to a hub of original AI innovation. Liang wasn’t just building a chatbot but a future where China could stand tall in the global AI race.

Breaking Barriers: The Open-Source Revolution

Liang's first breakthrough came with DeepSeek's R1 model. Unlike other AI models locked behind closed doors, Liang made R1 open-source and allowed free commercial use under the MIT license. His philosophy was simple: real progress happens through open collaboration, not closed systems.

And boy, did it work! In January 2025, DeepSeek's chatbot launched in the U.S. and became the most downloaded app in 50 countries within ten days. The impact of Liang's open-source revolution was so massive that it outperformed OpenAI's GPT-4 and caused NVIDIA's stock to drop by 18%. Talk about shaking up the tech world!

A Unique Approach to Innovation

His unique approach to innovation made Liang's success even more remarkable. Instead of following the traditional corporate playbook, he did things differently. He hired fresh graduates, eliminated KPIs and rigid hierarchies, and gave researchers unrestricted access to GPU clusters, a strategy that piqued the interest of the tech world.

This bottom-up innovation model allowed DeepSeek to develop its V3 models for just $5.58 million—a fraction of OpenAI's GPT-4 cost. In May 2024, DeepSeek slashed API costs to 1/53rd of OpenAI's Cloud 3.5, forcing industry giants like Alibaba and Tencent to lower their prices. Liang wasn't just playing the game; he was rewriting the rules.

The Road to AGI: Liang's Ultimate Vision

Liang's ultimate goal is to achieve Artificial General Intelligence (AGI)—a form of AI that can perform any intellectual task a human can. He believes that language, mathematics, and multimodal intelligence are the key pillars to achieving this. He's confident that AGI could become a reality within 2 to 10 years, especially with China's young scientists leading the charge.

Despite U.S. chip restrictions, DeepSeek has proven that it can train models on just 2,000 chips, compared to competitors who require 16,000 chips. This groundbreaking efficiency is a testament to Liang's vision and ingenuity.

A Self-Funded Powerhouse

One of the most fascinating aspects of Liang's journey is that he never took venture capital funding. DeepSeek thrives on High-Flyer's financial backing and philosophy of long-term research, keeping it free from market pressures. When DeepSeek V3 stunned experts worldwide, NVIDIA scientist Jim Fan called it "the biggest black horse in the AI world."

DeepSeek: A Symbol of China's Technological Independence

Today, DeepSeek is more than just a company—it symbolises China's technological self-sufficiency. As Professor Marina Zhang from the University of Sydney says, "Overcoming hardware shortages through software innovation—that's DeepSeek's real achievement."

Liang Wenfeng's story powerfully reminds us that curiosity, courage, and vision can overcome any resource constraints. He firmly believes, "As long as China copies the U.S., it will remain second best. Now is the time to create something original—so that the world copies us."

So, what's next for Liang and DeepSeek? Only time will tell. But one thing's for sure: this is just the beginning of a thrilling journey inspiring the world to dream bigger and bolder.