Liang Wenfeng: The Visionary Behind DeepSeekās AI Rise
In recent years, Liang Wenfeng has emerged as one of the most influential figures shaping the AI space.
His position at No. 5 on AI Magazineās list of the Top 100 Leaders 2026 reflects the global impact of his ongoing work at DeepSeek.
In a sector dominated by Silicon Valley giants, Liang has helped redefine what is possible in frontier AI development, proving breakthrough innovation is no longer solely dependent on billion-dollar budgets or the largest GPU clusters.
Through DeepSeek, Liang has sparked a shift in assumptions around cost, accessibility and open innovation, positioning himself at the centre of one of the most disruptive AI stories of the decade.
To see the Top 100 list in full, click here.
Smart architecture and optimisation
Liang Wenfeng is not a traditional tech founder.
Launching DeepSeek in 2023, he brought an unconventional perspective to AI leadership, backed by the Chinese hedge fund High-Flyer. His approach combined financial discipline with research ambition, prioritising algorithmic efficiency over brute-force compute.
While many competitors poured hundreds of millions into large language model training, Liang focused on achieving comparable performance through smarter architecture and optimisation.
This mindset has helped to transform DeepSeek from a new entrant into one of the most closely watched AI labs in the world, signalling the growing influence of Chinese firms in the global AI race.
Liangās ability to fuse strategic investment thinking with technical innovation has become a defining part of his leadership style.
DeepSeekās rapid rise
DeepSeek has quickly garnered international attention for building low-cost, high-performance open-source models that rival established players such as OpenAI and Anthropic. Its AI assistant climbed to the top of Appleās US App Store within a week of launch, underlining the speed of its adoption.
DeepSeekās models are particularly strong in mathematical reasoning, scientific analysis and software development, with variants designed specifically for chat and coding tasks.
Despite operating under export restrictions that limit access to the most advanced chips, the company has pioneered training methods using less powerful hardware, demonstrating that efficiency can outperform scale.
This disruptive momentum has unsettled markets and challenged long-held assumptions about competitive advantage in AI.
Pioneering AI innovation
At the heart of DeepSeek’s breakthroughs is Liang’s insistence that frontier AI can be built differently.
DeepSeek-V3, trained on 14.8 trillion tokens for just US$6m, used only a fraction of the compute typically required for models at GPT-4 level. By comparison, OpenAI is estimated to have spent around US$100m training GPT-4.
Perhaps most significantly, DeepSeek-V3 was released under the permissive MIT License, reinforcing Liang’s commitment to research openness. His belief is that the future of AI will be shaped less by who has the biggest infrastructure and more by who designs the most efficient systems.
Industry leaders are taking notice. Alexandr Wang, CAIO at Meta, previously described DeepSeek as “on par with the best American models,” while Marc Andreessen, Co-Founder of venture capital firm Andreessen Horowitz, has called DeepSeek's R1 model as "AI's Sputnik moment".
Ultimately, Liang Wenfeng’s influence lies in proving that innovation, not spending, will define the evolution of artificial intelligence.
The Top 100 AI Leaders 2026
The Top 100 AI Leaders 2026 recognises the individuals driving real-world impact through artificial intelligence, from scaling enterprise adoption to advancing responsible and ethical AI.
Our inaugural leaders list highlights the experts and executives who are shaping how AI is developed, governed and deployed across industries including technology, finance, healthcare, manufacturing and energy.
From CEOs embedding AI into core strategy to researchers and innovators pushing the boundaries of what’s possible, the Top 100 celebrates those defining the AI era.
To see the Top 100 list in full, click here.
