DeepSeek R1 — How a $6M Chinese Model is Challenging the $157B American AI Giant
China’s New AI is Faster, Cheaper, and Raising Eyebrows.
Hey System Thinker,
A whale just dove into the AI ocean 🐋.
This week, a major splash hit the AI world—DeepSeek arrived, and it’s making waves.
Outstanding performance. Economical costs. Nimble execution.
Here are 5 things you need to know about DeepSeek, the new (and free) AI reasoning model shaking up the industry.
1/ Where Did It Come From?
DeepSeek is a Chinese AI company based in Hangzhou, Zhejiang 🇨🇳, fully funded by the High-Flyer hedge fund.
Their mission? Advance AI research and development, creating models that compete with or surpass those built by U.S. tech giants.
Their approach is simple: make powerful AI more accessible and affordable.
2/ How Did It Happen?
DeepSeek’s rise was fast and calculated:
April 2023 – High-Flyer announced an independent AI research group.
May 2023 – DeepSeek was officially launched.
May 2024 – DeepSeek-V2 sparked a price war in China’s AI market.
January 2025 – DeepSeek-R1 dropped, challenging U.S. dominance in AI.
Fun fact: Founder Liang Wenfeng started as a hedge fund trader during the 2008-2009 financial crisis.
3/ What’s the Breakthrough?
DeepSeek is competing on performance, cost, and accessibility by leveraging:
Innovative architecture – Uses MoE (Mixture of Experts) to activate only the necessary neural networks, cutting computational costs.
Efficient resource use – Achieves top-tier AI performance on a fraction of the budget.
Open-source philosophy – Making AI more accessible for global research and development.
Top-tier benchmark scores – Competes with OpenAI and Google in coding (HumanEval) and problem-solving (GSM8K).
Despite limited hardware and U.S. export restrictions, DeepSeek is proving that software innovation alone can create world-class AI models.
4/ Why Is It Controversial?
DeepSeek directly challenges the U.S. AI stronghold, and that’s making waves:
Performance Parity – DeepSeek-R1 matches or beats OpenAI, Google, and Anthropic in key AI benchmarks.
Cost Efficiency – DeepSeek reportedly built its model for just $6 millions in compute costs, a fraction of what U.S. companies spend.
Resource Optimization – Achieving state-of-the-art results without bleeding-edge hardware.
Just to put things in perspective: OpenAI is valued at $157 billion—26,000 times the size of DeepSeek.
5/ The Bigger Picture
DeepSeek’s rapid progress raises serious questions about the future of AI dominance:
Shifting AI Power – Some experts believe the “center of gravity” in AI is shifting towards China.
Economic Implications – U.S. companies may struggle to compete with China’s ability to deliver high-performance AI at lower costs.
National Security Concerns – Advanced AI capabilities in China are raising red flags in U.S. policy circles.
Online discussions are heating up over reports that AI models shut down when asked about sensitive political topics related to China, sparking concerns over embedded biases.
Final Thoughts
DeepSeek isn’t just another AI model—it’s a wake-up call.
Its rise is forcing a reassessment of U.S. AI strategy, questioning whether policy restrictions and funding models are enough to maintain technological leadership.
Nonetheless, tough week to be the CEO of a non-profit.
Anyway, I digress…
Cheers,
Thiago V Ricieri
System Thinker @ Systematic Success
Maker @ Making of a Maker
Engineering Manager, Apps @ Pluto TV / Paramount Global
Founder @ Ghost Ship & Co.
Digital Nomad @ Instagram, Threads, X.com, LinkedIn, GitHub, Website
PS... I started Systematic Success to share the secrets I’ve accumulated over the past decade—leading teams, working remotely, and building proven systems. Turns out, AI is one of the best real-world applications of system thinking. Expect more AI deep dives in future editions.
PPS… Ever feel like AI is moving too fast to catch up? It’s not as daunting as it seems. With just 20 hours of focused effort, you can master the basics of AI agents and start building impactful tools. The gold rush is here—don’t miss your chance.