What is DeepSeek and Will It Matter?
Artificial intelligence (AI) is evolving rapidly, with companies in the U.S. and China competing for dominance in developing large language models (LLMs).
Among the most notable players is DeepSeek, a Chinese AI firm that has stunned the world by producing state-of-the-art models with significantly lower costs and efficiency improvements.
DeepSeek’s latest reasoning model and LLM, DeepSeek v3, is proving to be a game-changer, demonstrating that China is now only weeks behind its American counterparts, such as OpenAI and Google. The firm’s efficiency-driven approach challenges the AI industry’s traditional development model, making high-quality AI more affordable and accessible.
What is DeepSeek?
DeepSeek is an AI research firm spun off from High-Flyer, a Chinese hedge fund that originally used AI for financial modeling. Unlike major Chinese tech giants like Alibaba and Tencent, DeepSeek was not originally an internet company but rather an AI-first organisation focused on developing cutting-edge models for general use.
The company emerged as a major player in 2024 when it released DeepSeek v3, a large language model (LLM) with 685 billion parameters—making it larger than Meta’s Llama 3.1 and one of the most powerful AI models available.
DeepSeek also launched DeepSeek R1, a reasoning model that competes directly with OpenAI’s o1 and Google’s Gemini Flash Thinking. Reasoning models are a new frontier in AI that allow models to “think” step-by-step before answering complex questions, making them more accurate for scientific and mathematical reasoning.
Despite U.S. efforts to restrict China’s access to advanced AI chips, DeepSeek has found innovative ways to train models efficiently, drastically reducing costs while maintaining competitive performance.
How is DeepSeek Different from other AI Models?
Cheap and Efficient
DeepSeek has redefined AI economics by proving that cutting-edge models do not require massive investments. Training DeepSeek v3 cost under $6 million, compared to the tens of millions spent by U.S. firms like OpenAI and Meta according to The Economist. The model was trained using just 2,000 lower-quality chips, while Meta’s Llama 3.1 required 16,000 high-end chips. Despite using older or downgraded chips due to U.S. export bans, DeepSeek still matched or outperformed some Western models. This efficiency-first approach challenged the traditional belief that only firms with enormous resources can develop frontier AI models.
Unlike OpenAI and Google, which keep their top models closed, DeepSeek follows an open-source model, making its AI widely available. It aligns more with Meta’s strategy, but DeepSeek has taken openness even further by releasing more research details than any Western company. This approach democratises AI development, allowing more companies, researchers and developers to innovate on top of DeepSeek’s models. By fostering collaboration and transparency, DeepSeek is accelerating global AI progress while positioning China as an AI hub.
How is DeepSeek Different from other AI Models?
Low Running Costs
Not only is DeepSeek’s model cheaper to train, but it is also cheaper to operate, making AI services more affordable. Unlike OpenAI and Google, which charge premium prices for model access, DeepSeek plans to offer its AI services for a fraction of the cost. By optimising how it splits tasks across chips and minimizing redundant calculations, DeepSeek maximises efficiency and reduces computational waste. This means businesses, researchers and developers can access high-quality AI without massive financial barriers, further disrupting the AI industry.
How is DeepSeek Different from other AI Models?
DeepSeek is Reasoning-Able
DeepSeek is one of the first companies to successfully implement reasoning AI, a new frontier that improves problem-solving skills. Traditional AI models generate responses based on probability, often making mistakes when handling complex problems. Reasoning models “think” step by step, analysing different possibilities before answering, improving accuracy in maths, science and logic-based queries. OpenAI pioneered reasoning models with o1, followed by Google’s Gemini Flash Thinking but DeepSeek was one of the first to replicate this technology, closing the innovation gap between China and U.S models.
Why DeepSeek Matters
DeepSeek’s rise challenges the dominance of OpenAI, Google, and other Western AI firms. By proving that AI development can be cheaper and more efficient, DeepSeek is shifting the industry’s economic model. If high-quality models become widely available at lower costs, AI will become more accessible to smaller firms, startups and researchers. According to The Economist, China is now only weeks behind the U.S. in AI innovation, meaning American companies can no longer assume they have a lasting lead. With reasoning models, cost efficiency, and an open-source philosophy, DeepSeek has positioned itself as a serious global AI contender.
More than Another AI Model
DeepSeek is more than just another AI model—it represents a fundamental shift in AI development. By delivering powerful, open-source AI models at a fraction of the cost, DeepSeek is proving that Western dominance in AI is no longer guaranteed. Its breakthroughs in reasoning AI, cost efficiency, and training methods are forcing competitors to rethink their strategies. If DeepSeek’s approach continues to gain traction, the AI industry could move away from expensive, closed-source models toward affordable and widely accessible AI systems, benefiting businesses, researchers and users worldwide. In short, DeepSeek isn’t just catching up – it’s changing the game.
Digital Literacy for Leaders
Bill Owens, Managing Director of Veracity, has decades of experience in global business consulting and technology and is committed to raising digital literacy levels of business leaders across Australia. Bill often presents to Boards and senior leaders on data governance, privacy, AI, and cybersecurity to help leaders feel at ease discussing and making decisions about tech and IT.
Continue the conversation on LinkedIn at Digital Literacy for Leaders.
About The Author
With decades of experience in global business consulting and technology, Bill excels in governance, strategy development, risk management, and financial management. He serves as the non-executive Chair of Relationships Australia Queensland and is a member of the Gymnastics Australia Commercial Committee. Additionally, Bill contributes to the Tech Council of Australia and acts as a Technology Industry Expert for Queensland Leaders. He is a graduate of the Australian Institute of Company Directors course.
