DeepSeekAI: What is it and why is it a threat to the current AI Goliaths?

A new player has emerged that’s causing quite a stir in the tech industry. DeepSeek, a Chinese AI company, has recently launched an AI-powered chatbot that’s taking the market by storm. What does it mean for your portfolio? 

What is DeepSeek?

DeepSeek is an artificial intelligence company based in Hangzhou, China. Founded by entrepreneur Liang Wenfeng and backed by his hedge fund, the company has quickly risen to prominence with its AI chatbot application. The app, which was released in the US on January 10, 2025, has become the most downloaded free app on Apple’s App Store.

What sets DeepSeek apart is its latest AI model, DeepSeek R1, which the company claims is on par with industry-leading models like OpenAI’s GPT-o4, were built for just $6 million, a fraction of the billions spent by US-based AI companies, threatening the demand for Nvidia’s high grade chips

Why is DeepSeek Important?

1. Cost-Effective Innovation: DeepSeek’s ability to develop high-performance AI models at a fraction of the cost of US competitors challenges the notion that massive investments are necessary for cutting-edge AI technology.
2. Open-Source Approach: By making its models open-source, DeepSeek is contributing to the democratisation of AI technology, potentially accelerating innovation in the field.
3. Market Disruption: The company’s success is forcing US tech giants to reevaluate their AI development strategies and justify their massive investments.
4. Efficiency in Computing: DeepSeek’s approach may lead to more efficient AI models, potentially reducing the energy consumption associated with AI development and usage.
5. Global Competition: DeepSeek’s rise showcases the growing capabilities of Chinese AI companies, intensifying the global race for AI dominance.

Market Impact

This development challenges the prevailing notion that massive investments in data centres packed with NVIDIA (NASDAQ: NVDA) chips are essential for developing advanced AI systems. DeepSeek's achievement suggests that companies may be able to create competitive AI models with a fraction of the hardware, potentially reducing the need for Nvidia's expensive GPUs.

This also means we will need less power to run the AI data centers which has rocked the Uranium sector Global X Uranium ETF (NYSE: URA) and utilities providers like Constellation Energy (NYSE: CEG) as the outlook for power hungry AI chips is now uncertain.

Historical Lessons - Cisco, 2001

The comparison is especially timely given DeepSeek’s recent emergence as a low-cost competitor in the AI space, similar to how Cisco faced increasing competition in the early 2000s. DeepSeek’s ability to develop competitive AI models for just $6 million, compared to the billions spent by U.S. companies, mirrors historical patterns of market disruption.

Market Impact

Just as Cisco’s dominance was challenged, Nvidia is now experiencing market pressure, with its stock dropping 17% following DeepSeek’s emergence. This market reaction suggests investors are concerned about potential parallels to Cisco’s historical pattern, particularly regarding the sustainability of high valuations in the face of more cost-efficient competitors.

Where to from here?

We favour that the focus is shifting from mere capacity to capability in a new era of business productivity, focusing on the amplification of human capabilities rather than semi-conductors.

Productivity Gains

The potential productivity gains from AI are substantial. McKinsey estimates that generative AI alone could add $2.6 trillion to $4.4 trillion annually to the global economy.

Companies like Autodesk (NYSE: ADSK) have been investing in AI for a decade and the company's software is widely used across industries that are increasingly adopting AI for productivity gains in engineering, manufacturing, media and construction. 
In the construction sector, the benefits will be felt immediately through waste reduction, with which the industry has a multi-trillion dollar problem.

Data Center demand

While the market has initially sold off listed data centres, we see the barrier to entry being significantly lowered a positive for SME Data centre demand, taking AI from the realm of the Mega-rich, Magnificent 7, to many start-ups and medium sized enterprises that will be more agile in their application of AI to productivity gains and consumer based applications.

Who will be the long-term winners

While the current AI boom has created significant opportunities, it's still too early to accurately determine who the long-term winners will be.

Many of the most impactful AI companies of the future may not even exist yet. Just as Google emerged in the late 1990s to revolutionise internet search, we may see entirely new entrants disrupt the AI space in ways we can't currently imagine.

These future pioneers could emerge from any sector, any country, a mega-cap company or a start-up in a garage to become the next mega-cap. All we know is, like the portfolio rebalances last night in the US, AI will move FAST and unpredictable.