What investors can learn from the DeepSeek market frenzy
By Ryan Johnson
Chinese startup DeepSeek sent markets into a frenzy this week with the launch of its new AI chatbot that rivals its US counterparts in performance for a fraction of the cost.
The tech-heavy Nasdaq 100 index fell nearly 3% on Monday, with Nvidia alone erasing almost $1 trillion in market value-a record-breaking single-day loss in market capitalisation.
To put that in perspective, that's more than market caps of the Commonwealth Bank ($267 billion), McDonald's ($333 billion), and Samsung ($394 billion) combined, lost in 24 hours.
In the same way that OpenAI burst into the public consciousness, DeepSeek has done the same thing albeit with a twist, says Betashares senior investment strategist, Cameron Gleeson.
"The bang was the same with the difference being that it was built by a small AI team in China via an open-source model and 30 times cheaper - three key differences that have the potential to disrupt several market narratives in the coming weeks."
What is DeepSeek and why is it shaking up AI?
Founded in 2023 and based in Hangzhou, China, DeepSeek is an AI company specialising in large language models.
Its AI assistant, R1, is a generative chatbot similar to OpenAI's ChatGPT or Google's Gemini, designed to generate responses using vast datasets.
What sets DeepSeek apart is its affordability. Reports show that R1's platform charges just US$0.55 per million input tokens (the text you provide when asking a question), compared to ChatGPT's $15. On the output side-what the AI generates in response-R1 costs $2.19 per million tokens, while ChatGPT charges a steep $60.
This makes DeepSeek a significantly cheaper alternative while delivering comparable or even superior results, all achieved using less expensive hardware and software.
While concerns exist over its privacy and ability to answer certain queries, the platform raises critical questions about the rest of the AI industry and the companies that help build it.
"Even if DeepSeek does not maintain its current level of popularity, this development serves as a reminder that competition in the global AI arena is intensifying, and Nvidia may not be in the pole position forever," says Saxo chief investment strategist Charu Chanana.

Nvidia: A lesson in volatility
Across the market, there was collateral damage across the semi-conductor supply chain (TSMC, ASML) and even energy sector (Siemens) - as markets digested the news.
However, it was Nvidia, a US chipmaker and the world's third-largest company in terms of market cap, that experienced the biggest shakeup.
Gleeson says it's a timely reminder of the volatility of Nvidia stock, which is up 90% over the past year. In fact, Nvidia makes up eight of the top 10 biggest daily stock crashes.

"The US market has become reliant on big tech to drive returns in the last few years. It's perhaps a little early to call an end to the AI market narrative, but investors should be mindful that the US market is far broader than just big tech," says Gleeson. "For example, the S&P500 on an equal weight basis was flat, despite the mega sell off in Nvidia."
"As part of this changing of the guard, US financials are having a very strong Q4 2024 earnings season, with expected 28% earnings growth."
While Nvidia's stock may have fallen off a cliff, its stock price rallied 8.8% on Tuesday - recovering around $400 billion - about the value of Coca-Cola.
At the end of the day, Gleeson says diversification remains an important bulwark for investor portfolios while also considering the longer-term market implications.
"The development of more efficient AI training, is at face value, a net negative for the hardware (semiconductor industry) and longer term could be a very strong net positive for software," he says.
For example, Meta and Salesforce has experienced market gains in the past two days.
"End-user AI application development becomes cheaper and therefore has wider applications, so anyone who owns the customer by providing AI embedded solutions wins," Gleeson says.

Five upshots for investors
Saxo chief investment strategist Charu Chanana lists five key upshots of DeepSeek's model DeepSeek-R1 for investors:
- If DeepSeek's cost-efficient AI model gains traction, high costs and stretched valuations could put US AI leaders under pressure.
- DeepSeek's success could spark renewed interest in China's relatively undervalued AI sector.
- Investors should consider broader diversification within and beyond AI, including across sectors including renewable energy, semiconductors, and healthcare.
- Geographic diversification remains key for all investors over the longer term.
- If US export controls fail to restrict Chinese innovation success, expect US policymakers to explore new tactics in this space.
"Even if DeepSeek does not maintain its current level of popularity, this development serves as a reminder that competition in the global AI arena is intensifying, and Nvidia may not be in the pole position forever," says Chanana.
Get stories like this in our newsletters.