AI30.07.2025

Chinese AI cheat code for South Africans

South Africans who wish to invest in China’s burgeoning artificial intelligence sector should consider the KraneShares China Internet ETF, portfolio manager Grant Nader has said.

Nader, a senior portfolio manager at Benguela Global Fund Managers, recently selected the exchange-traded fund (ETF) as his stock pick on Business Day TV.

He said it was a “little bit of a cheat code” as it offered exposure to several Chinese technology stocks. Tencent, Alibaba, Baidu, JD.com, and Pinduoduo, which owns Temu, are among its largest holdings.

“My view is that the AI revolution — if you want to call it that — in China is kicking off with a vengeance. They’re about a year or two behind the US, but it is going to be gathering steam,” said Nader.

“You want to be able to participate in that, but it’s difficult to pick the winner, and it’s difficult to invest directly in China. This gives you a nice diversified exposure to that broad theme.”

Nader explained that many Chinese hyperscalers and cloud leaders are involved in artificial intelligence research and products that compete with ChatGPT, Gemini, Meta AI, and Grok.

“I think it’s a massive opportunity over the next few years in China. Especially for a DIY investor, this is a great way to play it,” he said.

“You can just participate through the ETF and get a handful of great businesses, and mitigate some of that regulatory risk that comes with China.”

KraneShares said the ETF offers access to Chinese Internet companies that provide similar services as Google, Facebook, Twitter, eBay, and Amazon.

Investors will benefit from exposure to companies listed in the United States and Hong Kong that stand to gain from increasing domestic consumption by China’s growing middle class.

Most U.S. artificial intelligence products are fighting for more mainstream dominance in the hopes of attracting more people willing to pay monthly subscription fees for their premium services.

China has taken a different approach. Its top AI companies are releasing models into the public domain, allowing enthusiasts and rivals alike to download and use them in their own systems.

Although the “open sourcing” of large language models developed by big technology companies at great cost can be traced to the leak of Meta’s original Llama model in 2023, China has embraced the practice.

Google, which is currently trailing in the AI arms race, also started releasing public versions of its language models, dubbed Gemma, in February 2024.

However, Chinese startup DeepSeek made waves in early 2025 when it published models that rivalled the best OpenAI’s ChatGPT had to offer, and which it said were developed at a fraction of the cost.

E-commerce, fintech, and cloud computing titan Alibaba released its new Qwen3 large language model (LLM) in April 2025, which has received wide acclaim among people who run LLMs on their own computers.

Warning about U.S. AI “monopoly”

Solly Malatsi, South Africa’s Minister of Communications and Digital Technologies, who rejects the assumption that the choice between U.S. and Chinese AI technology needs to be binary.

Bloomberg reported on Wednesday that China’s answer to the United States was the launch of a new global organisation that will convene countries to foster safe and inclusive use of AI technology.

The new group, known as the World AI Cooperation Organisation, embodies China’s plan to jostle with the U.S. for sway by positioning itself as a champion of AI for all.

At the annual World AI Conference over the weekend, Chinese Premier Li Qiang warned of AI “monopoly” and called on foreign officials in the room — mostly from developing countries — to cooperate on governance.

“The Chinese are coming to the table with a very different AI product mix,” said Eric Olander of the China-Global South Project. 

Olander predicted China’s approach would be extremely appealing to lower-income countries that lack the computing and power infrastructure needed for large-scale implementation of OpenAI-like systems.

While there are no binding global rules for AI development, China’s plan calls for building more digital infrastructure that uses clean power and unifying computing power standards.

China also said it supported the role of business in creating technical standards in security, industry, and ethics.

Bloomberg reported that Beijing’s emphasis on openness — a word used 15 times in its governance action plan — appears to ride on the success of Deepseek earlier this year.

China also emphasises Internet sovereignty, something the publication said may appeal to more autocratic regimes around the world. 

“We should respect other countries’ national sovereignty and strictly abide by their laws when providing them with AI products and services,” China’s Global AI Governance Initiative said in 2023. 

In contrast,  the US government has vowed to only work with engineers who “ensure that their systems are objective and free from top-down ideological bias.”

The US-China rivalry presents a familiar dilemma for countries that may feel pressured to choose a side, but Solly Malatsi, South Africa’s minister of communications and digital technologies, rejects the binary choice.

“It’s not a case of one model over the other,” Malatsi said from the conference. “It’s about an integration of the best of both worlds.”

Show comments

Latest news

More news

Trending news

Poll

Which is your go-to online retailer for tech products?

View Results

Loading ... Loading ...
Sign up to the MyBroadband newsletter