The rapid rise of Chinese artificial intelligence developers such as DeepSeek, Z.ai, and Moonshot AI has triggered a global tech market sell-off and forced American tech giants to rethink soaring infrastructure spending. According to IG Market analyst Tony Sycamore, recent developments show that Chinese AI and infrastructure match some US frontier models, shaking Wall Street and international markets.
Market Volatility and the Cost of AI Spending
Global financial markets have experienced sharp turbulence due to questions surrounding whether hundreds of billions of dollars in tech sector AI investments will yield adequate returns. According to market data, the Nasdaq 100 briefly entered correction territory, while South Korea’s Kospi index fell as much as 38 percent in July before staging a partial recovery. Analysts attribute this pressure directly to the emergence of lower-cost competitors from China. “The broad thinking was the US was about six to nine months ahead of Chinese competitors, but after developments recently, we now know that Chinese AI, Chinese infrastructure, are as good as some of the US frontier models,” Tony Sycamore said. He noted that the pullback is felt acutely in South Korea and could eventually impact Australian superannuation returns if US tech markets continue to decline.
Shifting Economics for Australian Businesses
Fierce international competition is rapidly changing the economics of artificial intelligence for Australian startups and enterprises building AI into supply chains. Jacky Koh, co-founder and co-CEO of Sydney-based startup Relevance AI, stated that his firm is increasingly shifting traffic toward open-weight models that can be downloaded and operated locally. Relevance AI builds autonomous AI agents for major clients like Canva, KPMG, and Autodesk. Koh noted that traffic on open-weight models has jumped from roughly 5 or 7.5 percent at the start of the year to between 20 and 25 percent today. While self-hosting allows businesses to bypass per-use developer fees, companies must still cover the underlying computing power required to run the systems.
By utilizing top-tier proprietary large language models for high-level planning and routing simpler, repetitive subtasks to cheaper open-weight alternatives, companies can achieve massive operational savings.
Model Routing and Token Cost Savings
As corporate deployment shifts from basic chatbots to multi-step autonomous AI agents, token consumption rates climb dramatically. Deepak John Joseph, who operates the product engineering business Pupa Clic Technologies, reported a surge in client inquiries driven by soaring token bills. According to Joseph, some customers consuming up to $20,000 worth of tokens daily are seeking drastic reductions. “We are looking at almost 80 per cent of savings in using hosted models in compared to the cloud hosted variants,” Joseph said. To manage these expenses, Sydney-based advisory firm Amplify AI advocates for model routing. Will Liang from Amplify AI explained that businesses reserve expensive systems for difficult jobs while sending simpler tasks to cheaper models.
Security Concerns and Regulatory Barriers
Despite the financial advantages, Chinese-developed AI models face scrutiny regarding political bias and security risks. UNSW AI Institute chief scientist Toby Walsh pointed out that models developed in China generally avoid unfavorable commentary regarding the Chinese regime or topics like Tiananmen Square. Furthermore, Will Liang noted that a lack of trust and regulatory uncertainty remain major adoption barriers. Deepak John Joseph confirmed that there is pushback to using Chinese-developed AI, especially in healthcare as well as on the finance-based domains. Trivium China technology partner Kendra Schaefer cautioned that downloading open-weight models locally does not eliminate cyber risks. “Like any locally installed software, a locally installed model is a vector for compromise,” Schaefer said, emphasizing that organizations must invest in purpose-built safety protocols and access models through trusted domestic cloud providers or secured on-premises hardware.
Did You Know? Tokens serve as the fundamental currency of the AI economy — every question AI reads and answers uses them.
Frequently Asked Questions
Why are Chinese AI models impacting global stock markets?
Investors fear that cheaper Chinese alternatives will undermine the return on investment for hundreds of billions of dollars spent by US technology giants, leading to recent corrections in indexes like the Nasdaq 100 and South Korea’s Kospi.
What is model routing in artificial intelligence?
Model routing is an architecture strategy where an organization uses expensive, high-tier large language models for complex planning tasks while delegating simpler operational steps to cheaper, open-weight models to reduce token costs.
Do open-weight models eliminate cybersecurity risks?
Related reading