Moonshot is reportedly preparing to release a larger open-weight version on July 27 and is also said to be preparing for an IPO in Hong Kong in the coming months.

Why does this matter for global financial markets?

US stock markets have been carried higher by a handful of technology companies, their massive AI investment plans and two startups in particular — OpenAI and Anthropic. Their success has echoed across chipmakers, cloud providers, data-centre operators and the broader supply chain, including energy, raw materials, construction companies and even banks helping to finance billions of dollars of AI investment.  The world's largest technology companies have signed tens of billions of dollars' worth of deals with OpenAI and Anthropic through direct investments and strategic partnerships, often tied to multi-year commitments to purchase chips or computing power.  OpenAI and Anthropic have reached extraordinary valuations and are both reportedly considering public listings in the coming months — unless market turbulence delays those plans. Those valuations rest on the assumption that AI spending will continue expanding well beyond today's circle of hyperscalers and into the broader corporate world.

The problem is that Chinese AI models are said to cost up to 100 times less than those developed by OpenAI and Anthropic, while many businesses simply do not need the world's most sophisticated model to run day-to-day operations. Take Airbnb, which recently announced that it would limit employees' token usage because costs had become too high. It can perfectly well operate with a model that is not necessarily the world's best at discovering new medicines — you get the idea.  And, of course, the fact that many of these Chinese models are open-source makes them even easier to adopt.  Then the rest is textbook economics. Lower demand for US frontier models should force US providers to reduce prices to remain competitive. Lower prices would compress revenue expectations, making today's eye-watering valuations increasingly difficult to justify.

The mismatch is already striking. OpenAI is generating roughly $25 billion in annualised revenue against an $852 billion valuation, while Anthropic is on track for around $47 billion in annualised revenue despite being valued at nearly $1 trillion. That means investors are paying roughly 34 times revenue for OpenAI and around 21 times revenue for Anthropic — multiples that leave very little room for pricing pressure or competitive disruption.  Because these two companies sit at the centre of the US AI ecosystem, any downward revaluation could send shockwaves through chipmakers, cloud providers, data-centre operators, construction companies and ultimately the banks that have helped finance much of the AI boom.  Incidentally, many of those banks posted record results last quarter thanks in part to equity and debt issuance linked to AI investment.

Very bad timing

Moonshot AI's moonshot also came at a particularly bad time. Earlier last week, TSMC announced 77% profit growth, yet the stock closed the session after earnings more than 7% lower. VanEck's Semiconductor ETF fell more than 2%, slipping below its 50-day moving average. SK Hynix is down another 2% this morning after attempting to recover earlier losses, while South Korea's Kospi has opened the week with a 4% sell-off after being closed on Friday.

On the other side of the trade, Alibaba is rallying more than 5% this Monday after previewing its latest Qwen model, which is said to be comparable with the world's leading AI models and ‘second only to Anthropic's Fable 5’. The broader Hang Seng is up more than 2%, with Chinese technology stocks outperforming.

Ugly macro setup

Nasdaq futures are calmer, trading slightly higher ahead of the weekly opening bell, but the broader macroeconomic backdrop continues to deteriorate. US crude opened the week near $85 per barrel, while Brent is trading close to $90 this morning following reports that Iranian strikes targeted Kuwaiti energy infrastructure and a desalination plant.  Nor is that the only concern. Ukrainian attacks on Russian energy infrastructure have also been tightening supply for months. Reliable data remain scarce because of Russia's opacity, but the IEA recently reported that since August 2025 at least 100 strikes have hit Russian refineries, with the pace of attacks increasing in recent months. In June alone, at least 10 refinery strikes were reported, and ‘almost every major refinery in western Russia has been targeted by drones’, forcing Russia to curb exports to safeguard domestic energy supplies.

Rising energy prices are pushing global bond yields higher. The US two-year Treasury yield begins the week under renewed upward pressure, as higher oil prices threaten to make last week's softer inflation data quickly obsolete.  This week will bring a fresh round of economic data and technology earnings. Volatility is therefore likely to remain elevated, but not necessarily to the upside. Markets could see another leg lower in both the US and Europe if geopolitical tensions intensify further and/or investors remain focused on AI spending rather than earnings growth.

The key question is whether renewed Middle East tensions could accelerate investors' interest in China's AI ecosystem. Chinese technology companies are not only closing the performance gap with their US peers while competing aggressively on price, but China has also shown remarkable resilience to the Iran-led energy shock. That combination could make Chinese technology an increasingly attractive refuge from both geopolitical uncertainty and growing doubts surrounding the economics of the US AI trade. Remains the question of: are investors ready to stomach broader Chinese risks: unfavourable demographics, weak consumer spending, slowing growth.