Deutsche Bank strategists say China’s AI models are narrowing the capability gap with US peers while being priced closer to mid-tier systems such as Anthropic Sonnet, pointing to a lower cost-to-intelligence ratio. They describe a shift in distribution as Chinese releases increasingly come as open-weight models that can be downloaded, modified and run locally, while US leaders have tended to rely on closed, proprietary products delivered via APIs. The bank links this pricing and accessibility to a reassessment of whether current AI capex plans can hold if comparable performance becomes available more cheaply.
Markets have already reacted, with AI and semiconductor shares under pressure. The Philly semiconductor index fell 9.97% last week and, after a further 1.63% drop on Friday, recorded its biggest weekly decline since the week of the Liberation Day tariff announcements last year. Deutsche Bank adds that cheaper, more widely available models could accelerate global AI adoption and support productivity, but may compress US AI capex returns and margins, encourage faster commoditisation, and intensify geopolitical fragmentation as rival technology stacks develop.
Opportunities from the US-China AI Valuation Gap
We see a prime opportunity for derivative traders to exploit the growing valuation gap between US AI infrastructure giants and highly efficient Chinese open-weight models. Historically, when disruptors slash intelligence costs—similar to the market shakeups we saw with ultra-low-cost Chinese model releases earlier this year—volatility in semiconductor indices spikes. We recommend utilizing near-term put options on major US chip manufacturers to hedge against immediate downside risk as valuations are reassessed.
Recent market data shows the Philadelphia Semiconductor Index (SOX) has already faced intense pressure, suffering a sharp 10% drop in a single week as investors question the returns on massive AI capital expenditures. To capitalize on this trend, we should look at trading bear put spreads on tech-heavy ETFs like the VanEck Semiconductor ETF (SMH). This approach allows us to profit from the downward momentum in overvalued chip stocks while keeping our capital risk defined.
Adjusting Strategies for Global Technology Volatility
We must also prepare for a potential capex overcycle as US tech giants struggle to justify their massive hardware spending against cheaper, open-source alternatives. Using long-dated straddles on major US cloud and software providers will allow us to capture massive price swings, regardless of which way the market breaks. This strategy is crucial for the coming weeks as geopolitical shifts and commoditisation pressure global technology margins.
