Moore Threads shares jump on swinging to profit in Q1

Moore Threads swung to a Q1 profit of 29.4 million yuan from a 112.5 million yuan loss a year earlier, while revenue surged 155% year-on-year to 737.6 million yuan on strong AI-related chip demand. The Shanghai-listed GPU maker also said it is scaling production as China pushes semiconductor self-sufficiency and U.S. export restrictions limit Nvidia’s advanced AI chip sales to China. Shares jumped as much as 12.8% to a record 720 yuan, signaling a strong investor response to the earnings turnaround.

Analysis

The key second-order read-through is not just that a domestic GPU vendor is monetizing AI demand, but that China’s AI capex is beginning to bifurcate from frontier-model compute into a protected, lower-performance ecosystem. That matters because once procurement standards and software stacks are adapted to local chips, the competitive moat shifts from raw silicon performance to supply certainty, integration, and state-backed buying power — a structural positive for domestic semiconductor ASPs and utilization rates over the next 4-8 quarters.

For NVDA, the immediate earnings hit is likely modest, but the strategic drag is real: every incremental domestic substitute reduces the addressable install base for restricted products and increases the risk that China revenue becomes more of a replacement cycle than a growth engine. The bigger medium-term risk is ecosystem leakage — if local AI workloads get optimized around domestic accelerators, software developers and cloud buyers may gradually de-risk away from NVIDIA’s stack, which can compress future mix even if headline China demand remains resilient.

The market is likely underestimating how cyclical this remains. Domestic chip winners can re-rate fast on earnings surprises, but the durability of that margin profile depends on sustained policy support, yield improvements, and supply chain continuity; any softness in end-demand or inventory digestion would hit names like this harder than the headline move implies. Conversely, if U.S. export controls tighten again, the domestic-vendor premium extends, but that also raises the probability of supply bottlenecks and execution misses — a classic case where the strongest price action may precede the widest fundamental dispersion.

Contrarian view: the move may be over-interpreted as pure AI secular growth when part of it is likely policy-driven substitution and forced localization. That can support above-average growth for a few quarters, but it also means valuation should be discounted for political risk and customer concentration. The better way to express the theme is not outright chasing the local champion, but owning the policy beneficiaries with the strongest balance sheets while fading any assumption that China AI demand will translate 1:1 into global AI leadership.

Key Decisions for Investors

• Reduce NVDA exposure tactically into strength over the next 1-2 weeks; treat China-related upside as capped until visibility on export policy and China mix improves. If holding, prefer a collar or covered call structure to monetize elevated event risk.
• Pair trade: long a basket of China domestic semiconductor enablers / equipment names versus short NVDA for 1-3 months. The thesis is relative multiple compression if domestic substitution momentum broadens, while NVDA’s downside is cushioned by non-China demand.
• If looking for a cleaner expression, buy 3-6 month call spreads on leading Chinese AI infrastructure beneficiaries rather than chasing the latest mover. Risk/reward is better where policy support is underappreciated and expectations are still lower.
• Watch for a 4-8 week follow-through in reported utilization, order backlog, and gross margin sustainability before adding to any domestic chip exposure. Fade the move if revenue growth decelerates sequentially or if inventory days start rising.
• For longer-term capital, prefer ecosystem picks-and-shovels over single-name hardware stories; the winners are likely to be vendors with supply-chain leverage, software compatibility, and state procurement visibility, not necessarily the highest-beta chip designers.