#AI股持续上涨还有哪些投资机会 AI sector continues to trend upward; possible sub-segment investment opportunities to watch

1. Compute hardware (higher certainty; prioritize earnings delivery)

1. Optical modules/silicon photonics/CPO: AI model inference surges, with 1.6T iteration and co-packaged optics upgrades. Bandwidth is a rigid requirement for compute clusters, and overseas cloud providers’ capital spending continues to drive momentum.

2. HBM and the storage supply chain: One of AI’s biggest bottlenecks. Demand-supply for high-bandwidth memory is tight, and storage chip cycles are moving upward.

3. Liquid cooling and heat dissipation, high-speed PCBs, and advanced packaging: AI racks see a significant increase in power consumption. Liquid cooling is gradually becoming standard equipment for intelligent computing centers, and high-speed substrates and packaging are key hardware upgrade needs.

4. AI-powered electric grid equipment: power consumption for intelligent computing centers is surging. Demand for transformers, high-voltage power distribution, and energy storage support equipment is growing.

II. Software and AI Applications (From Concept to Paid, Real-World Deployment)

1. Enterprise AI intelligent agents (Agent): digital office employees, workflow automation. The government-and-enterprise paid business model is already working. AI-SaaS with clearly defined cost-reduction needs.

2. AI coding/R&D tools: code generation and defect detection. Strong B-end willingness to pay and a mature commercialization closed loop.

3. Multimodal AIGC: AI videos, digital humans, AI short dramas, and marketing assets. Monetization follows two paths: C-end memberships and B-end API revenues.

4. Vertical-industry AI: pharmaceutical R&D, industrial quality inspection, and financial risk control. By binding to industry data, the barriers to entry are higher.

III. Embodied intelligence (AI + the physical world; a long-term main theme)

Industrial robots, humanoid robots, and warehouse logistics automation. Large models give robots a “brain,” moving from labs to factory deployments. Prioritize integrated hardware-and-algorithm targets, focusing on order and project execution.

IV. Edge opportunities that are easy to overlook

1. Inference compute service providers: not just training—provide enterprises with inference scheduling and Token output. As downstream applications explode, inference demand is directly boosted.

2. AI data supply chain: high-quality training datasets, data annotation, and data cleaning—large-model iteration is inseparable from high-quality data.

3. On-device AI: large models on mobile phones, PCs, and in-vehicle systems. On-device chips and software optimization, with hardware terminals equipped with AI capabilities.

Risk warning

1. Some AI sector valuations are already too high. Market sentiment is gradually shifting from “buying the story” to performance delivery. Stocks with no revenue aside from narratives carry high volatility risk.

2. Tightening by the Federal Reserve and rising yields on U.S. Treasuries will suppress the overall valuation of growth sectors. Macroeconomic liquidity changes will disrupt AI sectors.

3. Industry technology iterates quickly. Companies that choose the wrong technical route will be eliminated fast.