Two major negative factors impact but fall short of expectations; AI hardware stocks find support, Philadelphia Semiconductor Index rises for two consecutive days.
```
Resistance to data center construction and open-source models are seen as overinterpreted; the fundamentals of semiconductor demand remain solid
On Tuesday, the semiconductor sector led by memory chip manufacturers surged over 5%, outperforming the broader market for two consecutive trading days.
According to Bloomberg analysis, although short covering provided short-term support, the more convincing argument is that neither the local opposition to data center construction nor the emergence of open-source large models are sufficient to signal the end of the AI capital expenditure cycle.
Recent Asian export data, as well as price increases by companies such as TSMC, both confirm that underlying demand in computing power, storage, network, and electricity remains intact, and that market sentiment had previously been overly pessimistic.
As the above two major concerns gradually prove unfounded, the rebound since Friday's low may be much more than just a technical correction after overselling.
Resistance to Data Center Construction is Mostly a Delay
Local resistance to data center construction indeed poses real pressure.
According to Morgan Stanley citing third-party data, around $156 billion in data center projects for 2025 have already been canceled or delayed, and another $130 billion are affected in the first quarter of 2026.
The bank estimates AI capital expenditure in 2026 to reach $877 billion, facing multiple downside risks such as power grid constraints, construction suspensions, and stricter regulations on electricity and water usage.
However, given the strategic importance of AI, the most likely outcome of this rebound is delay and redesign, rather than a complete halt.
From another perspective, construction restrictions may instead boost semiconductor procurement demand within existing sites.
Operators can replace inefficient servers, increase rack density, and introduce advanced cooling systems to produce more computing power output under the same building area and power supply conditions.
On-site power generation, battery storage, and grid support services provide an alternative path around transmission bottlenecks.
These constraints objectively shorten the lifespan of old equipment, thereby increasing replacement demand for new chips, memory, and cooling systems.
The Decline in API Pricing for Open-Source Models Does Not Mean Hardware Demand Is Falling
The impact of the latest batch of open-source large models from China on hardware demand is also limited.
Take Kimi K3 as an example; reports indicate each instance of this model consumes a significant amount of high-bandwidth memory (HBM) and many accelerators. Though it can handle individual tasks with lower computing power, overall demand for storage capacity remains substantial.
Bank of America research further points out that falling API prices in China should not be interpreted as a sign of declining hardware costs.
Low pricing reflects improved architectural efficiency, China’s more efficient electricity rates, labor costs, and land costs, plus aggressive market share strategies, rather than a systemic collapse in semiconductor hardware costs.
From a longer-term perspective, the proliferation of open-source models may even broaden the overall semiconductor market space.
Closed models concentrate hardware deployment within a few cloud facilities; open models allow enterprises, governments, and sovereign clouds to download and deploy independently, thereby generating new demand for HBM, DRAM, and NAND storage at each endpoint, driving incremental hardware purchases.
Overall, the rebound of AI capital expenditure beneficiaries from recent lows may be driven by more than just technical oversold corrections.
Earnings expectations may have peaked for now, but based on current data, the underlying economic logic of AI infrastructure construction is more resilient than the market previously feared.
Market positioning had adjusted to overly pessimistic levels, while fundamentals have not seen material deterioration, providing space for the continuation of the rebound.
Risk DisclaimerThe market has risks, and investment must be prudent. This article does not constitute personal investment advice, nor does it take into account individual users' specific investment goals, financial situations, or needs. Users should consider whether any opinions, views, or conclusions in this article fit their particular circumstances. Investment is at your own risk. ```