TradingKey - Recently, due to tight CPU supply, Intel (INTC) and AMD have successively raised prices for their CPU products, covering servers and some high-performance consumer products; on the news, both Intel and AMD rose more than 7% on Wednesday. Arm also surged 16.38% in the previous trading session after announcing its self-developed AGI CPU and providing a staggering revenue growth outlook.
Advanced manufacturing capacity continues to shift toward AI-related chips, with the "crowding-out effect" of GPUs becoming increasingly apparent. Meanwhile, demand for traditional CPUs has not weakened significantly, leaving them in a dilemma of stable demand but limited supply, which has granted CPU manufacturers stronger bargaining power.
From a production line perspective, the uneven allocation of capacity in the wafer foundry sector has created a supply gap for CPUs. Advanced node manufacturers, represented by TSMC, are currently prioritizing high-margin AI chips, leading to a relative tightening of supply for general-purpose computing chips like CPUs. Against this backdrop, both Intel's internal production lines and the foundry ecosystem that AMD relies on are facing varying degrees of capacity allocation pressure, driving prices upward.
Looking at the demand structure, this round of CPU price increases is essentially a manifestation of "computing power repricing." As AI application penetration rises, data center demand for high-performance computing resources continues to expand. Although CPUs have been replaced by GPUs in some scenarios, they remain irreplaceable in general-purpose computing, edge computing, and enterprise applications, indicating that CPU demand has not been weakened.
It is worth noting that behind the price hikes lies a shift in manufacturers' profit models. During the stage dominated by the PC cycle, CPU prices were more sensitive to fluctuations in end-user demand; however, under the current demand structure centered on data centers and cloud computing, customers prioritize performance and stability over price, giving manufacturers stronger pricing power.
More stringent requirements also demand higher yields in CPU production. Previously, manufacturing yields below normal standards were the fundamental reason limiting Intel's supply. Lip-Bu Tan once publicly noted that while yields were in line with internal plans, they remained below desired levels and had not yet reached industry-leading standards. Therefore, greater pricing power in the future may be held by manufacturers with superior yields.
From a capital markets perspective, the impact of this price hike on the supply chain is showing clear divergence.
Profit margins for upstream chip design firms are expected to see immediate improvement, while downstream cloud service providers and corporate IT spending may face cost pressures. In the cloud computing sector, the cost pass-through mechanism is likely to materialize gradually through service price adjustments, which will in turn impact the cost of AI applications for end users.
However, it is important to note that price hikes are not a unilateral positive. If CPU prices continue to climb, it could accelerate the corporate shift toward heterogeneous computing architectures, further increasing reliance on GPUs and specialized AI chips, which may exert a substitution effect on CPU demand over the medium to long term.
Furthermore, macroeconomic uncertainties cannot be overlooked. If the pace of the global economic recovery falls short of expectations, corporate IT spending could come under renewed pressure, thereby undermining the support for CPU demand. Under these circumstances, the current price hikes may be difficult to sustain over the long term.