Inflation fears risk tipping into deflation
Memory chips serve as the hidden foundation of the global economy, enabling devices from smartphones to AI servers. A recent price surge has raised concerns about whether this signals the start of a new inflationary period or an early warning of deflation.
Dynamic Random Access Memory (DRAM) chips function as short-term memory for nearly all electronic devices. When a laptop starts or an AI model processes data, DRAM ensures smooth operation. Unlike storage chips, which retain information when powered off, DRAM clears its memory immediately after shutdown. This makes it essential—and particularly vulnerable to supply changes.
The market is controlled by three companies: Samsung Electronics, SK Hynix, and Micron Technology. Together, they supply chips to consumer electronics, cloud computing, and automotive industries. When supply tightens, prices don’t just increase—they can spike sharply. The Korean Stock Exchange has benefited from this trend, with Samsung and SK Hynix making up about 35% of its index value.
The impact reaches beyond financial markets. High Bandwidth Memory (HBM), a specialized DRAM variant, is vital for GPUs that train AI models. Companies like Microsoft, Tesla, and Meta depend on large volumes of high-performance memory to make their flagship accelerators function as designed. Rising memory costs directly affect the expenses of building and operating AI, cloud services, and electric vehicle infrastructure.
When supply shrinks, producers gain pricing power. DRAM manufacturers see their profit margins grow quickly as demand exceeds availability. Major buyers—such as hyperscalers with substantial financial resources—continue purchasing despite higher costs. These firms have easy access to capital, allowing them to absorb or pass along increased costs without immediate consequences. Smaller firms, however, face different challenges. Without similar cash reserves or credit access, they must decide whether to pay higher prices or reduce operations. If enough companies cut back, the outcome could be a sharp decline in demand rather than gradual adjustment.
This scenario supports the inflation argument. Higher memory chip costs affect everything from personal computers to data centers. If these costs persist, they may drive up prices across the tech sector, creating a cost-push inflation effect. Since AI infrastructure remains a priority for the largest corporations, the market is unlikely to correct itself soon.
Related: SpaceX asteroid threatens telco and cable giants
Another possibility exists. If the issue isn’t just price but availability, smaller manufacturers may struggle to obtain necessary chips. Production slows, revenue drops, and unfinished inventory accumulates. For companies already operating on thin margins, this can lead to financial trouble quickly.
Most firms lack the ability to absorb higher costs like hyperscalers can. When credit conditions tighten and unemployment rises, the effects multiply. Demand destruction isn’t hypothetical—it marks the beginning of recessions. With credit spreads at historic lows, even a minor disruption could have significant consequences.
Persistent high prices increase the likelihood of these secondary effects. The reaction isn’t in question—only its extent. Industries beyond tech, including automotive and consumer electronics, could experience pressure. If enough companies reduce spending, the result might not be inflation but a deflationary cycle driven by falling demand.
For now, major tech firms continue investing. The broader economy may not share the same resilience. The coming months will show whether this is a temporary constraint or the beginning of a larger shift—and which direction prices will move next.
This situation mirrors broader trends in corporate borrowing, where large tech firms drive debt markets to fund expansion.
