Samsung posts record profit as chip earnings jump nearly 49-fold

Samsung reported record quarterly profit, driven mainly by its memory and other chips division, which grew operating profit almost 49-fold amid the AI frenzy. The company also warned that the global memory shortage is not easing and, by its estimate, could worsen as early as 2027.

A record jump in profit

From January to March, operating profit at Samsung's chip division hit a record 53.7 trillion won, versus 1.1 trillion won a year earlier. That is almost 94% of the company's total quarterly operating profit of 57.2 trillion won. Total revenue for the quarter rose 69% year over year to 133.9 trillion won.

For Samsung this is not just a strong report. It shows how quickly AI demand is reshaping the economics of the semiconductor market.

The market reacted as expected: Samsung shares are up 88% since the start of the year, well ahead of the broader market. The stance of the biggest US tech companies added further momentum. Just beforehand, Alphabet, Amazon, Meta and Microsoft again signaled that they do not plan to cut AI infrastructure spending. For Samsung, that points to a long wave of demand rather than a one-off spike, and the company is already ready to sharply raise capital expenditure in 2026.

Why the shortage is growing

The main cause is the boom in AI data center construction. Memory and logic makers are increasingly shifting capacity to more complex chips needed for Nvidia accelerators and similar systems. The problem is that the same lines, people and materials no longer produce the previous volume of standard memory and ordinary components that smartphones, displays and other electronics need. As a result, bottlenecks appear not only in premium GPU clusters but across the whole memory supply chain.

"Our supply is far behind customer demand."

Samsung says plainly that the gap between demand and supply will last a long time. According to the company's senior executives, current orders already show that the shortage in 2027 may be even deeper than in 2026. To lock in volumes, customers are signing multi-year binding contracts, and Samsung itself warns that the strain cannot be relieved quickly. New fabs take too long to build for the market to absorb the surge in orders quickly.

• Building new fabs takes years, not quarters
• Demand is fueled by memory purchases for AI servers
• Multi-year contracts reserve future volumes in advance
• The shortage of expensive memory also pushes up prices of more ordinary chips

Pressure on the rest of the business

Paradoxically, the chip record does not mean everything is equally good across Samsung's businesses. Rising prices of ordinary components are already hitting other segments. Samsung warned that profitability of its mobile and networking division will decline in 2026, because components are getting more expensive faster than the company can pass those costs on to buyers. In other words, money earned on AI chips is partly eaten up by the rising cost of its own electronics.

In the first quarter, profit at the mobile and networking business fell 35% to 2.8 trillion won. The display division, which supplies panels to Apple among others, reported a 20% drop in operating profit to 400 billion won.

The company also separately flagged the risk of higher logistics costs due to more expensive oil, although the Middle East conflict has not yet disrupted production itself, it said: inventories and diversified supplies of industrial gases, including helium, helped.

What this means

The AI frenzy is affecting not only individual companies' revenue but the whole balance of the global chip market. While money and capacity flow into memory for data centers, the shortage will weigh on more mainstream electronics and push prices up. For the market, it is a signal that the AI boom is still far from saturation; for manufacturers, that the fight for capacity and long-term contracts is only intensifying.

*Meta has been designated an extremist organization and is banned in Russia.