Samsung’s AI Growth: From US$1tn Cap to 1,800% Profit Rise

Samsung Electronics is one of the leading global producers of the components that keep the wheels of AI moving. The rising demand for memory chips in the light of the AI boom is driving a historic 1,800% profit forecast for the South Korean technology giant.

The company is forecasting a 19-fold jump in its second-quarter operating profit of ₩89.4tn (US$58.4bn) between the start of April and the end of June. 

This figure marks its third record quarterly operating profit in a row and represents an astronomical leap from the ₩4.7tn (US$3.1bn) reported in the same period last year.

Samsung, which was once synonymous with conventional hardware, is now firmly positioned as a primary manufacturer of the hardware powering next-generation of AI infrastructure. 

Driven by this explosive demand, it was able to cross the US$1tn market capitalisation milestone in May 2026, along with Micron and SK Hynix, which are its direct competitors in the industry.  

Driving record pricing

Samsung released its earnings guidance ahead of its full detailed report due later in July. 

The latest preview comes as demand for semiconductors continues to outstrip supplies, an imbalance that has pushed up market prices significantly. Following this, Samsung hiked the prices of its memory chips as well.

The company brought in around ₩171tn (US$111.7bn) of sales during the quarter, which is more than double the amount for the same period last year. Revenue, on the other hand, rose 129% year-on-year, showing immense structural strength.

Marc Einstein, Industry Analyst at Counterpoint Research, tells BBC the projected earnings mark one of ‘the best quarterly performances ever’. He adds that the performance is close to the tech sector record set by NVIDIA earlier this year. 

“This has everything to do with the AI boom as memory companies continue to ride a tidal wave driven by limited supply and unprecedented demand,” Marc says.

Research firm IDC reports that demand for data centres and AI infrastructure is different from anything the memory industry has navigated before, which has severely impacted the supply of chips for everyday electronics.

HBM production shifts market dynamics

The global surge in gen AI relies heavily on high-bandwidth memory (HBM), a specialised component that Samsung Electronics is aggressively scaling up to manufacture.

Because the company has rapidly reallocated its production capacity toward building these high-performance, vertically stacked chips, the market availability of standard storage components has shrunk.

Consequently, memory chip prices continued to climb during the quarter as AI spending broadened into conventional DRAM and NAND products. Citi Research reports that average selling prices for DRAM and NAND rose 44% and 53% quarter-on-quarter, respectively.

Analysts say this rapid growth in HBM production has severely tightened the supply of conventional memory products. This impacts chips used in smartphones, personal computers and enterprise servers.

The high-bandwidth memory chips deliver faster performance and lower power consumption. They are tightly integrated with AI processors, creating higher barriers to entry and giving suppliers greater pricing power compared to standard memory products.

Samsung Electronics has been able to achieve its profit even after setting aside funds for sizable worker bonuses. The company agreed to a wage deal in May linking employee pay to operating profit following union worker strikes.

Analysts tell Reuters that without those provisions, operating profit would likely have exceeded ₩100tn (US$65.4bn). Losses at its foundry and logic chip businesses are likely to widen because bonus expenses are allocated across the entire semiconductor division.

Whether the AI boom can sustain this historic 1,800% profit rise remains the critical question for Samsung. While supply shortages and high-bandwidth memory demand continue to drive record pricing, market jitters are growing over the long-term discipline of tech hyperscalers. 

If spending on AI data centres slows down, replication of these massive triple-digit gains will prove difficult to maintain.