Samsung Electronics in-depth report: 89.4 trillion KRW quarterly profit, HBM4 catch-up and wafer foundry reduced losses

TL;DR

1.Samsung Electronics is a combination of a storage profit engine, terminal income base, display and automotive electronics buffer, and wafer foundry loss reduction option. Its business structure is different from the high-purity SK Hynix. In 1Q26, the DS division contributed 93.8% of the group’s operating profit, but only represents part of the company’s value. If all other businesses are seen as a drag, it underestimates net cash, internal synergy, and the buffer effect during downcycles; if all are seen as synergy, it ignores that mobile phones and TVs are being cannibalized by rising memory prices.

2.The quarterly operating profit of 89.4 trillion KRW first proves the strength of traditional storage prices, but it cannot alone prove that HBM leadership has been restored. The company's preliminary guidance indicates 2Q26 revenue of about 171 trillion KRW and an operating profit of about 89.4 trillion KRW; Bank of America estimates about 91.1 trillion KRW in storage operating profit, while wafer foundry, mobile phones, and TV/home appliances still run at a combined loss. Samsung Electronics has confirmed HBM4 mass production and sales to Nvidia’s Vera Rubin, but has not disclosed whether the customer has passed “certification,” nor has it revealed its market share or product-level profitability.

3.Profit expectations have undergone a sharp upward revision in half a year, while valuation multiples remain restrained. In January, HSBC and JPMorgan projected 2026 operating profits of about 169 trillion—173 trillion KRW; by July, Citi and Bank of America had raised estimates to 381 trillion—401 trillion KRW. During the same period, Citi’s storage valuation multiple dropped from 7.9x to 7.6x, and Bank of America’s P/E ratio fell from 11x to 10x. This indicates that the target price range of 530,000—550,000 KRW relies mainly on profit realization, rather than further multiple expansion.

4.The investment value of catching up with HBM4 lies in driving high-bandwidth memory, 4nm base dies, and advanced packaging together, but "internal orders" do not equate to "external competitiveness." Subsequent evaluations should separately verify HBM revenue, customer coverage, yield rate, 1c process capability, and external foundry customers. If any link fails, vertical integration could shift from synergy to a cycle of high capex and internal costs.

5.The real stress test comes after 2027. Bank of America expects operating profit in 2027 to reach 506.4 trillion KRW, with free cash flow of 335.5 trillion KRW, but this is not net profit, nor is it the company’s own guidance. Its 2028 scenario already assumes average DRAM and NAND prices will decline by 10%—12%. If profits only fall slightly then, it will indicate that long-term agreements, product mix, and cost improvements have increased the durability of earnings; if cash flow plummets with prices, we’re still witnessing an inflated cycle peak.