Samsung is not an AI company. It does not have a chatbot. It does not run a cloud platform or a social network. But right now, it is quietly winning the AI era harder than almost any company in the world. And the numbers just became impossible to ignore.
In early April 2026, Samsung released preliminary earnings for the first quarter of the year. The headline figure: a record operating profit of 57.2 trillion won, roughly $37.9 billion. That is a 755 percent jump compared to the same period a year ago. It also exceeds Samsung’s entire annual operating profit for 2025, achieved in a single three-month window.
But the more striking detail is what is driving it.
It Is All About the Chips
Samsung’s Device Solutions (DS) division – its semiconductor arm – is estimated to account for roughly 95 percent of those total profits, translating to somewhere between $33 and $36 billion from chips alone.
To put that in perspective: Amazon’s entire company posted $25 billion in operating income in Q4 2025. Meta’s entire company posted $24.7 billion for the same period. Samsung’s memory division, on its own, is earning more in a single quarter than either of those two tech giants makes across all of their products and services combined.
This is not Samsung the phone maker or Samsung the TV brand. This is just the part of the business that makes chips. Specifically, memory chips.
Why Memory Chips Are the New Oil
To understand what is happening here, you need to understand what AI actually runs on.
When you use ChatGPT, Gemini, or any large language model, somewhere in a data centre, an AI processor is doing billions of calculations per second. Those calculations need data, and that data has to come from somewhere very close by, very fast. That somewhere is memory.
We have written before about Samsung’s push into HBM — High Bandwidth Memory — the specialised type of memory that sits closest to AI processors and feeds them data at extraordinary speeds. The world’s biggest AI companies, including Google, Microsoft, Meta, Amazon, and OpenAI, all need this memory for their data centres. And Samsung is the world’s largest memory chip maker.
A global shortage of high-performance memory chips has allowed Samsung to charge significantly higher prices to data centre operators. Memory prices surged an estimated 40 to 50 percent in the final quarter of 2025 alone. Supply is tight because building these chips is extraordinarily difficult, and everyone wants them at once. Memory manufacturers are also redirecting production capacity toward HBM specifically, because it is far more profitable than the RAM that goes into consumer laptops and phones. Which only tightens supply further and pushes prices even higher.
A Comeback Story With Caveats
This result is also a redemption arc.
In 2024, Samsung was struggling. Its HBM3E chips failed to get certified by Nvidia, meaning they were shut out of the most profitable segment of the market. Rival SK Hynix, the other South Korean memory giant, pulled ahead. Samsung’s stock lagged. Analysts worried it had missed the AI wave entirely.
Since then, Samsung has rapidly narrowed the gap. It has qualified its next-generation HBM4 chips and secured design wins that position it for a larger slice of the market through the rest of 2026. Clients including Google, Meta, Amazon, Microsoft, and OpenAI are now lining up for Samsung’s HBM chips to stock their AI data centres.
It is worth being transparent: Samsung has not yet released its full Q1 2026 divisional breakdown. The $33 to $36 billion DS division figure is an analyst estimate based on the preliminary guidance. The full audited report will follow. But even directionally, the scale of the performance is significant enough to reframe how the industry sees this company.
What This Means
This is not just a good quarter for a South Korean electronics company. It is a signal about where value is accumulating in the AI economy.
The companies building models and AI interfaces; Meta, Google, Microsoft, OpenAI, are spending hundreds of billions on infrastructure. The companies supplying the raw materials that infrastructure runs on are collecting that spending as profit. Samsung sits firmly in the second group, and right now, that is an extraordinary place to be.
According to CNBC, Samsung’s memory business has been setting all-time highs for quarterly revenue and operating profit across consecutive quarters, driven by surging market prices, high-bandwidth memory shipments, and a broader shortage across the memory market. Analysts at several brokerages now forecast Samsung’s operating profit for full-year 2026 could exceed 100 trillion won. That’s more than double last year’s level if memory prices hold firm and HBM shipments continue to expand.
Samsung’s Q1 2026 operating profit already exceeds its entire annual operating profit for 2025. In a single quarter.
Samsung may not be the face of the AI revolution. But it is very much one of its biggest beneficiaries — and one of the most important companies in the entire AI supply chain, whether or not its name appears on any model, app, or product you use.
