Samsung's profit surges 1800%, is the semiconductor super cycle coming back?

If the tech industry were a concert, Nvidia has been in the spotlight for the past two years, but now Samsung Electronics suddenly rushes onto the stage with a microphone: Q2 profit surged 1800% year-on-year, with single-quarter earnings directly exceeding the total of the past three years.

What does this mean? It's like someone saving pocket money for three years straight, then suddenly winning the lottery this summer.

The core reason is only four words: AI storage exploded.

The larger the AI model, the more the server becomes a "memory-devouring monster." HBM high-bandwidth memory has become one of the hottest hardware components, and manufacturers like Samsung and SK Hynix have finally turned the tables from past price wars, entering a "supply shortage" mode.

Many investors worry: Could this be just a short-term boom? But looking at current global AI capital expenditure, Microsoft, Google, Amazon, and Meta are all frantically building data centers. As long as AI training and inference continue to expand, high-end storage demand is unlikely to suddenly die out.

More importantly, this cycle is different from traditional semiconductor cycles. In the past, the focus was on smartphone and PC sales, but now it's about "who can train a larger model." This means a broader source of demand and potentially a longer duration.

So, rather than saying Samsung's profit surged, it's more that AI is redefining the profit logic of the entire semiconductor industry.

In the past, the market thought "chips = cyclical stocks," but now more and more people are beginning to believe "AI chips + AI storage = new infrastructure."

If this logic holds, then Samsung's earnings report may not be the end, but the opening remarks of a super cycle. #SK海力士ADR获超额认购