Samsung’s Q2 profit surged 1800% year-over-year. New record. Stock dropped 3% on the news.
The data doesn’t lie. The market is not celebrating. It’s running from the peak.
Samsung is the largest memory chip maker globally. DRAM and NAND dominate its revenue. But the semiconductor cycle is brutal. Two years: 18 months up, 6 months down. We have been in the upcycle since late 2023. DRAM prices rose 30-40% in H1 2024. NAND rose 50-60%. That is the sugar high. Now the market smells the hangover.
In May 2024, DDR5 spot prices started slipping. Down 5%. Small move, but it broke the streak. That is the signal.
Based on my experience analyzing liquidity depth during DeFi Summer, I know that when volume slows before price, the pivot is near. Same here: spot price weakness precedes contract price correction.
Let me walk through the evidence chain.
First, the profit mix. Samsung’s 1800% jump is not from volume growth. It is from price recovery. Revenue grew 129%. That is impressive, but operating profit margins are still below 2021 peaks. The gap? HBM.
Samsung is second in HBM. SK Hynix leads. Samsung’s 12-layer HBM3E is still in development. Hynix shipped it to NVIDIA in Q1. Samsung missed that window. In AI memory, time-to-market is everything. That is why Samsung’s HBM share is 40-45% vs Hynix 50%+. And HBM carries 20 percentage points higher margin than traditional DRAM. Samsung’s late entry means it is leaving margin on the table.
Second, the inventory cycle. Channel inventories rose from 5 weeks in 2023 to 8-10 weeks now. That is healthy. But we are past the restocking phase. Once inventories hit 12 weeks, the destock begins. History shows that peak earnings coincide with peak inventory builds. Samsung’s profit peak is now. The question is when the fall begins.
Based on my work tracing USDT outflows during the Terra collapse, I learned that rapid growth often masks structural weaknesses. Same here: profit surge masks the coming inventory correction.
Third, the valuation trap. Samsung trades at 15-18x trailing PE. For a cyclical, that is high at the top. Typical cyclical peak PE is 8-12x. When profits peak, PE compresses. The market is paying a premium not for growth, but for the inevitable decline. That is a contrarian signal.
Data is the only witness that never sleeps. The price action after the earnings release confirms it: the market used the profit beat to sell. Classic "sell the news".
Fourth, capital expenditure. Samsung’s 2024 capex is estimated at $37 billion. Record high. They are building fabs in Pyeongtaek and Texas. High capex during profit peaks is dangerous. When the cycle turns, depreciation eats margins. Samsung’s EUV machines cost €150 million each. Depreciation over 7 years. If utilization drops below 70%, those machines become a liability.
Fifth, the geopolitical overlay. Samsung is caught between US export controls and Chinese market access. It lost market share in China from 30% to 20%. That is a permanent structural drag. The profit surge from current operations does not change that.
In the ashes of Terra, we found the pattern. Profit peaks that feel unsustainably high are usually just that. The market knows. The code doesn’t lie.
But wait. Could the market be wrong? Perhaps the cycle is different this time? HBM demand is structural. AI is not a fad. Samsung could catch up by HBM4 in 2026. The valuation might be justified if they execute.
However, the data contradicts that. The speed of the profit surge itself is a warning. When profits triple in a single quarter, it is not sustainable. It is a product of base effects and price spikes. The base effect will fade. Prices will normalize.
The contrarian view that "this time is different" is dangerous. Memory cycles have been consistent for decades. The only difference is the magnitude, not the direction.
We don’t need to predict the exact timing. We just need to read the signals. And the signals say: this is the top.
Watch DDR5 spot prices this week. If they break below $4.5, the downtrend confirms. Watch for Samsung’s 12-layer HBM3E certification from NVIDIA. If it comes by Q4, maybe there is a second wind. But for now, the data says one thing: the cycle is turning. Position accordingly.