From a Hard Drive Purchase to a Million-Dollar Trade
In August 2024, a former ByteDance employee bought two Seagate high-capacity mechanical hard drives on Pinduoduo for a quantitative trading platform. Upon delivery, he noticed the same model had been repeatedly priced up—only up—within a week. Such sustained one-way price movement in a highly standardized industrial product is anomalous. Rather than dismissing it as merchant opportunism, he treated it as a signal worth investigating.
Using price tracking tools like Manmanmai and Keepa, he plotted historical price curves for the drive and compared them with other Seagate and Western Digital models. The conclusion was consistent: the entire nearline hard drive product line was experiencing continuous, unidirectional price increases. This wasn't short-term promotional noise; something structural was driving it.
The Hidden Driver: AI's Insatiable Storage Demand
His research led to a clear logic: while the market fixates on AI's demand for GPUs, it underestimates the parallel demand for storage. Large-scale model training and inference generate massive data that must be cost-effectively retained long-term. Nearline enterprise mechanical hard drives—not SSDs—are the primary solution. Hyperscalers like Microsoft, Amazon, Google, and Meta are placing huge orders, tightening retail supply. Seagate's HAMR technology, which boosts per-drive capacity, aligns perfectly with data center needs.
Seagate's latest quarterly report confirmed the thesis: revenue up 39% year-over-year, gross margins hitting records. The market began pricing the storage sector as part of the AI supply chain. With logic confirmed, he bought 500 shares at around $150 and shared the rationale, cost, and position in ByteDance's internal US stock group.
Verifying the Thesis with 13F Filings
Personal conviction required institutional validation. The US 13F quarterly disclosure (for managers with >$100M AUM) provides a legal, public record of large funds' holdings. Instead of immediately adding to his position, he waited to observe multiple quarters of 13F data. In H2 2024, only about 800 institutions held Seagate; that number dipped slightly. But by Q2 2025, it reversed clearly; by Q3, it accelerated to over 1,200. The number of new institutions opening positions also rose sequentially.
While some of the growth in total market value (to $45.6B) came from price appreciation, the breadth indicators—number of holders and new entrants—were more telling. This wasn't a few funds speculating; it was broad institutional accumulation. Confident, he scaled up via common shares and later LEAPS calls on $STX and $SNDK. Seagate's stock surged from ~$150 to ~$965, a 6x gain that briefly made it the S&P 500's best performer in 2025, overtaking Palantir. The initial 500 shares alone yielded ~$400,000 in paper profit.
Methodology and Risk Disclosure
His approach can be summarized: (1) notice anomalous everyday signals (price hikes, shortages) that often precede news and earnings; (2) quantify the signal—plot price curves to distinguish trend from noise; (3) trace upstream to identify a long-term structural demand; (4) find the publicly traded company at the key node of the value chain; (5) validate with consecutive quarters of 13F data for institutional conviction.
He also cautions against survivorship bias: he has tracked similar price signals that turned out to be short-term fluctuations—those failures were not posted publicly but are equally real. This is a personal review, not investment advice. Trade at your own risk. But the core idea stands: the next time a frequently purchased item inexplicably rises in price, ask who captures that profit and whether that company is publicly traded.