Author:Leto Bao, the protagonist behind the '30 million stock investment leading to financial freedom' at ByteDance
All information below was shared in ByteDance's US stock group last year, including trading targets, positions, and P&L, all verifiable.
Last August, I just wanted to buy two hard drives.
I was building a small quantitative platform and wanted to pull some tick-level market data for local storage, which required tens of terabytes of space, so I ordered two high-capacity Seagate drives from Pinduoduo. Looking back later, that purchase was the starting point of my entry into the storage theme over the past two years.
Hard Drive Prices Started Changing Daily
After receiving the drives, I bookmarked the link, planning to buy more if needed. A few days later, the price had increased; a few more days, it rose again. The same model, from the same store, had its price adjusted multiple times within a week, only going up, never down.
For a high-volume, highly standardized industrial product, retail prices shouldn't show such a sustained, one-way upward trend. I saw this as an anomaly worth investigating, not just simple merchant price hikes.
Using price comparison tools like Manmanbuy and Keepa, I pulled up the price curve for this specific drive over the past few months and compared it with other high-capacity models from Seagate and Western Digital. The conclusion was consistent: it wasn't an issue with a single model. The entire line of high-capacity mechanical hard drives was rising, and doing so continuously and unidirectionally, unlike short-term fluctuations caused by promotions.
At this point, it was almost certain there was a larger reason behind it.
Following the Thread, It Was AI Snapping Up Hard Drives
Digging deeper, the logic gradually became clear.
The market focused more on AI's demand for GPUs, but its demand for storage was equally massive. Large model training and inference generate vast amounts of data that need long-term preservation. For long-term, low-cost data storage, the primary reliance isn't SSDs, but high-capacity mechanical hard drives. The drives used in data centers for this are called nearline enterprise-grade hard drives, which major cloud providers like Microsoft, Amazon, Google, and Meta are purchasing in large quantities.
Seagate happened to be positioned perfectly for this wave of demand. Its key HAMR technology significantly increases per-drive capacity, aligning precisely with data center needs. With limited production capacity, manufacturers prioritized the more profitable enterprise-level orders, squeezing supply in the retail channel. The price increase I saw on Pinduoduo was essentially the procurement demand from AI data centers transmitting all the way to consumer pricing.
I checked Seagate's earnings report at the time: revenue for the latest quarter grew 39% year-over-year, gross margin hit a record, and the market also began pricing the data storage sector as part of the AI industrial chain.
After confirming the logic, I bought 500 shares at over $150, and posted my reasoning, cost basis, and position in the company's internal US stock discussion group.
What Gave Me the Conviction to Add Was the Subsequent 13F Filings
Being personally convinced wasn't enough; I needed to confirm if institutional money was making the same judgment.
The US market has a very useful public data set: institutions managing over $100 million are required to disclose their US equity holdings quarterly, known as 13F filings. This is essentially a legal, public record of institutional holdings that anyone can access.
I didn't add to my position immediately, wanting to observe the trend for another quarter or two. A change in a single quarter might be coincidental; a trend across consecutive quarters in the same direction is more credible. When the November filing for Q3 came out, plotting Seagate's institutional holdings over the past year into a line, the direction was already clear:
In the second half of 2024, this stock was largely ignored, with the number of institutions holding it hovering around eight hundred, even slightly declining. Entering Q2 2025, a clear inflection point emerged, accelerating further in Q3: the number of institutions holding it increased from over eight hundred to over twelve hundred, with the number of new entrants also increasing quarter by quarter.
It should be noted that the rise in total holding value to $45.6 billion within a year was largely due to the stock price appreciation itself, not entirely new capital inflows. However, 'breadth' indicators like the number of institutions and new positions tell a more important story. Their sequential increase meant this wasn't a bet by one or two funds, but a group of professional capital consistently entering the market.
After confirming this step, I added to my position on a larger scale with more confidence. I also began seriously researching the storage sector and later used LEAPS CALLs to continue adding to positions in $STX and $SNDK.
Looking Back Now
On the day I bought the hard drives, Seagate closed at over $150; today it's around $965, a gain of over six times. It once surpassed Palantir last year to become the top performer in the S&P 500 for the full year. Just the initial 500 shares alone represent a paper profit of around $400,000; the profits from subsequent additions aren't detailed here.
It's still somewhat surprising that two hard drives ultimately led to such a trade.
To Summarize This Framework
It's actually not complicated:
-
Anomalies encountered in daily life (price increases, shortages, queues) often appear earlier than news or earnings reports. Ordinary people sometimes receive first-hand signals earlier than professional institutions.
-
Don't stop at the impression of "seems more expensive." Plot the price into a curve; you can usually judge whether it's a trend or just noise.
-
Ask the next question: Is this driven by a long-term, structural demand? Then find the listed company that directly benefits and is in a key position in the industrial chain.
-
Finally, use 13F filings to verify institutional sentiment, focusing on the trend across consecutive quarters, not just a single quarter.
This method doesn't guarantee being right every time, but at least it allows your buy decision to be based on logic, not just feeling.
Finally, Being Clear About the Risks
This was one successful instance. I've also had cases where I focused on a price increase signal that turned out to be just short-term volatility. Those weren't posted in the group, but they are equally real—survivorship bias is evident.
The above is a personal review and does not constitute investment advice. Invest at your own risk.
However, I endorse this framework: pay attention to anomalies encountered daily, ask the next 'why', find the underlying listed company, and then use 13F to see if professional capital agrees. Next time something you regularly buy increases in price for no apparent reason, you might want to think: who is making that money, and is that company already public.
