Jason of NextGen Digital Venture wrote in a republished article that South Korea’s benchmark stock index fell 8.77% from July 10 to July 16, 2026. The Korean market was closed on July 17. Over roughly the same window, the Philadelphia Semiconductor Index in the United States fell 9.97% from July 10 to July 17.
His point was not that the artificial intelligence trade is over. He argued instead that a technology changing the world and a stock being worth buying at a given price are separate questions. In his framing, the more useful question after a sharp drawdown is not whether the long-term direction still sounds compelling, but whether an investor’s position was built in a way that allows them to stay in the trade long enough for that direction to play out.
That is the thread running through the whole piece. Jason writes that even if the thesis is eventually right, an uncontrolled position can prevent an investor from reaching the day when the thesis is finally validated. Price, time, funding cost, and leverage all matter. If a stock drops by half before the story is recognized by the market, the practical question becomes simple: can the investor still remain in the position?
Why “AI will change the world” is not enough
Jason says one tempting line he heard in the first half of 2026 was that AI will change the world, so buying the most representative companies should be enough. He says the first half of that statement may be right. The second half leaves out the factors that determine the actual investment result.
What price did you pay? How long are you willing to wait? What is your cost of capital? Did you use leverage? If the stock falls sharply before the thesis is realized, do you still have the financial capacity to stay in the market? In his telling, a powerful long-term trend does not make every entry price reasonable, and a company with durable competitiveness does not guarantee that all of its shareholders will make money.
He uses the latest moves in the Korean market to make that point concrete. As of July 15, 2026, Samsung Electronics and SK Hynix together accounted for 52% of the KOSPI by market capitalization, up from 34% at the end of 2025.
At the same time, leveraged products tied to single stocks expanded quickly. Jason cites disclosures from the Financial Services Commission showing that the market value of these products increased from KRW 4.4 trillion on May 27 to KRW 11.9 trillion on July 15. That was close to a threefold rise in less than two months.
On July 16, the Financial Services Commission and other related agencies announced restrictions, including a suspension on new listings and advertising for the products, along with a higher cash threshold for retail investors taking part. Jason is careful on causality. He says the figures do not prove every sell order in South Korea came from leveraged liquidation. The Bank of Korea raised rates on July 16. Global oil prices and geopolitical risks were also changing, and AI-related equities were facing their own repricing pressure.
Still, he argues that when two companies account for half of an index and leverage linked to individual names has expanded fast, the market structure can magnify the first wave of stress. Once prices fall, some borrowers may need to post more funds, and some structured or leveraged products may need to rebalance. That can create new selling, which pressures prices further and pushes more participants into action. The sequence is familiar. In his view, what has changed is the speed: more financial products exist, more adjustments are rule-based, and the transmission can move faster than before.
Jason adds that he does not know where the next “first gust of wind” will come from or on which day it will arrive. That uncertainty is why he says he kept warning in several programs during the first half of 2026 about valuation and crowding risks in AI trades. Some viewers thought those warnings came too early. He says he understands the reaction. In a real portfolio, being early by a day and being early by a year are very different experiences.
He does not use the July decline as proof that he had been right all along. He says he did not know in advance what would happen on July 16, and that a risk warning is not a timer. Its value, in his words, is much narrower: it helps investors avoid putting themselves in a situation where they must guess the exact top.
He also explains the thinking behind his program, 20 Minutes of Non-Consensus. “Non-consensus,” as he describes it, does not mean taking the opposite side of the market by default. Being in the minority is not inherently correct, and going against consensus is not an investment edge on its own. What matters is asking harder questions when everyone feels comfortable: how much good news is already in the price, and who would be forced to sell if something unexpected happens? Only when a dissenting view is supported by both facts and odds does it become useful.
Deep research does not remove liquidation risk
In the second section, Jason reflects on his own record during sharp sell-offs. Sometimes he handled them well. Sometimes he did not. When he got it wrong, he says he tended to move with market emotion. He would hesitate when he should have admitted that his view had changed. At the point of maximum price pressure, he might suddenly lose patience. In other cases, the problem was already visible, but he held on too long because he did not want to admit error.
Those mistakes look inconsistent on the surface. One reacts too quickly, another too slowly. Jason says the root cause was the same: he was deciding his discipline only after the decline had already started. The experience that pushed him to think more systematically about stop-losses came from a private investor he deeply respects.
According to Jason’s account, this investor built wealth from zero to a fortune in the tens of millions of dollars through his own research and investing. To understand one company thoroughly, Jason says the investor told him that he had even worked at that company and at competitors. Yet despite that depth of research, the investor still went through two or three episodes in which his account was almost entirely wiped out.
The example that left the strongest impression on Jason involved Meituan. Based on the investor’s retelling, shortly after Meituan was listed and while the stock was trading near HK$80, he was so confident in the business that he added borrowed money roughly equal to his own principal. In Jason’s simplified math, he put HK$100 of his own capital together with nearly HK$100 borrowed, for a total position of HK$200 in assets.
Jason then points to public market data. Meituan’s 2018 IPO price was HK$69. On January 3, 2019, the stock’s public quote hit an intraday low of HK$40.25. Starting from around HK$80, that was close to a halving. For an account that had borrowed almost one times its own equity, a drawdown of that size would leave the investor’s own capital close to zero.
Based on the investor’s own review shared with him, that decline wiped out the position. Where Meituan later traded became irrelevant to that account. Jason adds an important caveat: he has not seen the broker statements, so the episode should be treated as the participant’s own recounting rather than an audited trading record. The math of margin, though, does not change because of that limitation.
What affected him most, he says, was that this was not a case of someone speculating in a company they did not understand. Quite the opposite. The investor may have known the company better than most participants in the market. Even when the long-term direction is sound, leverage can still take away a person’s ability to wait. Jason notes that the investor later started over and returned to a fortune in the tens of millions of dollars. That shows unusual skill and resilience, he writes, but he does not want to romanticize starting over. Most people neither have the capacity nor the need to treat a total wipeout as a required lesson in investing.
Four separate questions in every investment
That experience left Jason more convinced that every major investment requires four separate answers.
- Direction: Are you ultimately right about the thesis?
- Odds: At the current price, is the risk still worth taking?
- Time: How long might it take before the thesis is recognized?
- Position size: If the worst path shows up first, can you still remain in the game?
He argues that getting the direction right does not mean the price is right. Even if direction and price both look reasonable, timing can still be wrong. And even if the first three are broadly intact, leverage and oversized exposure can still force an investor out early.
Jason brings up the Kelly criterion as a related framework. In repeated-bet settings, he writes, the goal is not to make the most money on a single try. The goal is to stay in the game for the long run.
He does not ask readers to memorize the formula. Instead, he offers a restaurant analogy. Imagine running a small restaurant and deciding how much fresh inventory to buy for a weekend that you expect to be busy. If your goal is to keep the business alive, the first step is to set aside next month’s rent and employee wages. That money never enters the weekend purchasing budget. Even after that, you still would not spend all the remaining cash on ingredients. You would also think about the probability that customers show up and the loss you would take if the food does not sell.
The analogy is not presented as a mathematical substitute for Kelly. Jason uses it to reinforce a sequence: first take out the money you cannot afford to lose, then talk about how strong your conviction is. Position size is not a statement of confidence. It is a design choice about whether you can continue after being wrong.
What he does first after a major drop
Jason says that when a sharp sell-off hits, he first tries to separate urgency from emotion. If the account contains borrowings, margin obligations, near-term liabilities, or even cash needed for daily life, the issue is about survival. That cannot be solved by deep breathing, and it is not something that “just wait and see” can fix.
If there is no immediate risk of being forced into action, he steps away from the screen for 15 minutes. This is not, he says, a rule for timing a rebound. It is simply a way to keep himself from inventing a new rationale during the single most emotional minute of the day.
After that, he takes a sheet of paper and divides it into two sides. On the left: facts that have already happened. On the right: things he is afraid might happen. Facts and fear are often not the same thing, he writes. He then checks account balances and cash flow. Where is the money needed for life over the next six to twelve months? Is there a continuing source of cash flow? Are there short-term debts? If he decides to exit, is there enough liquidity in the asset to do so?
Not being forced to sell does not mean the price will definitely recover. It means something else that he views as critical: time. Time lets an investor continue to make judgments instead of letting the market decide for them.
He also runs what he calls a “zero-position test.” Assume he owns none of the asset right now. Ignore the cost basis. Ignore how long it may take to get back to breakeven. Looking only at the current price, the information available now, and future cash flow, would he still choose to take the same risk? He says the question is uncomfortable because it forces an admission that cost basis is a personal memory, not a market fact.
The test does not automatically say buy or sell. It shifts the question from “when do I get back to breakeven?” to “from here forward, is this still worth the risk?” In Jason’s view, real stop-loss planning should not begin on the day of a crash.
A checklist for stop-losses, profit-taking, and anxiety
Before taking on a major risk, Jason suggests writing down five sentences:
- Why am I willing to take this risk?
- Which facts does my original judgment depend on?
- What change in facts would require me to reassess?
- How long am I willing to wait?
- If I am wrong, do I still have another chance?
He says stop-losses do not need to be defined only by a percentage decline in price. A change in the underlying facts is one kind of stop-loss. Funding pressure and cash-flow pressure are another. Time is another: if the expected development still has not happened after a long period, that matters too. And if a single risk factor has grown so large that it now determines the whole outcome, that also qualifies.
He applies the same reasoning to taking profits. Booking gains is not, in his view, a certificate proving the judgment was correct. The more useful question is whether most of the good news is already in the price, whether a profitable position has grown too large as a share of total risk, and whether the original odds are still there from this point onward.
On anxiety, Jason writes that if a position keeps a person awake, makes them check prices every few minutes, and leaves them searching only for information that supports their existing view, that does not automatically mean they must exit. It does suggest, at minimum, that the risk may have exceeded their psychological budget. Emotion is an alarm, not a directional signal. One response is to turn off unnecessary price alerts for a while. Another is to ask someone who is not involved in the trade to check just one thing: have the facts and the invalidation conditions written down earlier actually changed?
These steps do not guarantee that the investor will be right. They serve a narrower purpose. They can reduce the number of improvisational decisions made at the point of maximum fear.
His closing point
Jason closes with a simple idea. The first step in non-consensus thinking is not to oppose others for the sake of it. It is to step back when everyone is crowded into the same direction and look again at price, leverage, and who may be forced to act.
He leaves readers with two questions: in a big drawdown, which mistake are you more likely to make? Do you move too quickly in panic, or wait too long because you do not want to admit you were wrong?
The article notes that the data run through the July 16, 2026 close in South Korea and the July 17, 2026 close in the United States. The main data sources listed are the Financial Services Commission, the Bank of Korea, FRED / Nasdaq, Yonhap, Asia Business Daily, and Hong Kong Exchanges and Clearing.