Tech stocks are not where investors get paid for doing nothing and waiting, according to a market analysis published by TechFlowPost and written by Chaoxiang Research.
The piece cites a recent strategy report from Guotou Securities titled Investment Methodology for the Technology Industry, which set out to answer a basic question for technology investors: how to buy and how to sell tech stocks.
The report argues that mature companies such as Yangtze Power can be approached with a more traditional framework. Analysts can estimate how much they may earn each year over the next decade, then discount those future cash flows back into a present valuation. Tech stocks do not fit neatly into that model, it says, because progress in technology tends to arrive in jumps. A single catalyst can rewrite the logic of an entire sector.
As one example, the article says the market in 2019 believed artificial general intelligence was still 80 years away. In 2022, that estimate was cut to eight years. Then ChatGPT arrived in 2023 and changed the timetable again.
The article also points to A-share data. A technology stock that doubled in one year went on to fall 40% on average the next year, while only 5% of tech stocks were able to maintain growth above 30% for five straight years. The conclusion drawn in the report is blunt: money in tech stocks comes from capturing swings, not from passively sitting on positions.
The N-shaped setup: two rallies and four points
The report reduces the full cycle of a technology rally to one chart: an N shape. It marks four key points along that path, labeled A, B, C and D.
The A-to-B leg is the first rally, moving from 0 to 1. That phase is driven by narrative. A company may have no earnings yet and may not even have a finished product, but if the story is compelling enough, valuation can rise quickly. The article says the valuation approach at this stage is rough: estimate the output value of the entire industry, assign slices of that market to different parts of the chain, and cap the market-value-to-output ratio at roughly 3 to 3.5 times.
According to the report, embodied intelligence, the low-altitude economy, commercial aerospace and AI applications are all currently in the A-to-B stage.
The B-to-C leg is the correction. After the first rally runs its course and the story has been fully traded, prices retreat. The article says most tech stocks die here and never produce a second move.
The C-to-D leg is the second rally, moving from 1 to 100. That phase is powered by earnings. Companies begin to post results, penetration rises quickly, and share prices resume their climb. At the same time, price-to-earnings ratios can actually fall because profit growth is running faster than the stock price.
The report lists optical modules, PCBs, AI compute chips and data centers as examples that have already gone through the C-to-D phase.
Why point C matters most
The article calls point C the real dividing line for institutional investors.
By that stage, the stock has already fallen from point B, market sentiment is weak, the upside ceiling is still hard to define and most investors are lightly positioned. Yet this is also when earnings begin to show up, orders start to land and industry fundamentals move into a real expansion phase.
The report says point C can be identified only when three conditions appear together.
Capital spending by industry giants
First, large companies need to be committing money to the direction. The article compares capital expenditure for an industry to credit for the economy: without funding, the industry cannot get off the ground.
It uses AI as an example. From 2023 to 2024, overseas cloud providers increased capital spending and bought from overseas supply chains, which the article says helped start Zhongji Innolight. In the second half of 2024, ByteDance increased capital spending and bought domestic computing power, which the article says helped start Cambricon.
A breakout hit product
Second, the market needs a product that tears open a path for penetration. The examples listed are iPhone 4, AirPods, Model 3, ChatGPT and DeepSeek. Each one, the article says, marked the starting point of a C-to-D move.
Supply-chain execution
Third, companies along the industrial chain need to be receiving orders. Once big-company capex and a hit product form a closed loop, supply-chain companies start generating revenue and a self-reinforcing cycle can begin.
The report sums it up in one line: when all three elements are in place, investors should move in quickly, because that is the key action in capturing a major rally.
How the report thinks about point D
On the sell side, the report introduces what it calls an M-top framework. The M has two peaks. The first is a trading peak, where sentiment tops out. The second is a fundamentals peak.
For that second peak, the article tracks three signals: whether the macro economy is entering recession, whether price wars are breaking out on the supply side and whether capital spending on the demand side is starting to come down. If two or more of those three appear, the report says the industry rally is basically over.
It also argues that when a sector leader falls because of macro pressure or another outside shock, but the industry trend remains intact, that drop can become the best buying point.
The examples it gives are 2010, when a fall in the Nasdaq created a buying point in Apple; 2020, when the pandemic-driven selloff created a buying point in Tesla; and trade-war pressure, which it says also created buying opportunities in the AI technology theme.
For NVIDIA, the article says the sell case mainly depends on two conditions. One is whether the U.S. economy suffers a hard landing. If that happens, the stock should be sold because the cash flow of the five major cloud providers is closely tied to consumption, and a problem in that source of cash flow would break the broader logic. The second is whether the competitive structure deteriorates. If competition worsens, that is also a sell signal.
U.S. stock mapping and supply-chain mapping
The article says mapping A-share opportunities from U.S. stocks has long been an important investment method.
It points to Japan as one historical case. From the 1990s to 2000, the Japanese stock market fell 67%, yet Tokyo Electron, Advantest and Toshiba still performed well.
In quantitative terms, the report says the logic behind U.S. stock mapping has less to do with earnings correlation and more to do with the magnitude of gains in the corresponding U.S. names.
It adds that the most effective approach is supply-chain mapping. Zhongji Innolight and Luxshare Precision are given as examples of companies that can turn into major winners when the industry trend is rising and earnings continue to be delivered.
The article is attributed to Chaoxiang Research, with the handle @chaoxiangooo.