I don't usually write about semiconductor stocks. My home base is oil and gas - E&P companies where cash flow per barrel and midstream operators where fee-based predictability drive every judgment call. But value discipline is not sector-specific. When a stock jumps 19 percent in a single day because of a 'breakthrough' narrative, the same questions apply: are those earnings real, is the cash flow durable, and is there a margin of safety between today's price and the underlying business?
Hong Kong-listed chipmakers surged last Friday after Huawei unveiled what it calls 'Tao's Law' - a new chip design framework that replaces the traditional race toward smaller transistors with system-level optimization, focusing on reducing signal latency, advanced packaging, and 3D stacking. SMIC, China's largest contract chipmaker and the beneficiary of all this enthusiasm, soared roughly 19 percent in one session. The market heard 'breakthrough.' The numbers tell a different story.
Let me start with what 'Tao's Law' actually is. This is not a manufacturing breakthrough. Huawei cannot access extreme ultraviolet lithography tools due to U.S. sanctions - the equipment that TSMC and Samsung use to etch chips at 3-nanometer and beyond. Instead of waiting for hardware that will not come, Huawei's engineers have shifted the problem: optimize chip performance at the system level rather than the transistor level. The company plans to launch its first Kirin chip using 'logic folding' technology in autumn 2026. By 2031, Huawei projects that chips built on this approach could reach density equivalent to roughly 1.4-nanometer process technology.
That is a five-year timeline, not a revenue event. And the company that has to actually manufacture these chips - SMIC - is a cash-flow black hole at its current valuation.
SMIC reported full-year 2025 revenue of $9.33 billion, up 16.2 percent from the prior year. Gross margin improved to 21 percent from 18 percent. On the surface, growth and margin expansion look like cause for celebration. Now let's talk about what that revenue pays for. SMIC spent $8.1 billion on capital expenditures in 2025 - and from 2020 through 2024, its total capex of $31.4 billion essentially equaled its total revenue of $31.0 billion. The company has spent nearly every dollar it earned just building the factories needed to stay relevant. Free cash flow - operating cash flow minus capex, the actual cash left for shareholders and debt service - remained negative in 2025.
Then in Q1 2026, SMIC eliminated its dividend entirely to maintain its capex spending pace. Q1 revenue was $2.5 billion, up 11.5 percent year over year, but profit missed analyst forecasts of $215 million, coming in at $197 million. The growth is real, but it does not reach the bottom line.
From a valuation perspective, shares trade at approximately 8.8 times trailing sales and roughly 19 times EV/EBITDA (enterprise value divided by earnings before interest, taxes, depreciation, and amortization - a cash-earnings proxy). With a $82 billion market cap against $9.3 billion in annual revenue, the market is pricing this company as though Tao's Law delivers profits tomorrow rather than five years out. Compare that to TSMC, which trades at roughly 15 times forward earnings and a materially lower EV/EBITDA multiple, while generating actual free cash flow in the tens of billions of dollars annually. TSMC's FY2025 revenue topped $100 billion with gross margins in the mid-50 percent range. The comparison is not flattering.
Hua Hong Semiconductor, the second-largest HK-listed beneficiary of this rally, is worse on the economics. Q1 2026 revenue was $660 million with gross margin at just 13 percent and net profit of $21 million - a 3 percent net margin on thin, cyclical work. Yet shares trade at 43 times EV/EBITDA and 12 times sales. That is not a valuation gap - it is a narrative premium with no cash-flow anchor.
The strongest argument for these names is geopolitical conviction: if China achieves semiconductor self-sufficiency, SMIC becomes an irreplaceable national asset, and market logic stops applying. While it's true that state support provides a floor for SMIC that a normal company would not have, I would argue that being bailed out is not the same as being well-run, and a stock that depends on subsidy rather than cash generation is a value trap dressed in patriotic certainty. Value investing is not about buying stocks with strategic importance - it is about buying stocks that generate cash, trade below their intrinsic value, and carry a margin of safety.
Even if Tao's Law works exactly as Huawei promises, the timeline is 2026 for a first product and 2031 for competitive density. That means at least five years of negative free cash flow, massive capex, and no dividend before the thesis delivers. At 8.8 times revenue, the market has already done the waiting. The upside is priced in; the risk is not.
This does not mean Chinese semiconductors are a bad bet. It means the risk-reward at today's prices does not justify the position. There are better opportunities elsewhere - in sectors where cash flow is positive, valuation is anchored to earnings, and the market has not already run up to a five-year-away promise.
All things considered, I would rate the current rally in HK-listed chipmakers a clear Sell for anyone who follows cash flow rather than headlines. The optimism is real. The margin of safety is not.