$1.3 trillion evaporated from the chipmaker market: the real threat is not in semiconductors
Over the past month, the market capitalization of leading chip manufacturers has shrunk by more than $1.3 trillion. However, contrary to the panic in the market, the fundamental problem lies not in the semiconductor manufacturing sector. The market, as often happens, is looking in the wrong place.
TSMC's Records Amidst a Sell-Off: Paradox or Signal?
Amidst a massive stock sell-off, Taiwanese giant TSMC reported outstanding quarterly results. The company's revenue reached $40.2 billion, and the growth forecast was raised to above 40%. Management described confidence in the artificial intelligence (AI) megatrend as "very high." TSMC's profit soared by 77%, hitting an all-time high, and margins exceeded all previous company records. Moreover, TSMC announced additional investments of $100 billion in its Arizona plants.
How is this possible? A simultaneous sector collapse and a brilliant report cannot both be true regarding demand. The reality reflects the report: chips are being shipped, demand is accelerating, not fading. Chips themselves have ceased to be a scarce resource for AI. Now, the industry's development hinges on steel, copper, and, most critically, time.
Bottlenecks: From Packaging to Transformers
The first bottleneck, according to independent analysts, has become chip packaging. The capacity of CoWoS technology (which connects the processor to memory) is already insufficient: Nvidia buys up about 60% of this capacity, and demand for it has tripled in two years. TSMC is doubling production in this area, but lines are still running at full capacity.
However, the second, much more serious problem is energy. A finished chip is useless without electricity, and the industry is acutely short of it. Data center statistics confirm the shortage of power capacity. Companies announced 16 gigawatts of capacity in the US for 2026, but are only building 5 of them: the remaining projects are frozen, and 25% don't even have a power supply plan.
The root of the problem lies in the high-voltage transformer, which converts grid power into data center supply. Currently, its manufacturing takes an average of 48-60 months, whereas before 2020, 12 months sufficed. The shortage will persist for a long time. The four manufacturers of such units — Hitachi, Siemens Energy, GE Vernova, and ABB — have taken orders for years ahead. At Siemens Energy alone, the order backlog reached nearly €136 billion.
The problem begins at the raw material level. Special electrical steel for transformers is produced by only five companies worldwide, and they are unable to quickly ramp up production.
Circumventing the shortage is also not working. Some data center operators decided to supply their own energy and ordered gas turbines, but GE Vernova, Siemens, and Mitsubishi have already sold their capacity for about five years ahead or more.
Conclusion: The AI Story is Now a Story About Electricity
Putting it all together, we get a frightening picture. Chips are being produced without delays, packaging capacity is occupied, new transformers will not appear before 2029, and turbines not before 2030. The largest construction project in economic history is being held back by heavy electrical equipment that is produced almost manually.
Traders were dumping stocks out of fear of falling demand, but the data suggests the opposite — a supply shortage. Transformer and turbine manufacturers, grid companies, and owners of sites with ready power connections could benefit from this.
The world's most advanced companies are forced to build their own power plants to run neural networks. The AI story is now not about intelligence, but about electricity: the market was counting chips, when it should have been counting substations. Investors who have realized this are already looking for assets in "heavy" energy infrastructure.
My expert opinion: The market made a classic mistake by focusing on the "digital" side of AI and ignoring the "physical" side. While attention is fixed on new processors, the real bottleneck is basic industrial components that cannot be scaled up quickly. Those currently betting on electrical equipment manufacturers and infrastructure companies may end up in a more advantageous position than those chasing volatile chipmaker stocks.