Post by VistaShares

AI has two halves, and the market is mispricing the one that matters. One half is brilliant and getting cheaper every quarter. The other is boring, and it is the only one with durable pricing power. The compute half is the one everyone watches. Better models, faster chips, falling cost per token. It is genuinely remarkable, and it is deflating. A new architecture or a cheaper accelerator can erase a pricing assumption overnight, and the whole complex re-rates on the next headline. The grid half is the one almost nobody is pricing. Transformers, switchgear, cable, substations. It is boring, and it is inflating. You can take delivery of an AI accelerator in about eight weeks. A large power transformer now runs as long as three years, and the US imports roughly 80 percent of the big ones. When demand like this lands on a supply chain that physically cannot flex, the scarce hardware does not deflate. It compounds. Look at the components, not the data centers. Transformer prices are up 60 to 80 percent since 2020. US demand has more than doubled since 2019, into a supply shortfall. And 70 percent of the transformers and lines already in the ground are 25 years or older. The grid doesn't just have to grow for AI. It has to be rebuilt for everything, which is a decade of demand the suppliers can already see on their order books. Here is the tell that this is structural and not a passing spike. The suppliers are filling order books and raising prices at the same time. GE Vernova's backlog has reached about 150 billion dollars. Eaton just set a record electrical backlog while its margins hit new highs. Prysmian's transmission margins went from roughly 14 to 21 percent in a single year. A record backlog normally means cutting price to win the work. Instead margins are expanding. In this buildout, the companies that make the scarce hardware do not compete on price. They set it. So here is the mispricing. The market is paying the higher multiple for the volatile, deflating half, and the lower multiple for the durable, inflating one. That gap is the opportunity. Own the cash flows that get more certain as the buildout scales, not less. We have run this play before. Rural electrification, the Interstate, Apollo. Every national buildout paid modest returns to the headline project, and durable, compounding returns to the companies that supplied the steel, the turbines, the cable, and the engineered parts. The AI era has the same shape. The models get the headlines. The grid gets the decade of backlog. This is the supercycle we have been mapping at VistaShares, owning the parts the buildout cannot happen without. Which half do you think is mispriced: the compute everyone is watching, or the grid almost nobody is? #electrification #AIinfrastructure #investing DVx Ventures Adam Patti