Whoa, only 2.2% of households in the U.S. are willing to pay for AI. Seeing this data really shocked me. AI feels like it’s everywhere now, but on the consumer side, it’s just getting started. Check out these comparisons to see how crazy it is: Only about 2.2% of U.S. households are paying for AI subscriptions. SiriusXM: 25%. Cloud storage: 55%. Streaming services: 91%. In other words, the current household AI subscription penetration rate is only 1/25 of cloud storage and about 1/40 of streaming services. The average monthly spending of U.S. AI paying users was around $22 in May 2024, and by May 2026, it’s projected to rise to $31—a 41% increase in just two years. Even though the pay rate is only 2.2%, individual spending by paying users is already noticeably increasing. Let’s do some quick math: with roughly 133 million households in the U.S., a 2.2% penetration rate means only about 2.9 million households are paying for AI. If AI adoption reaches SiriusXM’s 25% penetration rate, that’s over 33 million households—about 11x growth. If it hits cloud storage’s 55%, that’s over 73 million households—about 25x growth. If it matches streaming services’ 91%, that’s 120 million households—about 41x growth. Meanwhile, AI infrastructure spending is still skyrocketing. The five major hyperscale cloud providers are expected to spend around $780 billion in capital expenditures by 2026, and possibly over $1 trillion by 2027. The share of GDP spent on data center construction is already being compared by a16z to the U.S. railroad boom back in the day. On one side, we’ve got trillion-dollar capital expenditures. On the other, only 2.2% of households are actually paying for AI. This is just the beginning…