NVIDIA and….
There is an awful lot going on here, but Jensen Huang may have just given us the sentence that matters: “compute is revenue.” Nvidia has just delivered $96.2bn of quarterly revenue, up 106% year on year and 18% in a single quarter, with Data Center revenue of $89bn, up 117%, and is guiding the next quarter to $108bn. Put that growth against the valuation and the “Nvidia is insanely expensive” argument becomes rather less obvious: the shares are around 32x trailing earnings and roughly 21–24x forward earnings. This is a $5tn company growing quarterly revenues at triple digits and earnings even faster, yet the market initially throws its toys out of the pram because gross margin is 74.4% rather than the 74.8% analysts fancied. Monkeys. The bigger message is that AI has crossed from experimentation into useful, productive and increasingly profitable work. That is why I keep saying own the bottlenecks AND the applications. We need more chips, memory, power, data centres and connectivity, but increasingly the interesting question is what all this intelligence actually does. And this is where AI and digital assets start colliding. AI agents cannot swipe Visa cards, remember PINs or wait two days for a bank transfer; they need programmable wallets and instant machine-to-machine settlement. Stablecoins, blockchain rails and tokenised assets suddenly stop looking like crypto toys and start looking like infrastructure for an autonomous economy.
Meanwhile Bitcoin remains the scarce monetary asset sitting alongside an increasingly abundant world of intelligence. Forget whether Bitcoin reaches $150,000 or $300,000 on precisely somebody’s spreadsheeted date; the direction of travel matters more. Scarce compute, scarce energy, scarce digital money and an explosion in intelligent applications. Own the bottlenecks, own the rails and increasingly own what runs across them.
1
1August 26, 2026 162
