Investing

Watching the screen all day feels like diligence, but activity is not an investment process. Constant price checks amplify noise, reward emotional reactions, and make a short-term move feel like new information about a long-term thesis. A process has a schedule. It defines which data matters, how often the thesis is reviewed, and what would […]

The AI trade has quietly changed its unit of account. In 2023 and 2024, scarce GPUs determined who got allocation and who won the silicon trade. In 2025 and 2026, GPUs became purchasable, but power did not. The scarce unit is now the energized megawatt: sited, permitted, interconnected, and firm. A transmission queue can take

The world is structurally short compute. Every major efficiency gain in computing history lowered the cost of a unit of compute and increased total demand. Cheaper cycles do not end the need for infrastructure. They make more applications economically possible. That is the logic behind Jevons paradox, first described in 1865. When a resource becomes

Imagine knowing in 1996 that the internet was real, but not knowing which companies would become the winners. Google did not exist. Amazon was a bookstore. The obvious names were not yet obvious. The trade was in the layers every winner would need: fiber, routers, hosting, and the infrastructure that moved data. Those suppliers could

In 1996, you knew the internet was real, but Google did not exist and Amazon was still a bookstore. The trade was not picking the eventual logo; it was owning the layers every winner would need: fiber, routers, hosting, and the supporting infrastructure. The AI economy is at its own 1996. Applications are everywhere, business

What is the next great technology? I focus on AI investing, and my other half @FrankDPrestia focuses on AI technology and infrastructure. He’s an AI strategist and innovation theorist.

AST SpaceMobile just filed to test spectrum it doesn’t own. That’s not a mistake. That’s the signal. The bands817824 and 862869 MHz, sit directly against the spectrum AT&T and Verizon already leased them. Same antennas, same satellites, wider pipe. Nobody spends regulatory effort testing spectrum they have no path to using. Testing comes before deals.

A rising account balance can quietly rewrite an investors self-image. Good outcomes become evidence of superior judgment, sizing increases, and the bear case gets treated as an obstacle rather than a requirement. That is how a winning streak turns into a larger future mistake. Markets do not reward confidence by itself. They reward a process

The market often prices a delay before it prices the milestones that resolve the delay. That distinction matters in infrastructure businesses. A project can miss a near-term expectation while the underlying conversion remains on schedule. When the stock trades the delay, investors are given a chance to separate calendar risk from execution risk. For DGXX,