GOOGL

Google Cloud and Search together are an unusually strong pairing. Search remains a highly profitable core business, generating the cash that funds infrastructure investment elsewhere in the company. Cloud's improving profitability and large compute demand and backlog support continued investment in that infrastructure.

The two reinforce each other. Search throws off cash without needing much incremental capital, while Cloud consumes capital but now contributes operating income rather than only absorbing it. That shift from loss-making to profitable is what changed the investment case.

Capital expenditure has risen sharply to build AI compute capacity. The defensible version of that spend is that it is backed by compute demand and contracted backlog rather than speculative capacity-building. The bear version is that the industry is collectively over-building, and that depreciation on this capacity lands in operating expenses before the revenue does.

Both readings are live. What separates them is whether Cloud revenue and contracted backlog keep pace with the spending, not whether the spending itself is large.

YouTube is a large advertising and subscription business in its own right, and I count it in the base case. Waymo I do not. It is a real option on autonomous mobility, but its value is not something I can underwrite with any precision, so the thesis does not lean on it.

The valuation is considered reasonable relative to the quality and cash generation of the business, which is the main reason this is one of the larger positions in the book.

I do not want it to grow much further. The portfolio already carries heavy exposure to hyperscalers, AI infrastructure and mega-cap technology, and Alphabet correlates with several other holdings. Capping it is a construction decision, not a judgment about the company. A position can be a good investment and still be the wrong size.