GZC holds twenty positions across our BTT model portfolio, and we don't rank them by return expectation — bottlenecks don't work that way, and pretending otherwise would be a disservice to the framework. What we can say plainly is that GOOGL is our highest-conviction position: the name where the largest number of independent threads in our research — ecosystem, moat, and near-term setup — point in the same direction at the same time, with the fewest open questions relative to the others in our coverage universe.
This piece is not new analysis. It is a deliberate synthesis of three things we've already published on this name, written because we think the case is stronger stated together than scattered across separate reports.
Three Businesses, One Balance Sheet
Our first piece on Alphabet made the case that Search, Cloud, and custom silicon are not three separate businesses to value in isolation — they're a single flywheel where Search's cash funds the infrastructure buildout, that buildout defends Search against AI-native competition, and the same silicon investment turns Cloud into a seller of differentiated, proprietary compute rather than a reseller of someone else's GPUs. Most of the market still models these as three businesses with three separate growth rates and three separate risk profiles. We think that's the wrong lens, and it's the first reason this is our highest-conviction name rather than merely one of several well-run mega-cap holdings.
The Part the Market Still Misprices
Our second piece went further, into why that flywheel is hard to compete away: proprietary silicon that removes a margin toll most competitors still pay to external GPU suppliers, distribution across billions of devices that gives every new product a head start, and a regulatory overhang that is real but — based on the facts known to date, including the European Court of Justice's early-July final ruling on the Android case — bounded rather than existential. A bounded, quantified legal cost is a very different investment situation than an open-ended one, and we think the market has been slower to reprice that distinction than it should be.
Our most recent piece turned to the near-term setup: Cloud backlog continuing to outrun the market's model of the business, capital expenditure plans stepping up again into 2027 rather than moderating, and a technical structure that has held up through a volatile period for AI-adjacent names. None of that changes the long-term thesis — it's simply evidence, quarter after quarter, that the long-term thesis is playing out roughly on schedule.
What Would Change Our Mind
High conviction is not the same as unconditional conviction, and we think it's worth stating plainly what would move us off this position rather than only what supports it. We would reassess if Cloud backlog conversion began to slow rather than accelerate, if a future regulatory remedy moved from a bounded fine or data-sharing requirement toward something structurally closer to a breakup, or if a competitor demonstrated a genuine silicon-cost advantage that closed Alphabet's TPU-driven gap. None of those conditions are present today, but naming them is part of how we hold a high-conviction position honestly rather than dogmatically.
GZC Thesis Summary
GOOGL is our highest-conviction position because it is the rare name in our coverage universe where the ecosystem argument, the moat argument, and the near-term setup all point the same direction without requiring us to look past a significant open question in any one of them. We track GOOGL as our highest-conviction position in the Technology pool, and will continue to publish updated research on the name as the thesis develops.
