AI and Growth
Everyone is arguing about whether Google's AI models are “better” but that misses the real advantage entirely.
Google’s strength isn’t model quality. It’s the plumbing underneath: massive, stable, battle-tested infrastructure. Their TPUs aren’t flashy, but they rarely fail mid-training. That one boring detail lets Google iterate faster than teams chasing raw performance. In AI, steady progress beats brilliance that crashes halfway through a run.
Google can lose money on AI for a very long time and still win. Search pays for the whole operation. It’s the same move they pulled with Chrome and Android. They didn’t need to be first. They just needed to be the one that never runs out of fuel.
The contrarian truth is this: innovation doesn’t always beat scale. Sometimes the company with the most reliable infrastructure and the longest financial runway quietly walks away with the market.
Google doesn’t need Gemini to be perfect. Their structure does the heavy lifting, and by the time the industry notices, the gap is already wide.