The visual below shows The Business Engineer’s AI Capex Map, revealing how tech giants are placing fundamentally different bets on artificial intelligence infrastructure — as explored in the economics of AI compute infrastructure — . Three massive spending pools emerge: Microsoft at $190B, Amazon at $200B, and Alphabet at $185B.

But raw dollars tell only half the story. The capex intensity ratios expose who’s betting hardest relative to their revenue base.

Microsoft’s strategy centers on strategic partnerships, primarily its $13B OpenAI investment. They’re buying proven AI models rather than building from scratch, keeping their capex-to-revenue ratio at 13%.

Amazon takes the infrastructure-heavy approach, with $75B directed toward AWS data centers and $48B into logistics automation. Their 12% capex intensity reflects massive physical and digital infrastructure buildouts.

Google pursues the full-stack strategy, spending $52B on custom Tensor chips and $39B on research facilities. At 15% capex intensity, they’re building everything in-house from silicon to software.

The map reveals smaller players making outsized bets. Tesla allocates 18% of revenue to AI capex for autonomous driving, while traditional chipmaker Intel maintains just 8% intensity.

Meta’s $28B commitment represents 14% intensity, focused entirely on metaverse AI applications. Meanwhile, Apple’s secretive $31B AI investment accounts for 8% of their massive revenue base.

Each percentage point of capex intensity represents billions in strategic commitment. Microsoft’s partnership model requires less capital but creates vendor dependence.

Amazon’s infrastructure play demands enormous upfront investment but builds lasting competitive moats. Google’s vertical integration burns the most cash but promises maximum control.

The visual makes clear that spending levels alone don’t determine AI strategy — as explored in the AI stack war reshaping big tech — success. Capex intensity reveals conviction levels and strategic focus.

But which approach will generate the highest returns when AI revenue scales exponentially over the next five years?