GOOGL, AMZN: Plays on inference and bundling

We see the power moving back towards the Hyperscalers, and find the combination of less risk, high growth and reasonable valuations to be compelling.

AMZN: sensible upside to $305 (+20%)

• SOTP model and ~26x GAAP EPS multiple on 2027 EPS estimates of $11.75.

Google: sensible upside to $400 (+20%)

• Lots of ways to value Google higher. DCF, or EV/OIBDA multiple. However, we prefer simple PE. $15.30 in 2027 GAAP EPS, growing 24-25% CAGR for at least the next three years, thus 26x 2027 EPS.

There is a strong possibility that CAPEX will be higher than the market expects for the next two years, thus increasing the revenue growth that Google Cloud and AWS produce. They can clearly beat numbers, however this may come with a continued derating, such as we have seen with NVDA, because of a peak earnings assumption? 

What happens if there is a pause / digestion phase?

• Hyperscalers would assumedly slow capex growth, expanding FCF vs. semis which would assumedly get destroyed. 

Hence, we find the risk reward in the Hyperscalers in particular AWS and GOOG to be compelling, you get the upside, but with much lower risk on the 2nd derivative of growth.

Link to Note