Google Cloud’s blowout revenue growth of 82% y/y, compared to the Street at 63%, wasn’t enough to satisfy investors. Even more upside came from the backlog number. Cloud backlog grew by over $50B q/q and now sits at $514B. To put that number into perspective, Google Cloud is expected to do roughly $130B in revenue over the next four quarters.

The bottom line is Cloud growth is accelerating and would be growing even faster if not for supply constraints. I don’t know when that balances out.

The trade-off is margins. Cloud margins were around 36% in June, near a peak, and the CFO said to expect them to decline in the back half of the year. The reason is Google needs to use third-party capacity, including SpaceX, to meet demand. That capacity is less profitable than using its own infrastructure.

That is not ideal, but it is also not a demand problem. It is the opposite. Google cannot build fast enough. My sense is Google Cloud growth stays above the Street for the next few years. The Street is looking for 65% growth in CY26 and 54% in CY27. CY27 likely ends up closer to this year’s growth rate given the steady state of supply constraints.

Capex is the reason the stock didn’t work. Google raised its CY26 capex guide to $195B-$205B, up from the prior $180B-$190B. The midpoint moved from $185B to $200B, an 8% increase, and about 7% above the Street at $187B.

That 7% above the Street number may not sound dramatic. It is dramatic because the base is so large. Google went from spending $53B on capex in CY24 to $91B in CY25, and is now guiding to $200B in CY26. That is growth of 63% in CY24, 74% in CY25, and 119% in CY26.

The Street was already modeling $257B in CY27, up 37% from its prior CY26 estimate. Now the CFO is saying capex will increase significantly again next year. Even using the new $200B CY26 midpoint, that $257B CY27 estimate would still imply another 28% step-up. My sense is CY27 estimates are going higher.

The second issue is the back-half ramp. Google spent $35.7B in March and $44.9B in June. To hit the $200B midpoint, capex needs to be about $119B in the back half, or almost $60B per quarter. That is about 33% above the June quarter run rate. It is also roughly $12B higher than where FactSet had September and December combined before the updated guide.

That is why capex controls the stock. Investors are not saying the spend is wrong. They are saying they don’t know where the spending curve settles and what the impact will be to free cash flow. In the June quarter, free cash flow was negative $6B. That was the first time since the company’s IPO in 2004 that it was negative. The good news is even if you model negative free cash flow for the next six quarters, Google should still have cash on hand of around $180B at the end of 2027. With $242B in cash, cash equivalents, and marketable securities, and $98B in long-term debt, Google currently has a net cash position of roughly $145B. Even under a worst-case scenario, it would exit 2027 with around $90B in net cash, still ample liquidity to operate and grow the business.

Search growth was up 16.5% y/y, slightly below the Street at 17.1%. A fractional miss, but the broader commentary was positive. Google is starting to talk more about AI Mode ads, and that is an important lever to the evolution of Search. The idea is: imagine planning a trip inside Gemini and seeing a clearly marked ad that fits the conversation, like hotel or tour guide suggestions. That is the vision for the conversational Google Search.

This reminds me of the 2010 to 2015 mobile transition, when Google slowly increased mobile ad coverage and managed search growth through the shift. AI Mode and AI Overviews give them a similar set of growth levers.

On the other hand, Google’s CFO also talked about difficult search comps ahead. While true, the Street already has deceleration modeled in the numbers, with September and December expectations stepping down to around 15% and 14%, respectively.

The bottom line is the Search setup is not broken. It missed by a fraction, the core business is healthy, and Google has new monetization levers coming through AI.

Waymo continues to be interesting in Google’s toy department and continues to make meaningful headway in the robotaxi space. They announced they have now done 1m deliveries through partners like DoorDash and Walmart. While that is probably just under 5% of Waymo’s total usage, it shows another real use case for autonomy beyond a robotaxi service.

For now, autonomous deliveries are a drop in the bucket for an already small business segment. However, longer term, the convenience delivery market is big enough to pay attention to.

Disclaimer