Deutsche Bank: Google Cloud Revenue Significantly Underestimated, Earnings Focus on Cloud Demand
Claire Weston
Deutsche Bank raised its Q2 Google Cloud revenue growth forecast to 70% and projects 2027 cloud revenue at $190–195 billion — nearly $50 billion above the Street's $142 billion consensus, suggesting the market is drastically underpricing Alphabet's cloud trajectory.
What exactly is the market missing?
Deutsche Bank's core thesis: the market is fixated on capital-expenditure pressure and overlooking how fast Google Cloud revenue is converting. This means → investors are watching the spending line but ignoring the payback pace.
DB forecasts 2027 Google Cloud revenue at $190–195 billion; consensus sits at just $142 billion — a gap of nearly $50 billion.
At a 30% margin, that extra revenue alone would add roughly $15 billion in operating profit — about $1 per share in incremental earnings. In plain terms = the cloud revenue the Street hasn't modeled could lift EPS by a full dollar.
What justifies a 70% growth forecast?
AI compute demand remains supply-constrained — Google recently leased part of SpaceX's compute capacity to fill the gap.
As of Q1-end, Google Cloud's backlog reached $462 billion, up 400% year-on-year. This means → customers have already placed orders; revenue recognition is largely a matter of delivery schedule.
Several hyperscalers, including Amazon, have begun raising prices on select cloud services — a signal that pricing power is shifting toward sellers across the industry.
With capex this high, where does the money come from?
DB projects 2027 Alphabet capex at roughly $325 billion (up from a prior $250 billion estimate), rising to $365–370 billion in 2028.
Financing capacity is keeping pace: this quarter alone, Alphabet completed an estimated $65–70 billion in financing — including a $10 billion investment from Berkshire Hathaway, two equity raises totaling about $36 billion, an ATM facility of up to $40 billion, and multi-currency senior debt issuance.
As of Q1-end the company held roughly $127 billion in cash and investments; DB estimates cumulative operating cash flow of about $420 billion through 2027. In plain terms = spending is accelerating, but so are earnings and fundraising — the cash buffer is more than sufficient.
Could build costs come in lower than feared?
DB's capex math: roughly 11.5 GW of existing compute demand plus an estimated 10 GW of new capacity in 2027, at $30–35 billion per GW in construction cost.
Google uses a hybrid deployment of TPUs — its in-house AI chips — and Nvidia GPUs. This means → it is not fully reliant on the most expensive Nvidia hardware, so unit build costs may run below what the market fears.
Will Search and Gemini be a drag?
DB's channel checks show advertisers increasingly directing budgets toward AI Overview while also spending more on Google Search and ChatGPT-related AI ad placements.
SimilarWeb data shows Google's Q2 site visits and page views improved versus Q1. DB maintains its forecast of 16.5% constant-currency year-on-year Search revenue growth for Q2.
The delayed Gemini model release is a short-term event — Sensor Tower data shows Gemini App session volumes still growing. This reflects that a shift in model-release timing does not alter Alphabet's long-term AI competitiveness.
What should investors actually watch in this earnings report?
DB is explicit: the key is not whether capex keeps climbing.
What truly determines the valuation is how management frames three things: cloud demand trajectory, backlog trends, and the pace of future revenue conversion.
This means → if management confirms strong cloud demand and continued backlog growth on the earnings call, the market may reprice Alphabet's valuation framework for the AI investment cycle.
Content is for reference only, not financial advice.