Is Google really losing share to ChatGPT? Barclays answers

A Barclays report indicates Alphabet's AI inference token processing has surged nearly 50-fold year-over-year, reaching 480 trillion tokens per month, significantly outpacing Microsoft's Azure. The analysis estimates Alphabet spent $750 million on inference tokens in Q1 2025, representing approximately 1% of Search revenue, suggesting manageable margin pressure despite rising AI usage and compute capex now accounting for over half of total capex spend. While AI costs remain a small percentage of overall expenses, the report emphasizes the importance of cost management as token consumption continues to increase with the rollout of new AI-driven products and agents.

Analysis

A recent Barclays report provides a quantitative rebuttal to investor concerns regarding the financial impact of AI integration on Alphabet's (GOOGL) core business. The analysis reveals a nearly 50-fold year-over-year surge in AI inference token processing to 480 trillion tokens per month, a volume estimated to be 5-6 times greater than Microsoft's Azure, reinforcing Google's significant scale advantage in AI application. Critically, the associated costs appear manageable; the estimated $750 million spent on inference tokens in the first quarter of 2025 represents only about 1% of Search revenue and 1.6% of COGS and operating expenses. This suggests that the deleveraging effect on margins, a key point of debate, is currently minimal and has not yet impacted company-level operating income. However, the report also signals a structural shift in spending, with compute capex now accounting for over half of total capital expenditures and trending upwards. Future token consumption is expected to accelerate further with the rollout of more advanced AI agents and new products like Project Astra, underscoring that while current costs are contained, vigilant cost management will be crucial as AI workloads expand.

Key Decisions for Investors

• The report's finding that AI inference costs are a mere 1% of Search revenue may alleviate near-term concerns about margin erosion, potentially supporting a more constructive view on Alphabet's stock for those who were previously cautious on this specific issue.
• Investors should closely monitor Alphabet's capital expenditure trends, particularly the proportion allocated to compute, as this is a leading indicator of future AI-related operating costs that will likely rise with increased token consumption from new products.
• While Google's dominant token processing volume suggests a strong incumbent advantage, it is prudent to continue tracking the adoption and monetization strategies of both its AI Overviews and competing platforms to assess the long-term impact on market share and profitability.