Goldman Sachs strategists expect the five largest US hyperscalers — Amazon, Alphabet, Microsoft, Oracle and Meta — to spend a combined $1.2 trillion on AI infrastructure in 2027, more than 50 percent above the roughly $800 billion projected for this year and above Wall Street's consensus of $1.1 trillion, Bloomberg reported.
Relative to GDP, that would be the biggest investment cycle since 19th-century railway construction. The pace is nonetheless decelerating: growth of nearly 100 percent in 2026 drops to 54 percent in 2027 and 12 percent in 2028, according to strategist Ryan Hammond. To earn back the outlays, the companies would need roughly $300 billion a year in AI revenue. Cloud revenue growth has accelerated from 25 percent in 2024 to 48 percent in the second quarter of 2026, but it remains unclear whether revenue growth at AI labs such as OpenAI and Anthropic is fast enough to justify the buildout.
Goldman also notes that spending now exceeds what the companies generate from ongoing operations, which implies more debt financing. Bottlenecks in power supply, skilled labour and memory chips could slow the pace further. The bank had already argued in June that consensus estimates were far too low. The numbers land in the middle of a debate about whether the AI buildout is a durable investment cycle or the early stages of a bubble.
