Goldman Sachs spotlights $37B AI cost surge and chip stock calls
Goldman Sachs CIO Marco Argenti said AI bandwidth costs will rival personnel expenses as enterprise AI spending grew from $11.5 billion in 2024 to $37 billion in 2025, driving demand for LLM observability startups. Ahead of April 13 Q1 results, GS shares rose 4.2% over seven sessions and options imply a 5.8% post-earnings swing, while the firm flags Teradyne, Applied Materials and AMD for potential outperformance.
1. AI Cost Recognition by Goldman CIO
Goldman Sachs Chief Information Officer Marco Argenti highlighted that AI bandwidth costs will need to be compared directly to personnel expenses as enterprise AI deployments expand.
2. Surge in Observability Startup Funding
Enterprise AI spending jumped from $11.5 billion in 2024 to $37 billion in 2025, prompting LLM observability startups—such as Braintrust, LangChain and OpenRouter—to raise Series B rounds of $80 million, $125 million and seek funding of $120 million respectively to help clients monitor token usage.
3. Q1 Earnings Preview and Options Activity
Goldman Sachs is scheduled to report first-quarter results before market open on April 13 and its stock has climbed 4.2% in the past seven sessions, with options pricing in a 5.8% move following the release.
4. Semiconductor Stock Recommendations
In sector strategy, Goldman Sachs analysts recommend buying Teradyne, Applied Materials and Advanced Micro Devices into Q1 earnings while advising caution on KLA Corp., Onsemi and Arm Holdings based on sub-segment outlooks.