Meta and Microsoft report as investors question gargantuan AI investments
Coming in bot… Big Tech earnings. Microsoft’s AI-boosted cloud biz failed to make it rain last quarter. The tech titan managed to beat sales and profit growth expectations, but investors were unimpressed by its Azure cloud segment, which grew slower than expected. The stock tumbled after hours. Meantime, Meta shares spiked 5% after it crushed quarterly sales and profit expectations, with revenue up 21% year over year — to a record $48B — and profit up 50%.
All eyes on AI: Microsoft boss Satya Nadella said the company’s AI biz now does ~$13B in annualized revenue. Microsoft expects higher capital expenditures as it pours billions into AI (it committed $80B this year to AI data centers).
Meta said it expects to spend at least $114B this year in total, mostly on infrastructure (see: its plan for a city-sized AI data center). AI-powered recommendation algos have boosted its ad revenue, helping keep its 3.3B+ users hooked on its apps.
A deep sinking feeling… Since many see AI as the next big growth driver, investments in the tech have exploded. Between January and August of last year, Microsoft, Meta, Google, and Amazon shoveled a combined $125B into building out and running AI data centers, and they’re planning to spend a whole lot more (see: Stargate’s $500B plan).
But investors are starting to question the value of these massive investments after the roll-out of China’s DeepSeek model, which is said to match the performance of OpenAI’s most advanced GPT offering using many fewer and cheaper chips.
New standards require new strategies… DeepSeek pushed the boundaries with its success-for-less strategy, and it’s not the only one: yesterday Chinese tech powerhouse Alibaba touted a new AI model that it says outperforms both Meta’s Llama and DeepSeek’s V3. Now, Silicon Valley will have to rethink its strategy to justify its massive investments. Cue: chipmaker AMD published a blog post on how to run DeepSeek on AMD processors.