Sharp Tech with Ben Thompson (Preview) Meta’s Plans to Spend $135 Billion, The ‘AI Bubble’ Bubble?, Why Hyperscalers Should NOT Invest in TSMC
26 snips
Jan 30, 2026 A fast-paced look at Meta’s massive $115–$135B CapEx plans and why Wall Street suddenly cheered. A debate over whether AI investment is inflating its own hype cycle. A critique of hyperscalers co-investing with chipmakers like TSMC and what that means for industry dynamics. Brief takes on Apple’s role in the AI era and a VR sports viewing anecdote.
AI Snips
Chapters
Transcript
Episode notes
Wall Street Embraces Meta's AI Spend
- Meta has convinced Wall Street its AI spending is accretive after showing stronger-than-expected engagement and ad monetization.
- That narrative justified investors reversing last quarter's skepticism about large AI CapEx.
More Inventory In Existing Products
- Meta is finding more ad inventory in existing products by increasing engagement and ad load without losing users.
- This expands monetization similarly to past hits like Stories and Reels.
Compute Directly Improves Ads
- Meta's internal models like Gem scale with compute, linking more infrastructure spend to better ad performance.
- That provides a direct financial justification for large CapEx tied to model improvements.