U.S. vs. China AI spending gap widens
Speakers
Deirdre Bosa
— Anchor/Reporter, CNBC Tech Check
Carl Quintanilla
— Anchor, CNBC
Summary
- US Big Tech is projected to spend over $500 billion on AI infrastructure this year, funded by free cash flow and debt, while Chinese companies are spending a fraction ($70 billion).
- A massive divergence in market sentiment exists: AI breakthroughs are treated as "threats" to US incumbents (causing stock drops), while similar breakthroughs in China are treated as "opportunities" (causing rallies).
- Chinese development strategy focuses on "fast following" and efficiency rather than chasing the expensive "frontier," allowing them to produce competitive models at significantly lower costs.
- Databricks CEO Ali Ghodsi notes that Chinese models are "right behind" US models in capability but are "basically free," signaling potential margin compression for US firms over-spending on compute.
Trade Ideas
Deirdre notes that when Google released "Project Genie" (AI world generation), "You had Unity, Roblox, Take-Two all falling... The entire US software sector is in freefall." The market views generative AI as a displacement threat rather than an enhancement for US gaming and software incumbents. If AI can generate interactive worlds from a prompt, the value proposition of coding engines (Unity) or existing gaming ecosystems (Roblox) is perceived to be at risk of obsolescence. SHORT. Sentiment is currently punishing US software incumbents as "killable" rather than beneficiaries. AI tools could eventually prove to be margin-accretive for these companies if they successfully integrate them, reversing the narrative.
Deirdre notes that when Google released "Project Genie" (AI world generation), "You had Unity, Roblox, Take-Two all falling... The entire US software sector is in freefall." The market views generative AI as a displacement threat rather than an enhancement for US gaming and software incumbents. If AI can generate interactive worlds from a prompt, the value proposition of coding engines (Unity) or existing gaming ecosystems (Roblox) is perceived to be at risk of obsolescence. SHORT. Sentiment is currently punishing US software incumbents as "killable" rather than beneficiaries. AI tools could eventually prove to be margin-accretive for these companies if they successfully integrate them, reversing the narrative.
Up Next
This CNBC video, published February 10, 2026, features Deirdre Bosa discussing TTWO, RBLX, U. 2 trade ideas extracted by AI with direction and confidence scoring.
Speakers: Deirdre Bosa · Tickers: TTWO, RBLX, U