Information Technology · resolves 25.02.06 · closed

In January 2025, after DeepSeek R1, would a hyperscaler cut its 2025 AI capex plan?

DeepSeek's R1 model, reportedly trained for a fraction of US labs' cost, wiped about $600B off Nvidia on 2025-01-27 and raised the question whether Microsoft, Alphabet, Amazon and Meta still needed to spend so much on data centers.

▶ N0S Investor's videos on this

Outcome: No · resolved 25.02.06 · https://www.gdnonline.com/Details/1344313/Tech-megacaps-plan-to-spend-more-than-$300bn-in-2025-as-AI-race-intensifies · a stance dated before that is a call, marked right or wrong below

Where the experts stand

Yes

0 on record 0%

2 on record 100%

Where the public stands

Yes

0 answers —

0 public answers · one per account Closed to new answers.

Yes — 0 on record Wrong

No one on record yet.

No — 2 on record Right

Satya NadellaExecutive · BEATS 8

Jevons paradox strikes again! As AI gets more efficient and accessible, we will see its use skyrocket, turning it into a commodity we just can't get enough of.

said 25.01.27 on X (via NPR) · source

Mark ZuckerbergExecutive · BEATS —

I continue to think that investing very heavily in CapEx and infra is going to be a strategic advantage over time.

said 25.01.29 on Meta Q4 2024 earnings call · source

How it resolves · 25.02.06 · Yes if any of Microsoft, Alphabet, Amazon or Meta, on its late-January / early-February 2025 earnings call, lowered a previously stated 2025 capex plan or guided 2025 capex below its 2024 level. Actual: none did. Meta held $60-65B (01-29), Microsoft held about $80B for FY2025 (01-29), Alphabet guided about $75B vs about $52.5B in 2024 (02-04), Amazon said over $100B vs about $83B in 2024 (02-06); combined plans of roughly $320B. · by Q4 2024 earnings calls of Meta, Microsoft (2025-01-29), Alphabet (2025-02-04) and Amazon (2025-02-06) Dispute a quote