Nvidia CEO sees $3‑$4T AI data‑center spend by 2030
Get SkimNews daily
A curated daily digest across 7 categories: geopolitics, tech, finance, health, energy, sports, culture. Free.
- Jensen Huang said data‑center spending will reach $3 trillion to $4 trillion annually by 2030, implying 32%‑41% annual growth.
- Nvidia supplies GPUs and networking gear that account for over 50% of data‑center spend, according to McKinsey, Bernstein and TD Cowen estimates.
- Reasoning models such as GPT‑5 are more compute‑intensive than earlier models, driving higher demand for Nvidia GPUs for training and inference.
- Nvidia claims its systems deliver the lowest cost per token and the highest revenue per token‑watt for cloud providers.
- Nvidia provides full rack‑scale AI solutions (GPUs, CPUs, networking) plus a robust software ecosystem, giving it a moat over custom‑chip competitors.
- Wall Street projects Nvidia’s earnings to grow at 38% annually over the next three years, making its 37× earnings valuation appear attractive.
Why it matters: Investors gain confidence as Nvidia’s projected $3‑$4 trillion data‑center market could triple current spending, boosting demand for its GPUs and rack‑scale solutions, while rivals relying on custom chips may struggle to match Nvidia’s token‑per‑watt efficiency and integrated software ecosystem, and could see reduced margins.
