Nvidia remains the biggest name in AI hardware, but the infrastructure boom now extends far beyond GPUs. Investors can get exposure through networking, memory, power, cooling and data centers: areas that may benefit as AI computing demand grows.
Nvidia’s Data Center revenue reached $75.2 billion in fiscal Q1 2027, up 92% year over year. At the same time, the IEA expects global data-center electricity consumption to rise from roughly 485 TWh in 2025 to about 950 TWh by 2030.
That means the AI trade is increasingly becoming an infrastructure trade.
Where to invest beyond Nvidia
| AI layer | What it provides | Examples |
|---|---|---|
| Compute | GPUs and accelerators | Nvidia, AMD |
| Networking | Connects large AI clusters | Broadcom, Marvell |
| Memory | High-bandwidth memory | Micron |
| Power & cooling | Electricity and thermal management | Vertiv, Bloom Energy |
| Data centers | Physical AI capacity | Equinix, Digital Realty, CoreWeave |
Coinpaper’s AI chips guide covers the semiconductor side of this stack, while its data centers analysis looks at companies providing computing capacity.
Power may become the key bottleneck
Building more AI servers does not help if data centers cannot secure enough electricity.
The IEA expects data-center power consumption to nearly double by 2030. This could benefit companies supplying electrical equipment, cooling systems and onsite generation.
Vertiv provides power and thermal-management equipment for high-density data centers. Bloom Energy offers onsite fuel-cell generation, an increasingly relevant option where grid capacity is limited. Coinpaper recently examined the AI-driven opportunity for Bloom.
Rising memory costs also show how bottlenecks can move through the stack. Coinpaper’s recent server costs coverage highlighted higher Nvidia system prices as memory became more expensive.
A simple way to compare AI infrastructure stocks
Investors should look beyond revenue growth alone.
One useful screening metric is:
AI Growth Efficiency = AI-related revenue growth ÷ forward P/E
Example:
- Company A: 40% growth ÷ 40× P/E = 1.0
- Company B: 25% growth ÷ 20× P/E = 1.25
A higher score can indicate that investors are paying less for each percentage point of expected growth. It is not a complete valuation model, but it can help compare companies operating in different parts of the AI stack.
Other factors matter too:
- AI revenue growth
- Free cash flow
- Customer concentration
- Capital spending requirements
- Valuation
- Exposure to infrastructure bottlenecks
The main takeaway is simple: Nvidia may remain at the center of AI computing, but the buildout increasingly depends on companies supplying memory, networking, electricity, cooling and physical data-center capacity.
For investors looking beyond GPUs, those infrastructure bottlenecks may offer some of the most interesting long-term AI opportunities.
