# Who Gets Paid When AI Data Centers Are Built? The Supply Chain

> A layered map of the AI compute supply chain, from the companies that pay to the chips, memory, packaging, networks, power and buildings, with the latest dated number for each layer and where the bottlenecks and risks sit.

Source: https://www.exent.ai/resources/ai-compute-supply-chain · Reviewed by David, Senior Analyst, Exent · Published Oct 29, 2026 · Updated Oct 29, 2026

## Key takeaways

- When AI data centers are built, the money flows from a few buyers (Big Tech and AI labs) through clouds and neoclouds to chips, memory, packaging, networking, power equipment and buildings. Each layer has different growth, margins and risks.
- The chip layer is the largest: NVIDIA’s data-center revenue was $89.0B in the quarter to Jul 26, 2026, up 117% from a year earlier, and it guided to $108.0B of total revenue for the next quarter.
- Memory is one of the three tightest chokepoints in 2026: SK hynix, Samsung and Micron shared HBM revenue 50% / 33% / 18% in Q2 2026, and Micron’s quarterly revenue rose from $11.32B to $54.23B in a year.
- Power is the slowest layer to expand: GE Vernova’s gas-turbine backlog plus slot reservations rose to 116 GW by mid-2026, and Goldman Sachs projects US data-center power demand of 66 GW in 2027, from 31 GW in 2025.
- Same layer, opposite outcomes: in the year to Sep 30, 2026, Dell returned +284.3% and Supermicro −14.3% (Exent market data). Owning the theme does not make every supplier a winner.

## What is the AI compute supply chain?

It is the chain of companies that turns AI budgets into running GPU clusters: buyers sign contracts, clouds and neoclouds buy chips and rent buildings, and suppliers deliver processors, memory, servers, networks, cooling and electricity, funded by lenders and shareholders. The map below reads top to bottom, following the money. Scope: listed companies in the US and Asia that sell into AI data centers, with dated figures from their own releases. Excluded: software and model companies, and private suppliers.

**Where the money for an AI data center goes**

1. Who pays:
- Big Tech (Microsoft, Alphabet, Amazon, Meta, Oracle)
- AI labs (OpenAI, Anthropic and others)

2. Who runs the compute:
- Hyperscaler clouds (Own their data centers)
- Neoclouds (CoreWeave, Nebius, IREN)

3. Systems:
- Servers (Dell, Supermicro)
- Networking (NVIDIA, Arista, Broadcom)

4. Chips:
- GPUs and custom chips (NVIDIA, AMD, Broadcom)
- HBM memory (SK hynix, Samsung, Micron)
- Foundry and packaging (TSMC)

5. Sites and power:
- Data-center landlords (Applied Digital, Hut 8, TeraWulf)
- Power and cooling gear (Vertiv)
- Turbines and the grid (GE Vernova, utilities)

6. Capital:
- Lenders and bond investors
- Shareholders

Flows:
- Big Tech → Neoclouds: multi-year capacity contracts
- AI labs → Neoclouds: multi-year capacity contracts
- Neoclouds → Servers: buy GPU systems
- Hyperscaler clouds → GPUs and custom chips: buy chips and custom designs
- Servers → GPUs and custom chips: GPUs in every rack
- Neoclouds → Data-center landlords: 15–20 year leases
- GPUs and custom chips → Foundry and packaging: wafers and advanced packaging
- GPUs and custom chips → HBM memory: memory stacked next to the GPU
- Lenders and bond investors → Neoclouds: loans secured on GPUs and contracts
- GPUs and custom chips → Neoclouds: NVIDIA: equity stakes and capacity backstops

*Simplified. Arrows down follow the money; feedback loops show money or support flowing back up, such as NVIDIA investing in its own customers. Green marks the layers with the fastest reported growth in 2026; red marks visible stress (neocloud debt) or the slowest bottleneck to expand (power).*

## Who gets paid in each layer, and how much?

Every layer grew in 2026, at very different speeds. The latest figure each company reported, with its date:

| Layer | Company | Latest figure (reported) | Period | Bottleneck or risk |
| --- | --- | --- | --- | --- |
| GPUs | NVIDIA | Data Center revenue $89.0B (+117%); Q3 outlook $108.0B total | Quarter to Jul 26, 2026 | A few large buyers; fast product cycles |
| Custom chips, networking | Broadcom | AI semiconductor revenue $16.7B (+221%); Q4 outlook $21.7B | Quarter to Aug 2, 2026 | Concentrated in a handful of hyperscaler designs |
| Networking | Arista | Revenue $3.036B (+37.7%) | Quarter to Jun 30, 2026 | Competes with NVIDIA’s own networking |
| HBM memory | SK hynix | Revenue 79.3T won, operating profit 60.5T won | Q2 2026 | Memory has a long history of boom-and-bust pricing |
| HBM memory | Micron | Revenue $54.23B (a year earlier $11.32B) | Quarter to Sep 3, 2026 | Same cycle risk; capacity takes years to add |
| Foundry, packaging | TSMC | Revenue US$46.72B (+41.2%); HPC 68% of revenue | Q3 2026 | Advanced packaging capacity; geographic concentration |
| Power, cooling | Vertiv | Net sales $3,274M (+24%) | Q2 2026 | Order timing tied to data-center build schedules |
| Turbines | GE Vernova | Gas backlog + slot reservations 116 GW (83 GW at end-2025) | Jun 30, 2026 | Manufacturing capacity; years of lead time |
| Neoclouds | CoreWeave | Revenue backlog ~$104B; Q2 net loss $626M | Jun 30, 2026 | Debt, customer concentration, delivery |

The buyers at the top are few: what Microsoft, Alphabet, Amazon and Meta plan to spend is tracked in [big-tech AI spending in 2026](https://www.exent.ai/resources/big-tech-ai-spending-2026). The neocloud layer, where the financial stress showed in October 2026, is mapped in [what is a neocloud](https://www.exent.ai/resources/what-is-a-neocloud) and compared company by company in [neocloud stocks compared](https://www.exent.ai/compare/neocloud-stocks-compared).

## Where are the bottlenecks in AI data centers?

In 2026 the three tightest points are high-bandwidth memory, advanced packaging and electrical power. HBM is stacked next to each GPU, and only three companies make it at scale; Counterpoint Research put their Q2 2026 revenue shares at 50% (SK hynix), 33% (Samsung) and 18% (Micron). Packaging that joins GPUs and memory is concentrated at TSMC, whose high-performance computing business made up 68% of its Q3 2026 revenue. Power is slowest to add: GE Vernova’s gas-turbine backlog plus slot reservations rose from 83 GW at the end of 2025 to 116 GW by June 30, 2026, and the company expects at least 125 GW by year-end. On the demand side, [Goldman Sachs Research](https://www.goldmansachs.com/insights/articles/us-data-center-power-demand-projected-to-double-by-2027) projected in May 2026 that US data-center power demand would rise from 31 GW in 2025 to 66 GW in 2027 (a projection, not a measured figure). A bottleneck usually means pricing power for the supplier that owns it, until new capacity arrives.

## How do you read a supplier’s growth numbers? A worked example

Three ratios from the table tell you more than the headline: growth, concentration and margin. Each uses only the reported numbers.

| Question | Arithmetic | Result | What it tells you |
| --- | --- | --- | --- |
| How fast did Micron grow? | $54.23B ÷ $11.32B | 4.8× (+379%) | Memory prices and volumes both rose; cyclical sectors can reverse |
| How dependent is NVIDIA on data centers? | $89.0B ÷ $96.2B | 92.5% of revenue | One end market drives almost everything |
| How profitable is the HBM leader? | 60.5T won ÷ 79.3T won | 76% operating margin | Shortage-level margin; it falls if supply catches up |

Interpretation (ours): very high growth plus a very high margin is what a bottleneck looks like at its peak. It is also what a peak looks like in a cyclical industry, which is why the next step is always to ask what happens to price when capacity is added.

## How did AI supply-chain stocks perform?

Very unevenly: suppliers in the same layer moved in opposite directions, and the highest volatility was in the neocloud layer. Exent market data, year to September 30, 2026:

| Stock | Layer | 1-year total return | Largest drop in the year (daily) | Annualized volatility |
| --- | --- | --- | --- | --- |
| [NVIDIA (NVDA)](https://www.exent.ai/stocks/nvda) | GPUs | +22.7% | −20.2% | 33.6% |
| AMD | GPUs | +278.1% | −27.8% | 67.8% |
| Broadcom (AVGO) | Custom chips, networking | +7.2% | −33.0% | 46.2% |
| [Micron (MU)](https://www.exent.ai/stocks/mu) | HBM memory | +537.6% | −30.3% | 65.0% |
| [Dell (DELL)](https://www.exent.ai/stocks/dell) | Servers | +284.3% | −32.3% | 73.5% |
| [Supermicro (SMCI)](https://www.exent.ai/stocks/smci) | Servers | −14.3% | −66.2% | 81.7% |
| Arista (ANET) | Networking | +39.7% | −28.3% | 54.4% |
| [Vertiv (VRT)](https://www.exent.ai/stocks/vrt) | Power, cooling | +60.1% | −24.8% | 56.6% |
| GE Vernova (GEV) | Turbines | +54.9% | −30.6% | 53.4% |
| [Oracle (ORCL)](https://www.exent.ai/stocks/orcl) | Hyperscaler cloud | −50.6% | −58.3% | 60.7% |
| [CoreWeave (CRWV)](https://www.exent.ai/stocks/crwv) | Neocloud | −36.3% | −61.2% | 95.7% |

Past returns are history, not a forecast, and these figures end before the October 2026 selloff in neocloud shares.

## How do investors look for growth stocks in the AI supply chain?

By asking which layer is short of capacity, who owns that shortage, and how long it will last, rather than by buying the theme. The “picks and shovels” idea is explained in [picks-and-shovels investing](https://www.exent.ai/resources/picks-and-shovels-investing), and what past big winners had in common in [how to find the next Nvidia](https://www.exent.ai/resources/how-to-find-the-next-nvidia). A checklist for any supplier:

- Bottleneck: is the company the scarce input (memory, packaging, turbines) or an easily replaced one?
- Customer concentration: how much revenue comes from the four or five big buyers?
- Cycle history: memory and servers have boom-and-bust records; check what margins did in the last downturn.
- Balance sheet: suppliers paid upfront are safer than customers who borrow to buy (the neocloud layer).
- Valuation and expectations: a stock can fall on good results if the price assumed more.

## When does the supply-chain trade go wrong?

It goes wrong when spending slows or capacity catches up, and the layers with the most borrowing and the most cyclical pricing feel it first. Gartner said on October 14, 2026 that over $3.5 trillion of AI infrastructure spending could be exposed to impairment by the end of 2030 and that over 80% of neocloud providers would cease independent operations by then, which includes being acquired. A pullback by one of the few large buyers would travel down every arrow in the map. Who should not rely on this page: anyone looking for a list of stocks to buy; the map shows where to look, not what to own. Limitations: company figures use different fiscal quarters, private suppliers are missing, and market-share data come from one research firm. For sizing a position in a theme this volatile, see speculative sleeve size.

## Where Exent fits in the search

If you look for growth companies in a theme like this, the hard part is the shortlist; Exent tracks Wall Street analysts and ranks them by their past accuracy.

1. Gems are Exent’s top picks: stocks where several of the best‑rated analysts (4★ and above on past accuracy) agree. They are our hunt for the next big winners; past accuracy, not a prediction.
2. Analyst profiles show each analyst’s record, so you can see who is behind a rating on a supplier before you research it further.
3. Before buying, upload your holdings (a screenshot, CSV, Excel, PDF or Google Sheet; no brokerage login) and see whether an AI supplier joins a correlation cluster you already own.

Exent is our product: [Exent Gems](https://www.exent.ai/radar)

## Frequently asked questions

### Who supplies the AI data center boom?

Chipmakers (NVIDIA, AMD, Broadcom), memory makers (SK hynix, Samsung, Micron), TSMC for manufacturing and packaging, server makers (Dell, Supermicro), networking (NVIDIA, Arista, Broadcom), power and cooling (Vertiv, GE Vernova) and data-center landlords.

### Who makes HBM memory?

Three companies: SK hynix, Samsung and Micron. Counterpoint Research put their Q2 2026 HBM revenue shares at 50%, 33% and 18%.

### Why is TSMC a bottleneck for AI chips?

Manufacturing of advanced AI processors, and the packaging that joins them to HBM memory, is concentrated at TSMC. High-performance computing was 68% of TSMC’s Q3 2026 revenue of US$46.72B; packaging capacity takes years to expand.

### Why is power the main bottleneck for AI data centers?

Because generation and grid equipment take years to build. GE Vernova’s gas-turbine backlog plus slot reservations reached 116 GW at Jun 30, 2026, and Goldman Sachs projects US data-center power demand of 66 GW in 2027, up from 31 GW in 2025.

### Who supplies CoreWeave?

NVIDIA supplies its GPUs and is also an investor and a customer; server, networking and power-equipment makers supply the racks; landlords such as Applied Digital, Galaxy and Core Scientific lease it data-center space.

### Does owning Nvidia give me exposure to neoclouds?

Partly. Neoclouds are among NVIDIA’s customers, but data-center revenue of $89.0B in one quarter comes mostly from much larger buyers, so NVIDIA’s results depend more on Big Tech spending than on any single neocloud.

### Are AI supply-chain stocks a good investment?

There is no general answer. In the year to Sep 30, 2026, results ranged from +537.6% for Micron to −36.3% for CoreWeave in Exent market data, so the layer and the company matter more than the theme. Past returns are not a forecast.

### What is the difference between InfiniBand and Ethernet in AI clusters?

Both connect GPUs into one cluster. InfiniBand is NVIDIA’s specialized networking; Ethernet is the open standard used by Arista, Broadcom and others, and AI builders use both.

## Sources

1. [NVIDIA announces financial results for second quarter fiscal 2027](https://www.sec.gov/Archives/edgar/data/1045810/000104581026000073/q2fy27pr.htm) — NVIDIA (SEC), 2026-08-26
2. [Broadcom fiscal Q3 2026 results (Form 8-K, Exhibit 99)](https://www.sec.gov/Archives/edgar/data/1730168/000173016826000076/avgo-08022026x8kxex99.htm) — Broadcom (SEC)
3. [Arista Networks Q2 2026 results (Form 8-K, Exhibit 99.1)](https://www.sec.gov/Archives/edgar/data/1596532/000159653226000174/ex991q226-earningsrelease.htm) — Arista Networks (SEC), 2026-08-04
4. [Global DRAM and HBM market share: quarterly](https://counterpointresearch.com/insights/global-dram-and-hbm-market-share) — Counterpoint Research, 2026-09-02
5. [SK hynix announces 2Q26 financial results](https://news.skhynix.com/en/q2-2026-business-results/) — SK hynix, 2026-07-29
6. [Micron fiscal Q4 2026 results (Form 8-K, Exhibit 99.1)](https://www.sec.gov/Archives/edgar/data/723125/000072312526000018/a2026q4ex991-pressrelease.htm) — Micron (SEC), 2026-09-30
7. [TSMC 3Q26 earnings release (Form 6-K)](https://www.sec.gov/Archives/edgar/data/1046179/000104617926000726/a3q26e_withguidancexfinal.htm) — TSMC (SEC), 2026-10-15
8. [Vertiv second quarter 2026 results (Form 8-K, Exhibit 99.1)](https://www.sec.gov/Archives/edgar/data/1674101/000162828026050323/q22026exhibit991vrt07292026.htm) — Vertiv (SEC), 2026-07-29
9. [GE Vernova 2Q 2026 press release](https://www.gevernova.com/sites/default/files/gev_webcast_pressrelease_07222026.pdf) — GE Vernova, 2026-07-22
10. [GE Vernova 4Q 2025 press release](https://www.sec.gov/Archives/edgar/data/1996810/000199681026000012/gevpressrelease4q25.htm) — GE Vernova (SEC), 2026-01-28
11. [US data center power demand projected to double by 2027](https://www.goldmansachs.com/insights/articles/us-data-center-power-demand-projected-to-double-by-2027) — Goldman Sachs, 2026-05-20
12. [CoreWeave Reports Strong Second Quarter 2026 Results (Form 8-K, Exhibit 99.1)](https://www.sec.gov/Archives/edgar/data/1769628/000176962826000362/coreweave2q26earningspress.htm) — CoreWeave (SEC), 2026-08-11
13. [Gartner says over $3.5 trillion of AI infrastructure spending could be exposed to impairment by the end of 2030](https://www.gartner.com/en/newsroom/press-releases/2026-10-14-gartner-says-over-us-dollars-three-and-a-half-trillion-of-artificial-intelligence-infrastructure-spending-could-be-exposed-to-impairment-by-the-end-of-2030) — Gartner, 2026-10-14

---

Educational content, not investment advice. Past performance does not guarantee future results.