outfox

Explore the bigger story

One building. An entire supply chain.

Follow the industries behind the AI buildout, from power and materials to chips and software. Select a layer to explore its companies, prices, and context.

What’s behind the data-center boom?

One buildout. An entire economy behind it.

A data center is more than a building full of servers. It connects power producers, chipmakers, networking companies, and software businesses. Explore ten connected layers of the AI economy—and the companies behind them.

This map covers the broader AI economy, not just data-center construction. Select a layer to explore. Bar widths reflect approximate market caps with a minimum display width; bubbles represent subcategories.

Market caps are approximate sums of tracked companies per layer.
Beyond the headlines

Understand the buildout

Big questions. Plain-English answers.

What does a data center actually do?

It houses networked computers that process and store information. These facilities support services such as websites, cloud applications, and AI. Power, cooling, and networking keep the equipment working.

Source: U.S. Energy Information AdministrationExplore this layer →
Why does it need so much electricity?

Computers need power to perform their work, and cooling systems remove the heat they produce. Demand varies with the facility’s size, equipment, workload, and efficiency—not every data center has the same footprint.

Source: U.S. Department of EnergyExplore this layer →
Does every data center use water for cooling?

Cooling designs differ. Some consume water through evaporation; others use air-based heat rejection or different systems. Water circulating in a closed loop is not the same as water consumed. Electricity generation can also have a water footprint, so local design and power sources matter.

Source: Department of Energy: water and coolingExplore this layer →
What does a community gain—and give up?

Projects can bring construction work, ongoing jobs, and tax revenue. Residents may also face concerns about noise, land use, and demands on utilities. Ask for project-specific figures: permanent versus construction jobs, tax incentives, water needs, and who pays for infrastructure upgrades.

Source: Associated Press: community benefits and concernsExplore this layer →
Which companies are connected to the buildout?

Outfox’s map groups companies across power, chips, networks, software, and other parts of the AI economy. It is a research starting point—not a list of suppliers to every project or a promise that every company benefits equally.

Source: Outfox methodology and data sourcesExplore this layer →

General explanations; local projects differ. Reviewed September 3, 2026.

Selected · Processors

GPUs, TPUs, AI accelerators — the engines of intelligence

Why it matters

Accelerator competition depends on more than chip specifications: software ecosystems, developer adoption, networking, supply, and total operating cost all shape demand.

Key companies
NVDA
NVIDIA
n/a
AMD
AMD
n/a
AVGO
Broadcom
n/a
QCOM
Qualcomm
n/a

Follow the disclosures

Who's buying in this layer?

Selected · Processors. Click a stack layer to pin it while you inspect the evidence.

Traded 2026-01-012026-09-23 · House archive

Bars count distinct filers, not dollars. Purchases and sales stay separate.

AMD5 buyersAMDBuying5Selling312 purchase records · 3 sale records · View evidence ↓All AMD disclosures →
NVDA5 buyersNVIDIABuying5Selling711 purchase records · 13 sale records · View evidence ↓All NVDA disclosures →
AVGO3 buyersBroadcomBuying3Selling37 purchase records · 7 sale records · View evidence ↓All AVGO disclosures →
MRVL3 buyersMarvell TechnologyBuying3Selling04 purchase records · 0 sale records · View evidence ↓All MRVL disclosures →
INTC2 buyersIntelBuying2Selling32 purchase records · 4 sale records · View evidence ↓All INTC disclosures →

Archive refreshed 2026-09-08. Readable House filings only; Senate and insider activity are not connected. Options and unresolved stock identities are excluded. Filers may report family holdings. This shows disclosed activity, not live buying or a recommendation.

Compare investments

Market performance

Toggle equal-weight indexes for each supply-chain layer on or off.

|
Loading market history…

Layer lines are equal-weight averages of available constituent returns. Historical data via Yahoo Finance.

Processors · 9 companies

View all
N
NVDAUS
NVIDIA
AI accelerators, CUDA software, and high-speed networking
price n/a
A
AMDUS
AMD
MI series GPUs, ROCm ecosystem
price n/a
B
AVGOUS
Broadcom
Custom AI silicon for hyperscalers
price n/a
Q
QCOMUS
Qualcomm
Edge AI and mobile inference
price n/a
AH
ARMUK
ARM Holdings
Chip architecture licensing
price n/a
I
INTCUS
Intel
Gaudi AI accelerators, x86 processors
price n/a
MT
MRVLUS
Marvell Technology
Custom AI accelerators
price n/a
SM
SMCIUS
Super Micro Computer
AI server systems and GPU platforms
price n/a
G(
GOOGUS
Google (TPU)
TPU custom silicon for AI training
price n/a

Loading verified market quotes…

Layer Performance

Who's moving today

Average daily change of the companies in each layer, from live quotes.

Latest headlines

In the news

Loading headlines for NVDA