Home โ€บ Blog โ€บ Inside the AI Factory: How Electricity Becomes Intelligence
Topic Explainer ยท Infrastructure

Inside the AI Factory: How Electricity Becomes Intelligence

Nvidia calls them AI factories: buildings whose raw material is electricity and whose product is tokens. In 2026 the four biggest cloud companies plan to spend around $700 billion building them. This guide walks through the production line, the gigawatt campus, the 600-kilowatt rack, the power fight, the money and the backlash, and separates what is measured from what is marketing.

FL
FrontierAGI Team
โš ๏ธ How this was researched Everything here comes from web searches run up to 23 October 2026. Much of the technical framing comes from Nvidia and its partners, and performance claims such as "50 times more throughput per megawatt" are vendor figures. Site capacities come mainly from Epoch AI's tracker; spending figures come from company guidance as reported by the press, and some sources mix calendar and fiscal years. The economics calculator uses an illustrative efficiency figure you can change, not a measurement. See Reading the numbers for conflicts.

From data center to AI factory

A traditional data center stores and serves information: email, web pages, databases. Its servers sit mostly idle and its value is measured in uptime. An AI factory, in Nvidia's framing, does something different: it converts electricity into tokens, the units of text, code, images and actions that AI models produce. Jensen Huang's line at GTC 2026 was that data centers "are truly becoming AI factories", and his formula for their business was blunt: revenue = tokens per watt ร— available gigawatts. At Nvidia's 2026 shareholder meeting he called every token "a profit unit".

The metaphor is useful because it changes what people measure. Instead of servers and square metres, operators talk about tokens per second, tokens per watt, cost per million tokens and how much of the time the chips are actually busy. It also explains why power has become the scarce input: if output scales with watts, the company with the most energy wins.

๐Ÿ’ก Branding or real change? Both. As one independent outlet put it, Jensen Huang did not invent GPU-dense data centers; companies were already building them, and "AI factory" is partly a marketing frame that happens to sell more Nvidia hardware. But the engineering really has changed: rack power has gone up more than tenfold in four years, liquid cooling has become mandatory, and whole campuses are now designed around a single workload.
~$700B2026 capex planned by Amazon, Alphabet, Meta and Microsoft combined (press reports of guidance)
~946 MWoperating IT power at xAI's Colossus 2, the largest tracked AI site (Epoch AI, Oct 2026)
600 kWplanned power of one Nvidia Kyber rack in 2027, against about 40 kW for a Hopper rack
~945 TWhIEA projection for data center electricity use in 2030, about double 2024

The token production line

Follow one kilowatt-hour through the factory. Click a stage to see what happens there and what it is measured by.

From electrons to tokens

Simplified. The moving dashes show energy and data flowing; the bottom row is where most energy leaves as heat.

Anatomy of a gigawatt campus

A gigawatt is roughly the output of a large nuclear reactor. A campus built around it looks less like an office building and more like a power plant with computers attached. Click the parts of this schematic site.

A schematic AI campus

Not to scale and not any specific site. Based on the subsystems named in Nvidia's Vera Rubin DSX reference design and partner announcements.

Nvidia's March 2026 Vera Rubin DSX AI Factory reference design and Omniverse DSX Blueprint try to standardise this whole stack, with contributions from Schneider Electric, Siemens, Vertiv, Eaton, Trane, Jacobs and others. The blueprint is for "physically accurate digital twins": simulating the power, cooling and layout of a site before pouring concrete. Nvidia's stated reason is that energy is now the main constraint, citing more than $300 billion of equipment backlog and over 200 gigawatts of projects waiting in US grid interconnection queues.

The rack revolution

The single biggest physical change is how much power one rack draws. Air cooling runs out at a few tens of kilowatts per rack; everything above that needs liquid flowing directly to the chips.

Power per rack by Nvidia generation

Approximate figures from trade press and vendor-adjacent analyses. Vera Rubin is an estimate; Kyber is a target for 2027, not a shipping product.

Sources: DCD; CoreAdvisors density chart; Cyfuture and Introl analyses.

The four networks inside the factory

A factory is only as fast as its slowest link. Networking is now described in layers, each with its own technology. Click a layer.

Scale-up, scale-out, scale-across, scale-in

"Scale-in" is a 2026 Nvidia addition that analysts describe as mostly vendor framing.

Power is the product

If tokens scale with watts, electricity is the true bottleneck. The International Energy Agency's 2025 Energy and AI report put data centers at about 415 TWh in 2024, roughly 1.5% of world electricity, heading to about 945 TWh by 2030, with AI-optimised data centers more than quadrupling. Its 2026 follow-up, as reported, says data center demand grew 17% in 2025 against 3% for electricity overall, and that AI-specific demand will triple by 2030.

Data center electricity use, IEA

TWh per year. 2024 baseline and 2030 projection from the IEA's 2025 report; the 2025 value applies the reported 17% growth to the baseline and is our approximation.

Where the power comes from

Click an option.

The world's biggest AI factories

No single site runs a full gigawatt of IT power yet, according to Epoch AI's tracker, but several are close and the plans are far larger. Solid bars show estimated operating power; the lighter bar shows the planned figure.

Operating vs planned power (MW)

Epoch AI estimates as of October 2026 unless noted. Hyperion's 5 GW is a long-range goal; Epoch's own projection is 1,632 MW for 2028.

Sources: Epoch AI largest data centers, Epoch AI on Stargate; Oracle's September earnings call for Abilene.

Sovereign AI factories

Governments now treat AI factories as strategic infrastructure, like ports or power stations.

National and regional projects

WhereWhatStatus
European UnionAI Gigafactories call for up to seven sites, each with at least 100,000 advanced processors (about four times Europe's largest today), backed by up to โ‚ฌ10 billion of EU and national money and expected to unlock at least โ‚ฌ20 billion of private investment. Letters of intent signed with AMD, Nvidia and Qualcomm.Call opened late July 2026; closes 12 Nov 2026; awards expected early 2027
GermanyDeutsche Telekom and Nvidia's Industrial AI Cloud in Munich, starting with 10,000 Blackwell GPUs; described as Europe's first industrial AI cloud.Announced; exact launch date not confirmed in our searches
IndiaReliance's Jamnagar campus with Nvidia Blackwell chips: 1 GW in recent reports, earlier reports said 3 GW and $20โ€“30 billion; Reliance has said it will finish within 24 months.Under construction; scale figures vary by source
PortugalMicrosoft plans a $10 billion data center investment, cited as part of private gigafactory plans in Europe.Announced

One analysis expects Europe's first gigafactories to run mainly on US chips from Nvidia and AMD, which tempers the word "sovereign": the buildings and data are local, the silicon mostly is not.

The money

The spending is the largest private infrastructure build in modern history. Press reports of company guidance put 2026 capital expenditure for the four biggest US cloud companies at roughly $700 to $725 billion, up about 77% from around $410 billion in 2025.

2026 capex guidance by company (US$ billions, midpoints)

Microsoft's figure is disputed between sources (calendar vs fiscal year); Microsoft says about $25B of its total reflects component price inflation, not extra capacity. Amazon's includes its Leo satellite business.

Factory calculator: what must a token cost?

Work out the economics of a site yourself. The calculator asks what revenue each megawatt-hour of busy computing must earn to pay back the building and chips over their life, then turns that into a break-even price per million tokens using an efficiency figure you choose.

Break-even price of a token

Simplified: straight-line payback of capex, plus electricity and an operations allowance; no financing cost, taxes or profit. The tokens-per-kWh slider is an illustrative assumption: real efficiency varies by orders of magnitude with model size, hardware and batch settings.

One estimate: ~$38B/GW
The depreciation debate: 2โ€“3 years (Burry) vs 5โ€“6 (hyperscalers)
Share of the year the chips do paid work
Log scale, 10 thousand to 100 million

Two lessons fall out of playing with it. First, capex dominates: electricity is usually a small share of the cost per token, which is why operators pay almost anything to get power sooner. Second, payback period matters enormously: moving from six years to three roughly doubles the price each token must fetch, which is exactly what the depreciation debate below is about.

Training factories vs inference factories

Deloitte, as widely cited, estimates that inference (running models for users) is about two-thirds of all AI compute in 2026, up from a third in 2023 and half in 2025. Training still grows at the frontier, but the mix is shifting, and the two jobs want different factories.

Training factoryInference factory
ShapeOne enormous, tightly coupled clusterMany smaller pools serving independent requests
LocationWherever power is cheapest and largest; latency to users irrelevantCloser to users and networks; latency matters
NetworkHuge scale-out and scale-across bandwidth; one failure can stall the jobScale-up within a rack matters most; jobs fail independently
Power profileSynchronised swings as thousands of chips start and stop together, hard on the gridFollows user demand through the day
BusinessA cost centre whose payoff is a better modelSells tokens directly: the "factory" metaphor fits best here

The bear case: depreciation, debt and circular deals

The build-out has serious critics, and their arguments are mostly about money rather than technology.

Bull vs bear, point by point

Question๐Ÿ‚ Bull view๐Ÿป Bear view
How long do GPUs last?Hyperscalers depreciate servers over about 5 to 6 years; older chips keep earning on inference.Michael Burry (Nov 2025) argues the real economic life is 2 to 3 years and estimates about $176 billion of understated depreciation in 2026โ€“28 (his estimate). Goldman calculates that a 3-year life would raise 2026โ€“31 depreciation from about $3 trillion to nearly $4 trillion.
Does accounting matter?Depreciation is non-cash; free cash flow is the same whatever schedule is used.Reported profits, and the valuations built on them, look better than the cash economics.
How is it financed?The biggest builders are among the most profitable companies in history.Morgan Stanley expects about $3 trillion of spending through 2028 with internal cash covering roughly half; analysts cite about $121 billion of hyperscaler debt in a year, special-purpose vehicles that keep debt off balance sheets, and Oracle's roughly $248 billion of lease commitments not yet on its balance sheet.
Is demand real?Inference demand and token volumes are growing fast; capacity sells out."Circular deals", in which chipmakers, clouds and labs invest in each other, can make demand look stronger than end-customer spending; Moody's has flagged such arrangements.
What if growth slows?Spare capacity gets absorbed as models and users grow.Analyst Gil Luria warns that if demand merely levels off, the industry could be left with overbuilt capacity and stranded debt.

The neighbours: bills, water and moratoriums

The most immediate limit on AI factories may be local politics. Opposition in 2026 has spread across many US states and across party lines, driven by electricity bills, water use and secrecy around deals.

PJM capacity price, US$ per MW-day

PJM is the largest US grid market (13 states and DC). Capacity prices are paid by all customers; data center demand is the main reason cited for the jump.

Sources: IEEFA; NPR member stations (Oct 2026).

Enterprise AI factories

The term has also become a product category for companies that want AI on their own premises. Dell AI Factory 2.0 (May 2026) is built on Nvidia Blackwell Ultra, with up to 256 GPUs per rack in Dell's IR7000; a secondary report says Dell named Eli Lilly, Honeywell, Samsung and Ascension as reference customers. HPE sells Private Cloud AI, co-developed with Nvidia, and put networking at the centre of its 2026 pitch. Cisco builds servers around Nvidia's Vera Rubin platform. These are factories in the small: a few racks rather than a gigawatt, bought for data control, compliance and predictable costs rather than scale.

What to watch to 2030

โšก Power

Grid interconnection queues, gas-turbine delivery dates, the first small modular reactors at data centers (mostly 2030s), and whether 800 V DC becomes standard.

๐Ÿ’ฐ Money

Whether 2027 capex guidance keeps rising, depreciation-schedule changes in annual reports, credit spreads on data center debt, and token prices per million.

๐Ÿ˜๏ธ Politics

FERC and state rules on who pays for grid upgrades, local moratoriums, the EU gigafactory awards in early 2027, and export controls on chips.

Reading the numbers: where sources disagree

ClaimWhat the sources say
"50ร— more throughput per megawatt"Nvidia's figure for GB300 NVL72 vs Hopper, citing SemiAnalysis; workload-dependent and vendor-selected.
2026 hyperscaler capexAbout $700โ€“725B in most roundups; one June report says about $520B because it uses Microsoft's fiscal year. Microsoft's calendar-year figure ranges from $110โ€“120B to about $190B across sources.
Cost per gigawattAbout $38B per GW from one analysis; we found no authoritative figure. "$50B per GW" circulates but was not supported.
Stargate Abilene "1.2 GW"A target. Oracle reported 618 MW handed over in September; Epoch had expected completion in Q4 2026.
Meta Hyperion "5 GW"Meta's stated expansion goal; Epoch projects 1,632 MW by 2028.
Reliance Jamnagar size1 GW in recent reports; 3 GW and $20โ€“30B in earlier ones.
Kyber 600 kW per rackA 2027 target described consistently in trade press; "1 MW later" is a roadmap aim.
Inference = two-thirds of computeOne Deloitte estimate repeated by many sites; lifetime-spend figures of 80โ€“90% measure something different.
Burry's $176BHis own estimate of understated depreciation, not an accounting finding.
IEA 2026 update figuresSeen only in press coverage, not the IEA report itself.

๐Ÿ”— References

๐Ÿงญ Closing

๐ŸŽฏ The bottom line "AI factory" is part marketing and part accurate description. The buildings really have become power plants with computers attached: racks drawing hundreds of kilowatts, liquid everywhere, campuses approaching a gigawatt, and electricity as the binding constraint. Whether the factories pay off depends on two numbers the industry does not control alone, how fast chips lose value and how much people will pay per token, and on a third that is now political: who pays for the power. Watch payback periods, power deals and local votes as closely as chip launches.