A directional model of how the four leading AI engines answer the question every buyer, investor, and policymaker is now asking — and which companies own the answer.
Trillions in announced AI capex. Tens of billions in nuclear deals. Hundreds of billions in cloud contracts. And then the question every buyer, every investor, every policymaker eventually asks — Who's actually powering AI?
This study runs that question against the four engines where the answer now forms — ChatGPT, Claude, Perplexity, and Google AI Overviews — and reports who they cite, how prominently, and from which sources. The data is a snapshot of the second quarter of 2026. The patterns are durable.
More than a third of consumers and a growing share of B2B buyers begin research with AI engines, not Google. Procurement teams at hyperscalers and at hyperscaler customers are running these exact prompts. Wall Street analysts are running them before earnings. DOE staff are running them when shaping policy. The companies that show up in those answers are not always the companies spending the most or building the fastest. They are the companies whose narrative has structurally lodged itself in the source stack the engines retrieve from.
Citation share is the new shelf space. This study is the floor map.
Want a custom Citation Share audit for your brand? 5W produces tailored AI Visibility Index audits for individual companies across any category. Same methodology, different report.
Request a Visibility Audit →This study estimates citation share — how often, and how prominently, each of 25 named entities appears in answers generated by leading AI engines when users ask category-defining questions about powering AI. Citation share is the new shelf space. In a world where more than a third of consumers and a growing share of B2B buyers begin research with AI, the answer is the market.
This is not a survey of users. This is not a count of mentions on social media. This is not a ranking of AI capability or compute capacity. It is a structured estimate of which entities AI engines surface, name, and source when users ask category-defining questions.
ChatGPT (OpenAI), Claude (Anthropic), Perplexity, and Google AI Overviews. These four account for the dominant share of consumer and enterprise AI-assisted research traffic in the United States as of publication.
Sixty-four prompts across six subcategories — Big Five Hyperscalers, AI-Native Compute, AI Consortia & Sovereign Compute, Foreign Hyperscalers, Data Center Operators, and Power-Anchored Compute — plus a cross-category authority set. Prompts were written to mirror real buyer, investor, and policy-research intent. The full prompt set is published in the Appendix.
Citation share is estimated through a combination of underlying model knowledge and structured web-search cross-checking. Four independent estimation passes were conducted per prompt to surface variance and firm directional ranges. Each entity is scored on (1) frequency of appearance across the prompt set, (2) prominence of placement inside answers, and (3) the source stack each engine cites when the entity appears. All financial figures, deal terms, and capacity numbers were verified through primary sources and current web search.
AI engine outputs vary by user, session, geography, account history, and model version. This report does not present logged query runs or single-pull rankings. It presents a directional model of the citation landscape — sufficient for category strategy, insufficient for trading decisions.
Outputs shift week to week as models retrain and source graphs update. A retest in six months will produce different absolute numbers; the structural patterns are more durable than the precise rankings. Companies that begin investing in citation share now will move on the leaderboard. That is the point.
The question is asked daily. By Wall Street analysts before earnings. By policy staff inside the Department of Energy and the White House. By procurement teams sizing cloud spend. By journalists framing the next nuclear story. By the citizen reading about a new data center in their county.
Each of them now asks the question to an AI engine first. The answer the engine produces becomes their working model. That working model becomes the next public meeting, the next investment memo, the next policy draft.
Powering AI splits into three questions that are converging into one market:
This study answers the first. The generators show up in the source stack of every subcategory because they are the proof the hyperscalers point to. The hyperscaler is the buyer. The generator is the receipt.
What follows is the leaderboard.
Citation Share Index. 0–100 scale. Aggregate across 64 prompts × 4 engines × 4 estimation passes. Directional.
SCORES ARE DIRECTIONAL ESTIMATES. NOT LOGGED QUERY RUNS. SEE METHODOLOGY § II.
The Big Five — Microsoft, Google, Amazon, Meta, and Oracle — capture more than 70% of total citation share across the prompt set. That tracks their share of announced AI capex. What does not track is the internal ranking. Microsoft sits well above the others because of a single asset: the Three Mile Island restart, rebranded Crane Clean Energy Center, anchored by a 20-year power purchase agreement with Constellation Energy for 835 megawatts. The deal generates more inbound search and editorial coverage than any other AI-power story in the cycle. Oracle, the lowest-ranked Big Five name, is now overtaken on the overall leaderboard by xAI — the first non-Big-Five entity to break into the top five.
Microsoft owns the AI-power story. The Three Mile Island restart is the most-cited single AI-power deal across all four engines, and the OpenAI association compounds the surface area. Every prompt about ChatGPT loops Microsoft. Every prompt about nuclear and AI loops Crane. The 2027 grid-return timeline keeps the story alive for at least eighteen more months of editorial cycles.
Google's citation share is structurally elevated by the host-engine effect inside Google AI Overviews — its own content surfaces more often in its own engine. Strip that bias and Google still ranks high because the Kairos Power partnership is the only first-corporate-SMR-PPA on the board, and Hermes 2 in Oak Ridge is a real, named, breaking-ground project. Google has three additional 600 MW nuclear projects publicly attached to its data center buildout. The narrative is technical-credible and source-rich.
Amazon's Susquehanna restructure with Talen Energy is a $18 billion, 17-year, 1,920-megawatt PPA — the largest single hyperscaler-nuclear commitment by megawatt-hour by a wide margin. By spend and capacity, Amazon should rank above Microsoft. It does not. The story arc lost altitude when the original behind-the-meter arrangement got reconfigured front-of-the-meter in spring 2026. The X-Energy investment and the Energy Northwest project keep Amazon in the SMR conversation, but the narrative is fragmented across three structures. Spend without story arc loses to a single famous reactor.
The January 9, 2026 triple-deal — Oklo (1.2 GW Pike County), TerraPower (up to 2.8 GW), Vistra (immediate capacity) — totaled 6.6 gigawatts. The largest corporate nuclear commitment in American history. Recency works in Meta's favor on press density; the Prometheus campus in Ohio anchors the narrative geographically. Meta still trails because the deal slate has three counterparties to remember, none of them a famous reactor. Complexity costs citation share.
Oracle's citation share is almost entirely Stargate-borrowed. The 4.5 gigawatt OpenAI partnership and the Abilene flagship build keep Oracle inside every Stargate prompt and most OpenAI prompts. Strip Stargate and Oracle's standalone AI-power presence collapses to a tail signal. Larry Ellison's claim that Oracle will build a 1 GW campus backed by three SMRs has not produced source-stack-grade reporting yet. Borrowed citation share is real citation share — until the partner moves on.
This subcategory is where citation share concentrates fastest. The AI engines learn to recognize a small set of names — CoreWeave dominant, Crusoe second, the rest measurable but not memorable. The challenger tier is a winner-take-most market inside the answer engines.
CoreWeave is the AI-native default citation. Public stock (CRWV), named in every frontier-lab procurement portfolio, anchor in OpenAI's compute roster alongside Microsoft, Oracle, and AWS. The Anthropic compute portfolio disclosure — CoreWeave for production Claude workloads — was a structural source-stack event. Among AI-native clouds, CoreWeave is now the answer the engines reach for first.
Crusoe owns the stranded-power narrative and a meaningful slice of the Stargate Abilene build. The Q4 2025 Series E at ~$10 billion valuation and the topping-out of the final Abilene building in Q1 2026 keep Crusoe inside Stargate prompts as a named contractor — a stronger position than its size would otherwise earn.
Lambda Labs is the persistent second-name in "GPU cloud" prompts. Citation share is real but tail. The recent infrastructure expansion and the Microsoft partnership for inference-as-a-service add weight that the engines have not fully ingested yet. Lambda is the cleanest GEO opportunity in this subcategory.
Together AI shows up in "open-source AI infrastructure" prompts more than "AI cloud" prompts. The model-hosting business is well-known inside developer circles and barely-known inside the citation graph the engines retrieve from. The opportunity is to push source placement out of developer pubs and into the trade press the engines weight more heavily.
Nebius (the rebuilt Yandex spinout) is closer to invisible than any company at its actual capacity should be. The Finland and Israel data centers and the public listing on Nasdaq have not produced citation-stack durability. The brand history is a friction. The growth surface is enormous if a coherent source-placement program is run.
This subcategory did not exist three years ago. It now produces more cross-engine citation density than any other subcategory in the study except the Big Five. The reason is news velocity: Stargate, xAI, G42, and HUMAIN each carry a near-weekly editorial drumbeat that the engines treat as fresh authority.
Colossus 1 built in 122 days, 200 MW initial, expanded to ~250 MW. Colossus 2 under construction targeting first gigawatt-scale single-site AI data center. The third Memphis building pushes capacity toward 2 gigawatts total with 555,000 NVIDIA GPUs at approximately $18 billion. The "speed as a weapon" narrative is the single most-cited operational claim in the AI infrastructure conversation. Memphis runs on 35 gas turbines and Tesla Megapack batteries — off-grid, behind-the-meter, controversial, and structurally story-shaped. The local opposition coverage adds to citation density, not against it. xAI is the rare entity where reputational friction increases citation share.
The OpenAI / Oracle / SoftBank joint venture announced January 2025 stalled at the formal-JV level, then reconstituted as bilateral deals — the Oracle $300 billion compute commitment, the SoftBank/SB Energy 1.2 GW Texas site, the five-site September 2025 expansion. The Abilene Texas flagship is now operational with 1.2 GW capacity running Oracle Cloud Infrastructure. Stargate's citation share holds up even with the operational complications because the engines treat the name as the umbrella for an entire infrastructure category. A category name is more durable in citation share than the company that owns it.
Stargate UAE: 5 GW AI campus in Abu Dhabi, 200 MW first phase tracking 2026 delivery, OpenAI / Oracle / Nvidia consortium partners. Microsoft's prior $1.5 billion equity investment cemented G42 as the dominant Western-aligned sovereign-AI operator outside the United States. Khazna Data Centers handles execution. The narrative is well-structured. The citation share gap relative to actual capacity is the gap between Gulf-region trade press and US-engine source-stack weighting.
Saudi Arabia's PIF-backed sovereign-AI champion. Target of 1.9 GW capacity by 2030. The AMD/Cisco joint venture announced November 2025 commits up to 1 GW of AI infrastructure starting with a 100 MW phase 1 deployment in 2026. Anchor partnerships with NVIDIA, AMD, and Google Cloud. The Saudi narrative is newer in Western citation graphs than the UAE narrative; HUMAIN's citation share is structurally below G42's despite comparable scale. The newer sovereign-AI champion has the bigger citation-share opportunity.
This is the most asymmetric subcategory in the study. By compute capacity, by capex, and by domestic market share, Alibaba, Tencent, Huawei, and Baidu would rank among the top ten entities in any honest accounting of who's powering AI globally. Inside the four Western AI engines tested, they collectively account for less than 8% of total citation share.
The cause is mechanical, not editorial. The source stacks the engines retrieve from — WSJ, Bloomberg, Reuters, the New York Times, the trade press, SEC filings, Wikipedia — are Western-language and Western-author dominant. Chinese hyperscalers produce abundant infrastructure reporting; almost none of it lands in the source stack the Western engines retrieve from at scale.
The strongest Chinese cloud signal in the study. Qwen model family gives Alibaba a citation anchor inside open-source AI prompts. The infrastructure side trails the model side — a structural inversion of the Big Five pattern.
Huawei's Ascend chip line and MindSpore framework drive citation share in "non-NVIDIA AI infrastructure" prompts. US export-control framing is the single largest editorial frame the engines retrieve.
The weakest of the four on Western engines despite domestic dominance. Hunyuan and the gaming-adjacent compute build do not translate into Western citation share.
ERNIE model family carries the citation anchor. Baidu's AI infrastructure pace lags Alibaba and Huawei in Western coverage, even though domestic ERNIE adoption is competitive.
Chinese hyperscaler citation share inside Western engines is not under-investment in PR. It is a source-stack-language asymmetry. The engines retrieve from where they read. Until the trade press and the analyst class that the engines weight begin originating primary reporting on Chinese AI infrastructure at the same density as Western infrastructure, this gap is permanent.
Equinix, Digital Realty, NTT, and QTS run a large share of the physical data center footprint the hyperscalers occupy. They are nearly invisible in the AI-power citation graph. When the engines answer "who's powering AI," they jump straight to the hyperscaler-buyer or the nuclear-supplier and skip the colocation layer entirely.
Equinix is the highest-ranked operator because of one deal: the 500 MW PPA with Oklo for SMR-sourced power across Equinix's data center footprint. The deal moved Equinix from background infrastructure into the foreground of the SMR-data-center story. One deal can promote an entity an entire tier on the citation graph.
Digital Realty has more capacity than Equinix. It has less narrative. The hyperscaler-tenant model keeps Digital Realty named-but-not-foregrounded in coverage. Capacity without a deal narrative produces capacity without citation share.
Strong in Japan and Asia-Pacific prompts. Lighter Western-engine salience.
QTS operates at the scale of a top-three colocation operator. The Blackstone ownership keeps the brand layer thin in public-facing citation. The strongest names in QTS coverage are Blackstone, Microsoft (tenant), and the construction trades — not QTS itself.
This subcategory captures the operational hybrids — companies that anchor their AI compute strategy directly to power generation, not to grid connection. xAI is the marquee. Crusoe (counted in AI-Native) is the second name. Tesla shows up adjacent. The category is small in entity count and large in citation share-per-entity because the off-grid narrative is one of the most-cited frames in the AI-power story.
Tesla's citation share is almost entirely borrowed from xAI Colossus. The Megapack deployment at Memphis and the GPU-shipment controversy keep Tesla inside the xAI prompt cluster. As an AI-power entity in its own right, Tesla rarely surfaces. Tesla is a power asset attached to an AI story, not an AI-power entity.
Iron Mountain's secure-data positioning translates poorly into "AI-power" prompts. The brand parent association (records storage) suppresses citation share in this category.
Strong actual capacity at the lowest citation share in the study. Project Jupiter in Doña Ana County, New Mexico — a planned $165 billion data center campus partnership with BorderPlex Digital Assets — was named in the September 2025 Stargate site expansion. The engines have not consolidated the story around Stack. Stack is the largest citation-share opportunity in the bottom tier.
The aggregate leaderboard hides meaningful per-engine variance. Citation share is not a single market — it is four overlapping markets with different source-stack preferences, retrieval logics, and content biases.
Heavier on the Big Five hyperscalers and on Stargate. The OpenAI parent-company effect surfaces Stargate and Oracle in more answers than other engines. Wikipedia and Reuters carry disproportionate weight. xAI underrepresented relative to news density.
Most even distribution across the 25 entities. Stronger on smaller AI-native names (Together AI, Lambda). Lighter on Chinese hyperscalers. Strongest weighting of the Wikipedia / reference layer for biographical and corporate-history anchoring.
Heaviest on news-cycle entities. Whatever broke in the last 14 days dominates. Crusoe, xAI, and recent IPO entities (Deep Fission, X-Energy) surface aggressively. Best coverage of trade press (Data Center Dynamics, The Information, TechCrunch).
Google-content elevated in Google answers. Constellation Energy and SEC filing-driven answers come back strongest here. The SMR generators (Kairos, Oklo, TerraPower) get more prominence than in other engines. Press-release-grade content lands stronger here than analysis.
An entity that wants to grow citation share cannot run one campaign. It needs four citation-share strategies, one per engine, because the source stack each engine retrieves from is structurally different. Press releases work in Google AI Overviews and underperform in Perplexity. Trade press works in Perplexity and underperforms in ChatGPT. Wikipedia anchoring works in Claude and ChatGPT and underperforms in Google AI Overviews.
The source stack is more durable than any single ranking, because moving citation share requires changing what shows up in the source stack. The matrix below maps source types to citation weight, by engine, for the AI-power category.
| Source Type | Examples | Weight | Engine Bias |
|---|---|---|---|
| Primary Reporting | WSJ, Bloomberg, Reuters, Financial Times | Highest | All four; heaviest in Perplexity and Google AI Overviews |
| Trade Press | Data Center Dynamics, The Information, TechCrunch, Axios Pro, Power Magazine | High | Heavy in Perplexity; mid in ChatGPT and Claude |
| Vendor-Owned | Press releases, S-1 filings, investor presentations, blog posts | Mid | Heavy in Google AI Overviews; lighter in Claude |
| Reference Layer | Wikipedia, Crunchbase, PitchBook open profiles | Mid-High | Heavy in ChatGPT and Claude; thinner in Perplexity |
| Analyst & Research | Gartner, IDC, SemiAnalysis, Dell'Oro, Synergy Research, HyperFRAME | Mid | Heavy in B2B prompts; thin in consumer-shaped prompts |
| Earnings & Filings | 10-Ks, 8-Ks, earnings calls, investor day decks, S-1s | High | Heavy in all engines for capex and capacity questions |
| Long-Form Editorial | NYT, WaPo, The Atlantic, The Economist, New Yorker | High | Heavy in Google AI Overviews and Claude |
| Independent Trade Voices | Stratechery, Hard Fork, Ben Thompson, Casey Newton, Everything-PR | Mid | Growing in Perplexity and Claude; thin in ChatGPT |
| Government & Regulatory | DOE press, NRC docket filings, FERC, state PUCs | High on policy | Heavy in policy prompts; thin in product prompts |
When the source stack is the same across two subcategories, citation share is mostly a function of paid-PR muscle and press-cycle saturation. When the source stack diverges, citation share is a function of strategic source placement — which is the GEO playbook.
In this study, the Big Five Hyperscalers and the AI Consortia & Sovereign Compute subcategories share substantially the same source stack — both rely heavily on Primary Reporting and Vendor-Owned content. The AI-Native Compute subcategory diverges sharply, leaning on Analyst & Research and Independent Trade Voices. The Foreign Hyperscalers subcategory has an entirely different source stack — Bloomberg Asia, Reuters Asia, Nikkei, South China Morning Post — that the Western engines under-retrieve.
The under-represented entities are not failures of capacity. They are failures of source placement. Each name below operates AI-power infrastructure at a scale that should produce stronger citation share than it does.
Microsoft's citation lead is sticky for as long as the Three Mile Island story keeps cycling. The next major reactor restart or first-SMR-online event will move share. Google's host-engine bias inside Google AI Overviews is a real structural advantage and a measurable one — strip it and Google's overall lead narrows to a competitive margin. Amazon's underweight relative to spend is the most resolvable gap among the Big Five — the Susquehanna front-of-the-meter transition in spring 2026 is a publishable narrative event that has not been fully harvested.
Citation share inside AI engines is not a brand-awareness metric. It is a buyer-research surface. The companies named here are the companies whose customers, investors, regulators, and journalists now form working models of the AI-power market from AI-engine answers. Five operational implications follow.
The companies that rank highest are the companies whose deals show up in Reuters, WSJ, Bloomberg, DCD, TechCrunch, and the SEC filings in primary citations. Press releases hit Google AI Overviews and underperform everywhere else. Trade-press relationships and analyst placement matter more than the next funding announcement.
Microsoft's Three Mile Island position. xAI's Colossus 122-day build. Oklo's Pike County campus. The engines reward concentrated narrative gravity. Companies with twelve mid-sized deals trail companies with one famous one.
ChatGPT, Claude, Perplexity, and Google AI Overviews retrieve from different source stacks. A press release works in one engine and not in another. A Wikipedia update works in two and not in two others. Analyst placement works in some prompts and not others. The four-engine source-placement matrix is the playbook.
Data Center Operators (Equinix excepted), Foreign Hyperscalers in Western engines, and AI-Native Compute below the top-two all show structural under-representation relative to operational scale. These are the fastest-moving citation-share opportunities in the study.
Every entity in this study carries reputation risk that is one news cycle from the front page — a reactor incident, an outage, a permitting fight, a labor action, a security event. The citation share you have on a calm day is the citation share that retrieves on a crisis day. Build the infrastructure before the crisis, not during it.
This is a directional model. It is not a logged-query benchmark and never represents itself as one. The following limitations are explicit.
Who powers ChatGPT · Microsoft AI power strategy · How does Google power Gemini · AWS AI infrastructure · Meta AI data centers · Oracle AI cloud · Microsoft nuclear deals · Amazon Talen Energy deal · Google Kairos nuclear partnership · Meta nuclear RFP · Best AI cloud provider for enterprise · Hyperscaler power consumption ranked
CoreWeave vs AWS · Crusoe AI data centers · Best GPU cloud for AI training · Lambda Labs power consumption · Nebius AI infrastructure · AI-native cloud providers list · Where do AI startups get compute · Specialized AI compute providers · Together AI infrastructure · Best alternative to hyperscalers for AI
What is Stargate · OpenAI infrastructure partners · xAI Colossus cluster · Largest AI training cluster in the world · G42 AI infrastructure · UAE AI strategy · Saudi Arabia AI compute · HUMAIN AI · Who is building gigawatt AI data centers · AI compute build-out 2026
Alibaba AI cloud · Chinese hyperscaler AI · Tencent AI infrastructure · Huawei AI compute · Baidu AI Cloud · Asia AI hyperscaler · China AI data center capacity · How does China power its AI
Largest data center operator · Equinix AI customers · Digital Realty AI tenants · Best colocation for AI · NTT data centers AI · Who builds data centers for hyperscalers · Iron Mountain AI data center · QTS Blackstone AI
Tesla data center Memphis · xAI Memphis power · Behind-the-meter AI power · Off-grid AI compute · Most power-efficient hyperscaler · Renewable energy AI data centers · Stranded power AI compute · Gas turbine AI data center
Who is winning the AI infrastructure race · Most-cited AI hyperscaler · AI data center capacity by company · AI compute market share · Hyperscaler AI capex ranking · Where will AI compute come from · AI power crisis solutions · Next decade of AI infrastructure
Citation Share. Estimated frequency and prominence of named-entity appearance inside AI engine answers, normalized across the prompt set. Directional.
GEO. Generative Engine Optimization. The discipline of building brand authority and source-stack presence such that AI engines surface, name, and cite a brand inside category-defining answers.
Source Stack. The ordered set of publications, filings, and reference layers that AI engines retrieve from when answering a category-defining prompt.
Retrieval Anchor. A piece of content (article, deal release, filing, Wikipedia entry, analyst report) that AI engines reliably surface when a related entity is queried. The unit of GEO production.
SMR. Small Modular Reactor. Advanced nuclear reactor designs typically generating 50–300 MW per unit, intended for data center and industrial power.
PPA. Power Purchase Agreement. Long-term contract between an electricity generator and a buyer, typically 10–20 years.
BTM / FTM. Behind-the-Meter / Front-of-the-Meter. BTM bypasses the public grid; FTM delivers through the grid. Talen-Amazon converted BTM to FTM in spring 2026.