The 100 Most-Cited Websites in ChatGPT for High-Stakes Investing, Retirement, Mortgage and Financial Consumer Decisions

A three-month citation-authority study across 11 commercial financial vertical datasets, using response-level domain coverage as the primary ranking metric.

Answer Capsule

Across 1,373 deduplicated ChatGPT responses from July through September 2026, LLM Authority Index observed 6,700 visible citation events spanning 975 registrable domains in the Investing, Retirement, Mortgage and Financial Decisions research family. forbes.com ranked #1 at 30.8% response citation coverage, followed by nerdwallet.com at 18.6%, reddit.com at 17.8%, wsj.com at 13.1%, and bankrate.com at 11.8%. The category Top 100 captured 69.7% of all response-level domain appearances, substantially above ChatGPT's broader high-stakes Top 100 concentration of 58.8%. Only 50 domains remained in the monthly Top 100 in all three months. This is a platform-specific source map, not a universal ranking of financial authority.

Key Findings

• forbes.com ranked #1 with 30.8% response citation coverage, appearing in 423 of 1,373 eligible ChatGPT responses.
• The Top 5 were forbes.com, nerdwallet.com, reddit.com, wsj.com and bankrate.com.
• The Top 10 captured 31.6% of response-level domain appearances, the Top 25 captured 46.4%, and the Top 100 captured 69.7%.
• ChatGPT's financial Top 100 shared 54 domains with Google AI Overviews, 52 with Google AI Mode, 41 with Perplexity, 41 with Gemini and 40 with Microsoft Copilot. The two Google financial lists share 83 domains with each other.
• Only 50 domains overlap between this category Top 100 and ChatGPT's broader high-stakes Top 100. Category context changes the source leaderboard materially.
• 50 domains stayed in the category Top 100 in July, August and September.
• The monthly Top 100 overlap was 38.9% Jaccard for July versus August, 38.9% for July versus September, and 55.0% for August versus September.
• Source leadership varies sharply by vertical. ChatGPT's leading source was kraken.com for Crypto Exchanges, coinbureau.com for Crypto Wallets, reddit.com for Gold IRAs and Precious Metals Dealers, stockbrokers.com for Online Stock Brokers, money.com for Reverse Mortgage, and consumeraffairs.com for Structured Settlements.
• The financial category is more concentrated than ChatGPT's broader high-stakes source market at deeper ranking levels. The Top 100 capture 69.7% of category domain-response appearances versus 58.8% across ChatGPT overall.

Questions This Study Answers

• Which websites dominate ChatGPT citations for high-stakes financial decisions, and how does that source hierarchy change by vertical?
• How persistent is the ChatGPT financial source market over time, and how portable is that authority to other AI platforms?
• What should publishers, financial brands and CMOs measure before allocating resources to ChatGPT-focused content, PR, partnerships or earned media?

Why This Research Belongs in Daily AI Search Marketing Decisions

This study is research-based, but its purpose is commercial decision support.

This Article sits at the intersection of two commercially useful views of the same source market:

• the ChatGPT platform view, which asks which sources recur across high-stakes consumer decisions on one AI platform; and
• the Investing, Retirement, Mortgage and Financial Decisions category view, which asks which sources recur across six AI platform families for one financial decision family.

For publishers, that intersection creates a more precise claim than saying a domain is simply "visible in ChatGPT." A publisher can document whether its visibility is strong specifically inside ChatGPT for high-stakes financial decisions, whether that visibility persists over time, and whether its authority is broad across several financial verticals or concentrated in one specialist category.

For financial brands, CMOs, PR teams and agencies, the operating question is:

Which sources are already present around the exact financial decisions our customers ask ChatGPT to help them make?

Four findings matter for daily planning:

• the Top 100 account for 69.7% of response-level domain appearances in this ChatGPT financial slice, compared with 58.8% across ChatGPT's broader high-stakes source market;
• 50 domains remained in the monthly Top 100 in July, August and September, and the median of the three monthly Top 100 Jaccard comparisons was only 38.9%;
• the final financial child-study Top 100 lists show ChatGPT sharing 54 domains with Google AI Overviews, 52 with Google AI Mode, 41 with Perplexity, 41 with Gemini and 40 with Microsoft Copilot; and
• the two Google financial surfaces share 83 domains with each other, showing that ChatGPT's financial source market is substantially less portable to those surfaces than the Google surfaces are to one another.

The parent Investing, Retirement, Mortgage and Financial Decisions study provides the category-level Persistence-Portability Gap interpretation. Under its frozen summary-based methodology, it reports 66.7% temporal persistence, 28.2% cross-platform portability, and an approximate 38.5 percentage-point Persistence-Portability Gap for this decision family.

That parent gap is exploratory and should remain the category-level summary. Article 23 should not manufacture a new gap from one or two platform comparisons. Its own 38.9% median monthly persistence answers a narrower ChatGPT-over-time question, while its pairwise overlaps with other engines answer separate cross-platform questions.

The parent and child studies also use slightly different frozen analysis versions. The parent category table reports 1,348 ChatGPT responses, while this later child study reports 1,373. Those versions should not be silently merged. The parent study remains the authoritative category-level persistence-portability summary until the entire financial family is rerun through one consolidated pipeline.

Current outside research reinforces why methodology and prompt population have to travel with every commercial claim.

Ahrefs' September 2026 broad U.S. ChatGPT study ranks Reddit first at 16.8% mention share, Wikipedia second at 7.0%, Consumer Reports third at 3.7%, and Forbes fourth at 3.1%. In this high-stakes financial panel, Forbes ranks #1 at 30.8% response citation coverage. The two studies use different metrics and prompt populations, so the percentages should not be merged. Ahrefs ChatGPT citation study

Wellows analyzed 22.7 million citations across more than 1.1 million questions and found that 76.3% of the sources ChatGPT cited were not cited by any of the other four measured engines for the same question. In its pairwise comparisons, the other engines failed to use roughly nine in ten of the websites ChatGPT cited on the same question. Wellows AI Citation Overlap Study

Machine Relations' September 2026 six-engine open-web index found that the broad Top 10 accounted for only 7.41% of domain-run citations and that 1,357 domains were needed to reach half of observed citation volume. That broad denominator is fundamentally different from this narrow ChatGPT financial decision set, where the Top 100 account for 69.7% of domain-response appearances. Machine Relations citation concentration study

The useful agreement across those outside studies is structural: ChatGPT source behavior depends heavily on prompt population, source role, engine, metric and denominator.

Commercial Action Matrix: What Publishers and CMOs Can Do With the Findings

The daily operating model is therefore four-layered:

1. Current ChatGPT financial core: maintain the sources that matter now, with the collection date attached.
2. Vertical and prompt-cluster layer: identify which sources dominate the exact mortgage, investing, retirement, crypto or other financial decision.
3. Longitudinal check: monitor whether priority sources persist, decline, rotate or change rank.
4. Cross-platform check: determine whether the same sources also matter in Google AI Overviews, Google AI Mode, Perplexity, Gemini and Microsoft Copilot.

Citation visibility is evidence of source presence. It is not proof that obtaining coverage, advertising, a partnership or a placement on a cited domain will cause ChatGPT to recommend a financial company or product.

How This Study Fits the High-Stakes Consumer Decisions Research Corpus

This Article is the intersection of two parent studies:

• The ChatGPT Platform Study
• The Investing, Retirement, Mortgage and Financial Decisions Citation Authority Study

It should also be read with:

• the 2026 AI Citation Authority Study, which establishes the full high-stakes source ecosystem;
• the Persistence-Portability Gap, which separates temporal durability from cross-platform transferability;
• Do AI Platforms Cite the Same Websites?, which provides the overall platform-overlap framework;
• How Stable Are AI Citations?, which separates portfolio persistence from same-prompt stability;
• Reddit's Decline in AI Citations for Investing, Retirement, Mortgage and Financial Decisions; and
• How Reddit Citations Changed in ChatGPT for Financial Decisions, which provides one longitudinal source example inside this exact platform-category intersection.

What We Studied

This analysis uses the same frozen July through September 2026 parent corpus and anti-double-counting rules used throughout the LLM Authority Index 66-study series. It restricts that analytical corpus to the ChatGPT platform family and the Investing, Retirement, Mortgage and Financial Decisions family.

Study size

Monthly response composition

Verticals Covered

The source labels included in this family are:

1. Annuities
2. Crypto Exchanges
3. Crypto Wallets
4. Gold IRAs and Precious Metals Dealers
5. Mortgage
6. Mortgage Refinance Lenders
7. Online Financial Advisors
8. Online Stock Brokers
9. Reverse Mortgage
10. Robo-Advisors
11. Structured Settlements

These are source dataset labels, not claims that the eleven labels are statistically independent industries. Prompt overlap can occur, and the verticals have different sample sizes.

The 25 Most-Cited Websites in ChatGPT for Investing, Retirement, Mortgage and Financial Decisions

Response Citation Coverage is the percentage of eligible deduplicated responses in this platform/category slice that visibly cited a domain at least once. A domain can count only once per response for the primary metric.

The Top 25 combine broad financial publishers, general media, community sources, financial brands and specialist vertical authorities. The category-wide order should not be treated as a universal outreach priority list for every financial product.

The Full Top 100 ChatGPT Financial Citation Sources

The Top 100 table is domain-level. It does not imply that every URL on a domain performs equally well, and it does not measure sentiment, factual quality or recommendation strength.

ChatGPT's Financial Source Market Is More Concentrated Than ChatGPT Overall

The Top 10 concentration is almost identical to ChatGPT's broad high-stakes panel, but the category remains much more concentrated as the ranking extends into the Top 25, Top 50 and Top 100. That pattern suggests a familiar head of broad sources plus a more specialized financial middle tier.

For a CMO, this is commercially useful because it argues against building a ChatGPT citation strategy from a generic platform-wide list. Financial decisions pull a larger share of source visibility into recurring finance-specific publishers, financial institutions and specialist domains.

Which Sources Gain the Most Inside Financial Decisions?

The following table compares a domain's response coverage in this financial category with its response coverage across the broader ChatGPT high-stakes corpus. A lift above 1.0x means the source appears more often in these financial decisions than it does across ChatGPT overall.

Specialist lifts should be read carefully. A high lift does not mean ChatGPT "prefers" a domain. It means the domain appears much more often in this category than in the broader captured ChatGPT source market.

The Most-Cited Sources by Individual Financial Vertical

The vertical cuts materially change the practical source map:

• Annuities: Forbes leads at 47.8%, followed by Kiplinger and AnnuityJournal.org.
• Crypto Exchanges: Kraken leads at 32.9%, followed by Forbes and Coinbase.
• Crypto Wallets: CoinBureau leads at 27.9%, followed by Reddit. This is a different hierarchy from the exchange market.
• Gold IRAs and Precious Metals Dealers: Reddit leads at 26.7%, followed by Money.com and Costco.
• Mortgage: NerdWallet leads at 43.1%, followed closely by Forbes at 41.3% and Bankrate at 35.3%.
• Mortgage Refinance Lenders: Forbes leads at 49.3%, followed by NerdWallet and Bankrate.
• Online Financial Advisors: Forbes leads at 47.7%, while Fidelity, Vanguard and Schwab all have substantial visibility.
• Online Stock Brokers: StockBrokers.com leads at 35.0%, ahead of Forbes and NerdWallet.
• Reverse Mortgage: Money.com leads at 33.3%, narrowly ahead of Forbes.
• Robo-Advisors: Forbes leads at 51.2%, followed by Reddit and Fidelity.
• Structured Settlements: ConsumerAffairs leads at 40.4%, followed by Catalina Structured Funding.

This is why LLM Authority Index includes vertical tables inside broader category studies instead of publishing a thin page for every source label.

Broad Publisher Authority and Specialist Authority Are Different Assets

A broad publisher can win by appearing across many verticals. A specialist can win by being unusually strong in one buyer journey.

In this ChatGPT slice, reddit.com and investopedia.com appear across all 11 vertical labels, while Forbes, NerdWallet and The Wall Street Journal appear across ten. That breadth can matter for publishers selling cross-category financial reach.

But specialist sources can dominate narrower commercial decisions. StockBrokers.com ranks #10 category-wide but leads Online Stock Brokers. Kraken ranks #18 category-wide but leads Crypto Exchanges. CoinBureau ranks #12 category-wide but leads Crypto Wallets. ConsumerAffairs ranks #21 category-wide but leads Structured Settlements.

For CMOs, this means a lower-ranked specialist can be more strategically relevant than a higher-ranked broad publisher when the objective is a narrow prompt cluster.

ChatGPT's Financial Source Market Is Not Interchangeable With Other AI Platforms

The final published child-study Top 100 lists make the platform difference explicit.

The two Google financial surfaces are much more similar to each other than ChatGPT is to any of the five other measured platform families.

The same contrast appears in the Top 10. ChatGPT shares seven Top 10 domains with each Google financial surface, while the two Google surfaces contain the same ten Top 10 domains, only in a different order.

That does not mean nothing transfers between engines. Forbes, NerdWallet, Bankrate, Reddit, Investopedia and other large financial sources recur across multiple platform lists. It means the priority order, source mix and specialist layer change enough that a single universal publisher list is not defensible.

The implication is operational: maintain a cross-platform core, then preserve a ChatGPT-specific source layer for the exact financial prompt clusters that matter.

This pattern is directionally consistent with Wellows' matched-question research, which found unusually low website-level source overlap between ChatGPT and other AI engines. The Wellows percentages use same-question website sets rather than Top 100 category portfolios, so the numerical values should not be compared directly.

Month-to-Month Persistence

Only 50 domains stayed in the ChatGPT financial Top 100 in all three months. The median of the three monthly Top 100 Jaccard comparisons was 38.9%.

The category therefore has a stable core, but a much less stable perimeter. A few examples illustrate the movement:

• forbes.com remained #1 in July, August and September.
• nerdwallet.com moved 3 to 2 to 2.
• reddit.com moved 2 to 6 to 9.
• coinbureau.com moved 34 to 59 to 3, a major September surge.
• wikipedia.org moved 5 to 22 to 179, leaving the September Top 100.
• consumeraffairs.com moved 169 to 12 to 19, showing how quickly a specialist source can enter the visible head of the ranking.

This should be read alongside the separate citation-drift study. A domain can remain strategically important even while query-level source sets and exact URLs change substantially.

How This Study Relates to Other ChatGPT and AI Citation Research

Ahrefs: broad all-topic ChatGPT rankings

Ahrefs' September 2026 study ranked Reddit first in ChatGPT with 16.8% mention share across a broad set of US queries. Wikipedia ranked second, while Forbes ranked fourth. This study produces a different order because it restricts the prompt universe to high-stakes commercial financial decisions and ranks domains by response citation coverage, not mention share among the leading sources.

The two studies therefore answer different questions. Ahrefs describes the broad ChatGPT source market. This study describes a narrower financial decision market.

Wellows: matched-question cross-engine divergence

Wellows analyzed 22.7 million citations across more than 1.1 million questions and reported that 76.3% of the sources ChatGPT cited were not cited by any of the other four measured engines for the same question. Its pairwise website-level comparisons also place ChatGPT among the least overlapping source pools.

That is directionally consistent with the platform-specificity observed here. The numerical values should not be compared directly because the matching rules, prompt populations and denominators differ.

Machine Relations: concentration depends on source class and denominator

Machine Relations' September 2026 index found a much longer citation tail across six engines and more than 22,000 cited domains. Its broad open-web Top 10 accounted for only 7.41% of domain-run citations. That does not conflict with the higher concentration reported here. This article is a narrow ChatGPT financial slice and uses response-level domain appearances rather than a six-engine open-web denominator.

The comparison reinforces a core methodological point: citation concentration is a property of the prompt population, engine set and denominator, not a universal constant.

How the LLM Authority Index Study Is Different

This study is not positioned as the largest AI citation dataset. Its defensible distinction is the combination of:

• a high-stakes consumer-decision focus;
• eleven commercial financial source vertical labels;
• a three-month longitudinal window;
• a ChatGPT-specific slice;
• response-level coverage and citation-event metrics;
• explicit extraction-failure accounting;
• global exact-repeat handling;
• Public Suffix List registrable-domain normalization;
• a full Top 100 plus vertical-level source leaders;
• direct comparison with the two Google AI financial source markets;
• commercial interpretation for publishers and CMOs.

Why AI Citation Studies Can Disagree

Two well-executed studies can produce different rankings without either being wrong. Major reasons include:

1. Prompt mix. Broad informational prompts, commercial comparison prompts and company-specific prompts surface different sources.
2. Platform. ChatGPT, Google AI Overviews and Google AI Mode use different source ecosystems.
3. Time window. Source rankings can move materially across weeks and months.
4. Metric. Mention share, raw citation-event share, response coverage, unique URLs and unique domains answer different questions.
5. Domain normalization. Subdomains can be consolidated differently.
6. Infrastructure filtering. Search wrappers, maps, thumbnails and technical URLs can distort rankings if treated as publishers.
7. Duplicate handling. Repeated prompt-surface observations can overweight some sources.
8. Geography and account state. Responses can vary by location, product tier, user state and rollout.
9. Citation visibility contract. Different tools may capture visible source links differently from the underlying product interface.

The correct comparison is therefore between studies with aligned definitions, not between two headline percentages stripped of methodology.

What the Results Mean for Financial Publishers

1. Prove the exact commercial footprint

A publisher can use the category and vertical tables to distinguish broad financial citation visibility from specialist depth. That is more useful to an advertiser than a generic statement that the publisher is "cited by ChatGPT."

2. Measure ChatGPT separately from Google

The Top 100 overlap results show that ChatGPT's source market is materially different from the two Google AI surfaces. Publishers should report platform-specific performance instead of pooling incompatible denominators.

3. Preserve pages that repeatedly earn citations

Domain-level authority is only the first layer. The platform still cites specific URLs. Publishers should retain and update pages that repeatedly earn citations, then measure URL persistence alongside domain coverage.

4. Sell breadth and depth honestly

A broad publisher can document cross-vertical reach. A specialist can document unusually high visibility in a narrow commercial decision. Both are useful, but they are different products.

What the Results Mean for CMOs and Financial Brands

1. Build a ChatGPT-specific source map

Do not assume that the publishers most visible in Google AI Overviews or Google AI Mode are the same sources ChatGPT uses most. Start with the ChatGPT category and vertical tables.

2. Separate direct citation from third-party earned media

A brand can appear through its own site, through independent publisher coverage, or both. Those are different levers and should be measured separately.

3. Benchmark before spending

Run the target prompt cluster first. Record company mentions, recommendations, cited domains and cited URLs. Then use the recurring source set to prioritize content, PR and publisher outreach.

4. Re-run the same prompt set

The monthly persistence data show enough movement that a one-time screenshot cannot establish campaign impact. Use matched prompts and repeated measurement.

Methodology

Parent dataset

The parent LLM Authority Index corpus contains July, August and September 2026 observations across six AI platform families and 53 source vertical labels. The broader flagship study analyzes 60,881 deduplicated responses and 278,499 visible citation events after the frozen eligibility and anti-double-counting rules.

Eligible platform and category

This article restricts the parent analytical corpus to ChatGPT and the Investing, Retirement, Mortgage and Financial Decisions family.

Explicit extraction failures

Observations explicitly marked as extraction failed are excluded from the analytical denominator. Missing source visibility is not interpreted as proof that ChatGPT used no sources.

Global exact-repeat handling

An observation is treated as an analytical repeat when all four conditions match:

1. same report month;
2. same exact raw platform/source surface;
3. same normalized prompt text;
4. same exact set of original citation URLs.

One analytical record is retained. This is an anti-double-counting rule, not a claim that the upstream system executed the prompt only once.

Prompt normalization

Prompt text is normalized with Unicode NFKC, case folding, whitespace collapse and trim for the repeat check.

Domain normalization

Citation URLs are parsed to hostname, lowercased, stripped of leading www., and consolidated to registrable domain using the Public Suffix List.

Public infrastructure exclusions

Known navigation or infrastructure artifacts are excluded from public publisher rankings, including designated map, thumbnail, image and activity endpoints. Raw URLs remain available in the research pipeline for auditability.

Primary ranking metric

Response Citation Coverage = eligible deduplicated responses citing domain / all eligible deduplicated responses in this platform/category slice.

A domain counts at most once per response. This prevents a response with several links to the same site from dominating the primary ranking.

Secondary metric

Citation Events count visible citation entries after public infrastructure filtering. This is useful for depth analysis but is not the primary ranking metric.

Vertical membership

A deduplicated response can retain source-dataset membership needed for vertical analysis. The vertical table reports response-level citation coverage within each source vertical label.

Monthly persistence

Monthly Top 100 lists use the same response-level ranking metric. Pairwise overlap uses Jaccard similarity: intersection divided by union. The persistent count is the number of domains present in all three monthly Top 100 lists.

Category lift

Category lift compares a domain's response coverage in this ChatGPT financial slice with its response coverage across the broader ChatGPT high-stakes corpus. Lift is descriptive and does not imply causal preference by the platform.

Limitations

• The study measures visible citations in the captured research pipeline, not hidden retrieval, ranking systems or model training data.
• Citation does not establish recommendation causality, endorsement, trust, factual accuracy or conversion impact.
• Prompt sets are research samples, not a census of every possible user query.
• Source vertical labels differ in sample size and may overlap semantically.
• Domain-level rankings can hide major URL-level differences.
• ChatGPT behavior can vary by geography, account state, model version and rollout.
• Visible citation interfaces and source-extraction contracts can change over time.
• A first-party versus third-party relationship depends on the evaluated brand, so this article does not apply a blanket domain taxonomy where the relationship has not been explicitly established.

Commercial Relationship Disclosure

LLM Authority Index is part of a business ecosystem that provides AI visibility, research, and marketing services. Affiliated businesses may have current or historical commercial relationships with companies or publishers that appear in the dataset.

Commercial relationships are not inputs to ranking methodology. Domains are included based on observed citation data. A ranking is not an endorsement, and a commercial relationship is not treated as a cause of citation visibility.

References

1. LLM Authority Index, 2026 AI Citation Authority Study
2. LLM Authority Index, Most-Cited Websites in AI for Investing, Retirement, Mortgage and Financial Decisions
3. LLM Authority Index, ChatGPT Top 100
4. LLM Authority Index, Google AI Overviews Financial Top 100
5. LLM Authority Index, Google AI Mode Financial Top 100
6. LLM Authority Index, Cross-Platform Citation Overlap
7. LLM Authority Index, Citation Drift Study
8. Ahrefs, The 50 Most-Cited Websites in ChatGPT, September 2026
9. Wellows, 89% of What ChatGPT Cites, Perplexity Never Touches
10. Machine Relations, How Concentrated Are AI Citations?

Related LLM Authority Index Research

Parent studies and measurement framework

• The 2026 AI Citation Authority Study
• The Investing, Retirement, Mortgage and Financial Decisions Citation Authority Study
• The ChatGPT Platform Study
• The Persistence-Portability Gap
• Do AI Platforms Cite the Same Websites?
• How Stable Are AI Citations?

Sibling financial platform studies

• Google AI Overviews
• Google AI Mode
• Perplexity
• Gemini
• Microsoft Copilot

Financial Reddit trend studies

• Financial Reddit parent study
• Google AI Overviews Reddit trend
• Google AI Mode Reddit trend
• ChatGPT Reddit trend
• Perplexity Reddit trend
• Gemini Reddit trend
• Microsoft Copilot Reddit trend