Skip to main content
USA-Based Digital Agency

Original research · Data you can download

How Well Do the 100 Largest U.S. Banks Protect Their Domains? A DNS Security Study

Most of the largest U.S. banks block spoofed email well. Far fewer lock down the domain itself. Only 30% of the 100 largest banks’ domains have a registry lock, 26% are signed with DNSSEC, and 88% depend on a single DNS operator. We checked the website domain the FDIC lists for each bank on November 17, 2026.

By George Shvaya · · 10 min read

have a registry lock on the domain
29.8%
28 of 94
rely on a single DNS operator
88.3%
83 of 94
sign their domain with DNSSEC
25.5%
24 of 94
cannot tell mail servers to block spoofed email
11.7%
11 of 94

Why we checked

On November 16, 2026, customers of Chase and several other banks lost online banking for hours. Chase said a software bug in its own systems caused it, and our own checks that evening found no sign of DNS tampering. We wrote that up in Was it a DNS attack or an outage? The questions readers asked next were broader: how well do banks protect the parts of their web presence an attacker could hijack, such as the domain, its DNS and its email? Those protections are public, so we measured them for the 100 largest banks.

Registry lock: 30% overall, 11% outside the top 50

Most domain hijacks start with a stolen registrar or DNS login, not with breaking DNS itself. Cisco Talos recommends a registry lock against exactly that, because changes then need an out-of-band check (Talos). ICANN describes the status codes a registry lock sets, including serverUpdateProhibited, as an extra protection against unauthorized changes (ICANN). We read each domain’s status from the registry’s own RDAP service.

Domain protections by bank size (FDIC rank by total assets), November 17, 2026. Counts are bank domains.
BanksDomainsRegistry lockDNSSECCAADMARC reject
Ranks 1–252415 (62.5%)3 (12.5%)10 (41.7%)23 (95.8%)
Ranks 26–50258 (32.0%)9 (36.0%)5 (20.0%)19 (76.0%)
Ranks 51–100455 (11.1%)12 (26.7%)5 (11.1%)32 (71.1%)
All9428 (29.8%)24 (25.5%)20 (21.3%)74 (78.7%)

Size matters most here. Large banks are far more likely to have the lock: 15 of the 24 top-25 banks we could check have it. The other 9 did not show it, including capitalone.com, bny.com, statestreet.com, 53.com and key.com. A registry lock is one protection among several; a bank can also lock its account at the registrar and restrict who can request changes, which public data cannot show.

88% rely on one DNS operator

Every lookup for a bank’s website starts at its authoritative name servers. 83 of 94 bank domains used a single operator for them; 10 used two and 1 used three. Three providers, Akamai, Cloudflare and UltraDNS, answered for 55 of the 94, and 22 used name servers of their own.

Who answers DNS for the 100 largest U.S. banks’ domains, November 17, 2026. A domain with two operators counts under both.
DNS operatorBank domains
Akamai24
Bank’s own name servers22
Cloudflare19
UltraDNS (Vercara)15
Amazon Route 535
AT&T5
CSC4
GoDaddy4
Others (7 operators)8

One operator is not one server: the large providers run many servers in many places. The risk is concentration. In October 2016 an attack on Dyn took down many unrelated sites at once because they shared the same DNS provider. A second operator removes that dependency, at the cost of keeping two systems in sync.

DNSSEC and CAA: a quarter and a fifth

DNSSEC lets resolvers detect forged DNS answers; ICANN says it “protects the user from getting bad data from a signed zone” (ICANN). 24 bank domains (25.5%) were signed, and all 24 validated at Cloudflare’s public resolver. The largest banks sign least: 3 of the 24 top-25 banks we could check (bankofamerica.com, firstcitizens.com and key.com) against 12 of the 45 banks ranked 51 to 100. The one bank on the .bank domain, provident.bank, was signed.

A CAA record names the certificate authorities allowed to issue certificates for a domain (RFC 8659). 20 bank domains (21.3%) published one. Taken together, 42 domains (44.7%) had none of the three: no registry lock, no DNSSEC and no CAA.

Email spoofing: most banks enforce, 11 do not

Outages bring phishing. A DMARC policy tells receiving mail servers what to do with mail that claims to come from the bank’s domain but fails authentication. A policy of p=none means the domain owner “requests no specific action be taken regarding delivery of messages” (RFC 7489). 74 bank domains (78.7%) published p=reject and 9 published p=quarantine. 10 published p=none and 1 had no DMARC record on the domain the FDIC lists:

Bank domains without an enforcing DMARC policy, November 17, 2026. The domain checked is the website the FDIC lists, which may not be every domain a bank sends mail from.
FDIC rankBankDomainDMARC
34Western Alliance Bankwesternalliancebancorporation.comNo DMARC record
39Old National Bankoldnational.comp=none
40UMB Bankumb.comp=none
47Frost Bankfrostbank.comp=none
62Banc of Californiabancofcal.comp=none
63BankUnitedbankunited.comp=none
91Bank of Hawaiiboh.comp=none
92First Hawaiian Bankfhb.comp=none
93OceanFirst Bankoceanfirst.comp=none
94TowneBanktownebank.comp=none
95WSFS Bankwsfsbank.comp=none

p=none is a normal stage while a domain owner checks that its legitimate mail passes before enforcing; it is not a sign of compromise. SPF was close to universal: 47 domains ended their SPF record with a hard fail (-all) and 41 with a soft fail (~all). Only 2 published MTA-STS (RFC 8461), which makes mail servers require an encrypted connection: fnbo.com and amerisbank.com.

HTTPS and certificates: universal

All 94 bank websites served a valid certificate on their www host. DigiCert issued 48 of them, Google Trust Services 14, Sectigo 11 and Let’s Encrypt 8. Of the 91 sites that answered our request, 74 (81.3%) sent a Strict-Transport-Security header, which tells browsers to use HTTPS only (RFC 6797).

What this means for any business with a domain

Banks are a hard target with large security teams, and they still leave some of these off. For a smaller business the same four settings are cheap and quick to add, and they cover the ways domains are actually hijacked or impersonated:

  • Registry or registrar lock on the domain, and multi-factor sign-in on the registrar and DNS host accounts.
  • DMARC at p=reject once your real mail passes SPF and DKIM, so nobody can send convincing mail as you.
  • A CAA record naming the certificate authority you use.
  • DNSSEC where your registrar and DNS host support it.

Our DNS guide shows how to check each one yourself, and the security hub covers the rest of a modern stack.

Method

  • Sample. The 100 largest active FDIC-insured institutions by total assets from the FDIC BankFind API (FDIC), report date June 30, 2026. Four list no website (Morgan Stanley Bank, Bank of China, Charles Schwab Premier Bank and State Bank of India) and were left out rather than guessed. Two domains are each listed for two banks (td.com, bny.com) and were counted once: 94 domains.
  • Domain. The registered domain of the website the FDIC lists, so JPMorgan Chase Bank is jpmorganchase.com, not chase.com. Our outage guide covers chase.com and six other consumer domains separately.
  • DNS. NS, DS, CAA, TXT, _dmarc and _mta-sts lookups through Cloudflare’s 1.1.1.1 resolver; DNSSEC counted when a DS record exists and the resolver returned an authenticated answer.
  • Registry lock. Status codes from the registry’s RDAP service (Verisign for .com); counted when server update prohibited is present.
  • HTTPS. One request to https://www.<domain>/ for the certificate and the Strict-Transport-Security header.
  • When. November 17, 2026, 06:13 to 06:14 UTC. Public lookups only; nothing was scanned or logged into.

Limitations

  • One domain per bank. Banks use many domains, and the FDIC-listed one may not be the login or mail domain.
  • Registry lock is read from public status codes. A bank may protect its domain in ways RDAP cannot show, such as registrar-level controls.
  • 13 sites answered our request with a 403 and 3 timed out; HSTS is measured on the 91 that answered, and a 403 page may carry different headers from the real homepage.
  • DNS operators are inferred from name server hostnames; a bank’s own name servers may be run by a vendor.
  • A snapshot from one moment and one resolver. Settings change.

Download the data and cite this study

The per-bank CSV has 96 rows (one per bank) and 14 columns. It is free to reuse under CC BY 4.0 with credit to Webvello Research and a link to this page.

Download the CSV (96 banks)

APA

Webvello Research. (2026, November 17). How well do the 100 largest U.S. banks protect their domains? A DNS security study. https://www.webvello.com/research/bank-dns-security

One line, for articles

Source: Webvello Research, DNS security study of the 100 largest U.S. banks, November 2026 (https://www.webvello.com/research/bank-dns-security)

Questions about this study

How many of the largest U.S. banks use a registry lock on their domain?

28 of 94 bank domains (29.8%) showed the server update prohibited status that a registry lock sets, on November 17, 2026. Among the 24 top-25 banks with a listed domain it was 15 (62.5%); among banks ranked 51 to 100 it was 5 of 45 (11.1%).

How many large banks use DNSSEC?

24 of 94 bank domains (25.5%) were signed with DNSSEC and validated at a public resolver on November 17, 2026. Only 3 of the 24 top-25 banks with a listed domain were signed: bankofamerica.com, firstcitizens.com and key.com.

Do the largest banks protect their email domain from spoofing?

83 of 94 bank domains (88.3%) published a DMARC policy that tells receiving mail servers to reject or quarantine mail that fails authentication: 74 reject and 9 quarantine. 10 published a monitoring-only policy (p=none) and 1 had no DMARC record on the domain the FDIC lists.

Does using one DNS provider make a bank easy to take offline?

Not by itself. Large DNS providers run many servers in many places. But a domain served by one operator depends on that operator, as sites on Dyn found in October 2016. 83 of 94 bank domains (88.3%) used a single DNS operator in November 2026.

Did the November 16, 2026 bank outage have anything to do with DNS?

There is no evidence it did. Chase said it was a software bug in its on-premise hardware systems, not a cyber issue. This study looks at how banks protect their domains in general, not at the cause of that outage.

Can I use this data?

Yes. The per-bank CSV is free to reuse under CC BY 4.0. Credit “Webvello Research” and link to this page.

Check your own domain’s defenses

We review your registrar lock, DNS hosting, DMARC, CAA and DNSSEC, and tell you what to fix first.