2.37 million: that was the peak daily volume Microsoft researchers recorded in late February for a phishing wave that hid finance-themed lures inside invisible Unicode characters, a technique the company ties to “ASCII smuggling.”
Microsoft traces a large-scale campaign using invisible Unicode
Microsoft threat researchers discovered a large-scale phishing campaign that used invisible Unicode characters drawn from the Unicode Tags block (U+E0000–U+E007F) to evade email filters. The company reported the operation reached as many as 2.37 million messages per day in late February and that the high-volume phase “persisted for roughly three months after February 9 and dropped sharply after May 15, 2026.” Microsoft stressed these dates “bound the observed use of the specific technique in our telemetry, not the broader campaign, which started earlier without it and continued without it.”
How the Unicode “split” defeats keyword filters
The attackers inserted an invisible Unicode character inside finance-related lure words so the words were split for keyword- or list-based detectors. Microsoft illustrates the method by showing how a keyword like “funding” can be transformed into “fun[invisible character]ding,” allowing messages to bypass defenses that rely on contiguous tokens. According to Microsoft, the method “has been used in millions of finance-themed phishing messages and works as intended.”
The cyber insurance questionnaire just landed. Now what?
SOC 2, HIPAA, insurance renewals - someone has to own security strategy. Nubivance provides fractional CISO leadership without the full-time salary.
Get a security leadInfrastructure and scope: ActiveCampaign, 148 sender domains, finance lures
Microsoft identified a cluster of 148 finance-themed sender domains on February 9 that accounted for about 96% of all messages its Defender for Office 365 hunting logic flagged for Unicode-tag signatures. The domains used words such as “funding,” “capital,” “loan,” “advance,” and “credit,” and the emails promoted business funding, loans, and credit services. Delivery occurred through infrastructure associated with the legitimate ActiveCampaign email-marketing platform. After Microsoft reported the abuse, ActiveCampaign responded that “its moderation systems detect invisible Unicode characters the same way they detect unobfuscated text and treat heavy use as suspicious.”
Defender performance and Microsoft’s mitigation advice
Despite the technique's ability to bypass simple keyword checks, Microsoft reported that Defender still identified over 99% of the messages using other signals — sender, IP, domain, and reputation checks. To reduce the specific risk from invisible Unicode characters, Microsoft recommends stripping or normalizing Unicode tag characters and other invisible code points before applying keyword, regex, or signature-based detection. The company also urges treating unexpected tag-block characters as a strong anomaly and recommends applying the same normalization before passing email content to AI assistants to mitigate prompt-injection attacks.
What this means for technologists, affected enterprises, and end users
- Technologists and security teams: Normalize and strip tag-block Unicode characters prior to tokenization and heuristic checks; treat unexpected tag-block characters as an anomaly. Microsoft’s telemetry shows the technique can scale to millions of messages but remains detectable by layered signals.
- Affected enterprises and procurement leaders: Note that delivery can come through legitimate third-party marketing platforms — in this case, ActiveCampaign — which those platforms say monitor for heavy use of invisible characters. Review third-party email marketing controls and moderation processes where phishing campaigns leverage legitimate infrastructure.
- End users and the general public: Finance-themed messages using words like “funding,” “capital,” “loan,” “advance,” and “credit” may be obfuscated by invisible characters; suspicious requests for business funding or loan services merit extra caution even if keyword-based filters do not flag them.
Context from the Blue Report 2026
The Blue Report 2026 — which “measures defenses technique by technique across 338 million simulations run in customer production environments” — highlights a broader operational point: “Overall prevention scores can hide what happens after initial access. Once attackers are using valid credentials, prevention drops sharply.” That observation underscores why detection and remediation beyond initial filtering remain essential.
Microsoft’s telemetry shows the invisible-Unicode trick was both simple and scalable: effective enough to be deployed in a mass campaign, visible enough for defenders to trace and advise countermeasures. The campaign’s high-volume window — bounded in telemetry between February 9 and a sharp decline after May 15, 2026 — and the clustering around a small set of sender domains tied to finance lures provide concrete signals defenders can use to hunt and harden controls. The remaining open question is whether attackers will pivot to new invisible code points or different delivery channels; defenders now have a specific, actionable countermeasure to apply.
