Several of the largest firms on Wall Street were targeted in early August by a coordinated AI voice-phishing campaign. Named as targets: Point72, Millennium, Two Sigma and Citadel, alongside several private-equity firms. FINRA was in contact with member firms about the attempted breaches.
The technique
Vishing is not new; the cloning is. Attackers synthesised the voices of colleagues an employee would recognise and used them by phone to request credentials or system access. Voice has functioned as an informal authentication factor inside finance for decades, and that is precisely the assumption being attacked.
Who has said what
Two Sigma (~$75bn AUM) said it blocked the attempt and has "no indication of any impact to our data or systems". Point72 told investors it was attacked and that an initial review found no client information stolen, with the review ongoing. Millennium and Citadel declined to comment — which is not a statement that nothing happened.
Volume is the change
Vinod Paul of Align Managed Services framed the economics bluntly: an operation that could previously mount 50 targeted attacks can now mount 1,000. Antithesis's Will Wilson described AI as having commoditised the capability. Google's security unit flagged a comparable vishing wave against law firms in June 2026.
What is unknown
No attribution has been published, no vendor has named the tooling, and no dollar loss has been reported at any firm. What is established is the targeting, the method, and one confirmed block.