Deepfake ads impersonating Nu founder spread investment scam
An organized operation deployed 51 fraudulent advertisements across a single 30-day window in mid-March to early April 2026, using AI-cloned face and voice of David Vélez, founder of the digital bank Nu, to promote a fake investment scheme on Facebook and Instagram, Revelum reported citing a Forbes Colombia investigation published April 9, 2026.
Deepfake Technique and Scale
The campaign weaponized synthetic media to impersonate a high-profile fintech founder, leveraging platform reach to target potential investors. The cloned face and voice mimicked Vélez's likeness and speech patterns in promotional content designed to lend false credibility to the investment pitch. By concentrating 51 ads in 30 days, the operators maximized exposure velocity before detection—a pattern consistent with time-bound fraud campaigns that exhaust platform detection budgets.
The use of a real founder's identity lowered victim friction; audiences familiar with Nu or Vélez's public profile faced reduced skepticism when encountering the synthetic endorsement. This represents a convergence of two agent-economy harms: automated ad placement at scale paired with deepfake synthesis to bypass human verification.
Platform Distribution and Reach
Both Facebook and Instagram—Meta's core ad networks serving billions of users—were the distribution channel. The choice reflects the platforms' scale, targeting granularity, and historical vulnerability to fraud rings. Neither platform has disclosed the scope of victim complaints, dollar volume defrauded, or enforcement action taken against the perpetrators as of this reporting.
Investigation and Reporting
Forbes Colombia published the investigation in April 2026, surfacing the campaign after it had run its 30-day cycle. The Revelum citation indicates the report documented the campaign structure and platform footprint; however, no court filing, regulatory action, criminal arrest, or victim recovery figures have been independently verified from law enforcement or financial regulators.
This incident aligns with a broader 2026 trend of organized fraud using deepfaked celebrity and executive identities to route investors into fraudulent cryptocurrency and investment schemes. The specificity of the 51-ad volume and 30-day window provides quantifiable evidence of campaign scale, distinguishing it from anecdotal reports of isolated deepfake fraud.
Implications for Agent Oversight
The case underscores the automation risk at the intersection of synthetic media generation and programmatic ad delivery. An agent-driven ad system—whether human-operated or algorithmic—can distribute deepfake content faster than manual review cycles catch it. Platform policy on synthetic media in ads has tightened since 2025, but enforcement remains reactive rather than preventive.