Why This Case Still Matters
The UK energy firm case might appear a bit historic. Not quite. It gave birth to every case that followed: UAE ($35M), Arup ($25M), Ferrari, WPP. Every case uses the exact same technique: a deepfake voice impersonating a trustworthy person.
In 2019, it was the first-ever deepfake fraud. In 2026, it is routine practice.
How the Attack Was Performed: 3 Calls in a Row. 3 Manipulations
The perpetrators of the attack called three times, each time with a slightly altered purpose [1].
Call one: instructions to urgently transfer €220,000 to the Hungarian supplier and pay off immediately. The CEO did.
Call two: assurance that a refund will be coming soon.
Call three: new instructions for another transfer, with the difference that now the call showed an Austrian phone number, even though it originated in Germany. The CEO stopped cooperating.
The money from the first transfer had been transferred to another country and could not be traced back.
What Nobody Saw Then. Nobody Sees Now
“His voice sounded like the voice of his German boss. Accent? Right. Melody? Perfect,” noted insurance expert Rüdiger Kirsch from Euler Hermes, when describing the case. Back in 2019, this seemed unbelievable. But it was far from [2].
Voice cloning software existed at that time. It needed about 30 seconds of good quality voice and some computational power. For the CEO of any company that holds conferences, interviews, or publishes YouTube videos, this is not an obstacle [3].
From $243K to $35M: An Increasingly Troubling Trend
2019: First deepfake voice, €220,000.
2020: First deepfake voice against a bank, $35,000,000.
2024: First deepfake video team, $25,600,000.
These three numbers are not random. They are milestones on a chart. The cost of technology is going down. The sophistication level of attacks is increasing. The victims are becoming bigger and richer.
What is consistent in all these cases: the companies that fell victim to attacks did not have technology to recognize deepfakes.
This technology is affordable now. It cost less than dinner for two back in 2019 and produced an $243K return. In 2024, it cost twice as much but generated $25M profit. Costs are decreasing. Benefits are increasing.
— SYNHAWK Analytics Team
How SYNHAWK Protects Against This Type of Attack
SYNHAWK PROTECTION: The British energy firm case was the first documented instance of a deepfake attack. It happened because of technological limitations. If this case took place today, HAWK 7, SYNHAWK’s foundational model for audio, would recognize the deepfake and stop the attack during the first phone call, before the money transfer took place.