FOR IMMEDIATE RELEASE
Boston, Massachusetts. (July 07, 2026) Microsoft, Google and AWS are exiting the voice biometrics voice verification space. All to be out by the beginning of 2028. Industry analysts concluded that while all three companies produced ‘good enough’ products at unbeatable price points but without the customization, fraud watchlists, or synthetic speech detection needed by top-tier banks, and critically, none of them even today can successfully detect and defeat AI Deep Fake voice.
Leaving Pindrop as the top voice biometric player according to the same analysts. However, VoiceVantage’s voice verification biometrics scores surpass Pindrop’s in PlayHT AI Deep Fake testing as well as in verification speed as demonstrated in independent tests by NPR for Pindrop and by a leading Canadian university for VoiceVantage.
Each company approached their testing using different methodologies; VoiceVantage used enrollment vs Pindrop nonenrollment, VoiceVantage 1.4 milliseconds verification vs Pindrop 6.24 seconds. Pindrop used television recordings, VoiceVantage used cell phones (comparable to television), landlines and laptop recordings.
Pindrop
https://www.pindrop.com/article/npr-audio-deepfake-study/
VoiceVantage went further and demonstrated in a total of 533 University tests to be able to not only immediately block the top five AI Deep Fake Voice products, FakeYou, ResembleAI, ElevenLabs, PlayHT, and Descript, but in further independent university tests clearly distinguished between identical twins using a single spoken word in any language in 1.4 milliseconds.
All VoiceVantage™ products provide unmatched security features and benefits. Our solutions strengthen identity protection, and our speed and accuracy accelerate verifications ensuring timely, accurate and secure facilities access as well as protecting all levels and methods of communication security. For more information visit our website: www.voicevantage.com
VoiceVantage™ VoiceCheck™ Specifications
Phrase Length and Speech Print Size:
1-2 seconds
7-10K
Verification Data/Time:
1 individual phrase 1.4 milliseconds
1 phrase, 2000+ verifications per second on Intel Core i7-7500U @ 2.70GHz
(based on simultaneous use).
Sample Types:
16-bit signed and unsigned linear
8-bit signed and unsigned linear
8-bit mu-law and a-law
Sample Rate:
8, 11.025 or 16 KHz
Interfaces:
Direct C++ API calls (Windows, Linux), .NET (Windows), Java (Windows, Linux, Android), Objective C (iOS,
macOS), Swift (iOS, macOS)
Database Access:
VoiceCheck™ ODBC allows storage of all VoiceCheck™ data in a Relational Database Management
System (RDBMS). Currently the following RDBMSs are supported:
| RDBMS Version | Windows | Linux | Notes |
|---|---|---|---|
| Microsoft SQL Server 2019 | ✔ | ✔ | Linux tested using connection to database instance running on Windows |
| MySQL 8.0.28 | ✔ | ✔ | Windows tested using connection to database instance running on Linux |
| Oracle 21c | ✔ | ✔ | Linux tested using connection to database instance running on Windows |
| PostgreSQL 12.10 | ✔ | ✔ | Windows tested using connection to database instance running on Windows |
Supported Languages:
C++, C#, Java, Objective C, Swift
Sample Code:
Demo applications in all supported languages are provided as usage examples to streamline integration
Supported Development Environments:
Microsoft Visual Studio 2022, GNU G++ 9.3.0 (Linux), Eclipse, Xcode release 12.3, Android SDK version 10.0 (Q) API Level 29 Revision 5, Android Studio release 4.1
Minimum System Requirements:
Windows: Intel i5 or equivalent or higher, 2GB RAM, Windows 10 or higher
Linux: Ubuntu 20.04 LTS
Android: Lollipop 5.0.2
macOS: OS X Yosemite