27 billion files. One trillion edges. A decade of digital stewardship. The 2025 Software Heritage Activity Report snapshots a defining moment. As the organization heads into its 10th anniversary year, this report captures an Archive scaling into its role as permanent, global infrastructure.

Highlights from the 2025 report:

• The Archive now protects over 27 billion unique source files from 421 million projects. To put that growth in perspective, that means ingesting 1.3 full projects every single second.
  
• Solving the « 2PB Problem »: As the Archive hit the 2-petabyte mark, our team, working with the Universities of Pisa and Milan, deployed new compression protocols that shrunk 78TB of raw data into a 3TB research dataset, finally making massive-scale code analysis accessible to academic labs everywhere.
  
• The SoftWare Hash IDentifier (SWHID) was officially published as ISO/IEC 18670. It’s now the international standard for verifiable, decentralized software identification.
  
• The CodeCommons project continues to build the transparent, traceable foundation needed for responsible, sovereign AI.
  
• OSPO-RADAR—a two-year, Sloan Foundation-funded project designed to help universities finally find their scattered software assets and build the dashboards they need to prove their research impact.
  
• With the mirror launched in Greece with GRNet, we’re entering our 10th year as a geographically distributed « preservation fabric » designed to outlast the platforms it tracks. More are in the works: UNIDue (Germany) is currently finalizing deployment and is expected to join the live network soon, alongside several more organizations.

Heading into 2026, the new Software Heritage Advisory Board—comprised of experts from the Eclipse Foundation, Amazon, Anaconda, and more—will guide our transition into a fully independent, global multi-stakeholder organization. We aren’t just archiving the past; we are securing the « source of truth » for the next century of innovation.