Project Silica: Data in Borosilicate Glass

Project Silica reports a Nature breakthrough: encoding data in ordinary borosilicate glass using femtosecond lasers, enabling phase-voxel single-pulse writes, parallel multi-beam writing, simplified single-camera readers, ML decoding, and accelerated aging tests indicating potential 10,000-year data retention.

Project Silica reports a Nature-published breakthrough in glass-based archival storage. The team extended femtosecond laser encoding from fused silica to borosilicate glass.

Main feature and impact

Project Silica introduces phase voxels and pseudo-single-pulse birefringent writing in borosilicate glass. Phase voxels require a single femtosecond pulse to record phase changes reliably. This reduces writer complexity and cost while enabling hundreds of data layers in two millimeters of glass. Readers now use a single camera, shrinking reader hardware and lowering retrieval expense.

Practical implications

The technique enables parallel multi-beam writing and faster encoding through pre-heating and post-heating models. Machine learning mitigates three-dimensional inter-symbol interference for phase voxels. Nondestructive optical aging tests and accelerated aging suggest data retention for at least 10,000 years. Media availability and lower cost support potential commercialization and easier manufacturing.

“Now, we are excited to report significant progress on Project Silica, our effort to encode data in glass using femtosecond lasers, a technology that could preserve information for 10,000 years. Glass is a permanent data storage material that is resistant to water, heat, and dust.”

The research phase has concluded and results are published for community use. Next steps include engineering prototypes, scaling multi-beam writers, and validating long-term retention under real-world conditions.

Key points from the article:

Uses common borosilicate glass, lowering media cost and increasing availability

Phase-voxel method creates data with a single femtosecond laser pulse

Multi-beam parallel writing boosts encoding throughput

Reader simplified to one camera, reducing hardware complexity

ML decoding and accelerated aging support projected 10,000-year data retention

Related Coverage:

• Today in Nature Magazine, we shared new results from Project Silica—our work to encode data in glass—pushing toward durable, immutable archival storage designed to last for millennia.
• Building Qiddiya City: How Copilot helps Abdulrahman AlAli navigate a project of unprecedented scale
• The data behind the design: How Pantone built agentic AI with an AI-ready database

From the Source

---