shuttle run

A shuttle run (Multi-Project Wafer or MPW) is a **cost-sharing arrangement** where multiple chip designs from different customers share the same set of photomasks and wafer lot, dramatically reducing the cost of prototype fabrication. **How It Works** Instead of one company paying for a full mask set ($1-10M+), multiple designs are **tiled onto the same reticle**. Each company gets a small area (a few mm²) on the mask. All designs are fabricated together on the same wafer lot. After processing, wafers are diced and each company receives their dies. **Cost Comparison** • **Full mask set (dedicated run)**: $1-10M+ for masks alone at advanced nodes, plus wafer processing costs • **Shuttle run**: $10K-200K depending on node and die area. Makes advanced-node prototyping accessible to startups and universities **Shuttle Providers** • **MOSIS** (now Synopsys): The original MPW service. Primarily academic and small-volume • **Europractice** (IMEC): European MPW service for multiple foundries • **TSMC**: Runs regular shuttle schedules (CyberShuttle) for all major nodes • **Samsung**: MPW services through Samsung Foundry • **GlobalFoundries**: MPW programs for their technology nodes **Typical Shuttle Schedule** Foundries offer shuttle runs on a **fixed calendar** (e.g., monthly or quarterly starts per node). Designers must submit their designs by the shuttle deadline. Turnaround time: **3-6 months** from tape-out to die delivery. **Limitations** Shuttle runs provide **small quantities** of dies (tens to hundreds, not thousands). Process conditions are shared (no custom recipe tweaks). Not suitable for volume production—just for prototyping, characterization, and design validation. Once the design is verified, companies order a dedicated production mask set.