Astera Labs (ALAB) UBS Global Technology and AI Conference summary

Company overview and market context

• Founded in 2017 to address AI infrastructure needs as compute shifted from CPU-centric to heterogeneous architectures.

• Technology enables interconnection of CPUs, GPUs, and ASICs, supporting protocols like PCI Express, Ethernet, and CXL.

• Headquartered in California with global offices.

• Went public earlier this year; technology widely used in hyperscaler AI deployments.

Growth drivers and business opportunities

• Over 80% of revenue comes from AI deployments, leveraging both third-party GPUs and custom ASICs.

• AI infrastructure investment expected to reach $300 billion next year, doubling from last year.

• Products address critical bottlenecks in data, networking, and memory, improving GPU utilization.

• Expanding product lines and increasing dollar content per GPU, especially with new fabric devices.

Product and technology differentiation

• Agnostic to GPU or ASIC platforms, but higher dollar content per GPU on ASIC-based platforms.

• Scorpio X-Series targets GPU-to-GPU interconnects, offering higher value and ASP.

• Taurus product line focuses on Ethernet AECs, using a module approach for flexibility and profitability.

• Leo product line addresses CXL memory expansion, with broader adoption expected from 2H 2025.