The Company
Astera Labs, Inc. (ALAB) is a semiconductor manufacturer specializing in connectivity solutions for cloud and AI infrastructure. Their products include smart retimers, riser cards, and GPU booster cards designed to enhance data center performance.
Financials
Bull Case
- Develops several products for high speed interconnects of AI rack scale deployments and therefore seeing strong tailwinds from the rapid build out of AI data centers.
- Scorpio P series has scored design wins with several hyperscalers during Q3 2025.
Total Addressable Market (TAM) and Compound Annual Growth Rate (CAGR)
Astera Labs (ALAB) has a total addressable market (TAM) of $17.2 billion, which it expects to grow to $27.4 billion by 2027. The company’s compound annual growth rate (CAGR) for revenue is projected to be 39% over the next four years.
Products
Astera Labs offers a range of products designed to optimize data center connectivity:
| Product category | What it includes | % of revenue (estimate) |
|---|---|---|
| Aries — PCIe / CXL retimers & gearboxes | Smart retimers, board/cable extenders and related PHY IP used to preserve signal integrity for PCIe Gen4/5/6 and CXL. (Aries family) | 50–60%. |
| Taurus — Smart Cable Modules / Ethernet connectivity | Active cable modules and optical/electrical interconnect modules for Ethernet and rack interconnect. (Taurus family) | 15–25%. |
| Scorpio — Fabric switches / rack-scale connectivity | Fabric switches and rack-level ASICs for PCIe/CXL fabric (aimed at AI rack/hyperscaler deployments). (Scorpio family) | 10–20% (growing fast as design wins ramp). |
| Leo — CXL / memory expansion controllers | Memory pooling/expansion controllers and CXL adapters (emerging product line). | 5–10% (early stage / ramping). |
| Software / COSMOS & services | Management/monitoring SDK (COSMOS), firmware, validation & customization services for hyperscalers. | 0–5% (minor today; value-add for customers). |
Astera Labs enables the PCIe Scale-Out as well as Ethernet Scale Up of these data centers.
Astera’s Role in PCIe Scale-Out:
- Scale-out means adding more nodes or devices to a system horizontally.
- Think of connecting more GPUs, CPUs, FPGAs, or SSDs to a central host.
- In PCIe-based systems, scale-out means using switches, retimers, and fabrics to expand PCIe lanes across multiple devices and boards.
- Products like Leo (PCIe/CXL Smart Retimers) and Aries (PCIe/CXL Smart Switches) enable scalable, high-speed connectivity between devices inside a server or across multiple servers.
- Use cases: AI training servers, SSD enclosures, GPU clusters.
Nvidia NVLink and ALAB products
NVLink (and its newer generation “NVLink Fusion”) enables high-bandwidth communication between GPUs, GPUs↔CPUs, and across many accelerators in a rack. For that interconnect to become physically deployed and scalable (chip-to-chip, board-to-board, module-to-module) you need supporting hardware: switches, retimers, cables, modules, backplanes. Astera provides many of those.
Astera’s Role in Ethernet Scale-Up:
- Scale-up means boosting the capabilities of individual systems or nodes — making each one faster and more powerful.
- In Ethernet-based networks, this often means increasing bandwidth and interconnect speed to support large workloads in a single system or tightly-coupled cluster.
- The Taurus Ethernet Smart Cable Modules (SCMs) allow servers or switches to scale bandwidth up to 100G/200G/400G/800G, while maintaining low latency and power consumption.
- These are used in AI clusters and cloud infrastructure to move data quickly between devices with minimal cost and complexity.
Scorpio
| Series | Focus / Use Case | Key features |
|---|---|---|
| P-Series | Mixed-traffic / “head-node” connectivity (GPU ↔ CPU/NIC/SSD) | Supports PCIe 6.x/5.x beside mixed devices. “For GPU-to-CPU/NIC/SSD PCIe 6 connectivity” as per the portfolio brief. |
| X-Series | Back-end / GPU-to-GPU / scale-up fabric connectivity (homogeneous accelerator clusters) | Built for “highest back-end GPU-to-GPU bandwidth” with platform-specific customization. |
Customers
Astera Labs serves major cloud service providers and data center operators, offering solutions that enhance performance and reliability in cloud and AI infrastructures.
Competitors
Credo Technology Group Holding Ltd.: Specializes in interconnect solutions for the data infrastructure market, offering products including PCIe retimers and Ethernet connectivity solutions.
Broadcom Inc.: A leading designer, developer, and global supplier of a broad range of semiconductor and infrastructure software solutions. Broadcom’s PCIe switches are widely used in data centers, directly competing with Astera Labs’ connectivity solutions.
Marvell Technology, Inc.: Competes in markets such as data center, carrier, enterprise, and automotive, providing a range of semiconductor solutions.
Founders and History
Astera Labs was founded in October 2017 by Jitendra Mohan, Casey Morrison, and Sanjay Gajendra. Company is headquartered in Santa Clara, California.
Jitendra Mohan serves as the Chief Executive Officer, bringing over two decades of experience in engineering and management roles within the datacenter and server markets. Prior to co-founding Astera Labs, he was the General Manager for Texas Instruments’ High-Speed Interface and Clocking Business.
Casey Morrison holds the position of Chief Product Officer, while Sanjay Gajendra serves as President and Chief Operating Officer.
The name “Astera” is derived from the founders’ affinity for astronomy, with “Astera” meaning “starlike.”
Since its inception, Astera Labs has focused on creating innovative semiconductor solutions to address data bottlenecks, serving major clients such as NVIDIA, Intel, AMD, and leading hyperscale cloud providers.
Invest Alpha 