“Astera Labs delivered strong results in Q3 with our revenue and profitability metrics coming in above our outlook.” (CEO)
“Our Aries 6 products are the industry's first and only PCIe 6 retimer solutions ramping in high volume today.” (CEO)
“We expect our overall dollar content opportunity per AI accelerator to significantly increase, representing another step-up from a baseline revenue standpoint.” (COO)
“Q3 non-GAAP operating margins for Q3 reached a new record level of 41.7%, up 250 basis points from the previous quarter.” (CFO)
Topic: Scorpio revenue ramp and customer base clarification
Key points:
Scorpio full-year revenue target is ~10% of total revenue; Q4 exit rate is closer to 20%.
P-Series will continue growing with new designs; X-Series is in low initial volumes now, with a material ramp in 2026.
X-Series is ultimately a bigger opportunity than P-Series and will surpass it at some point (timing not specified).
"10 AI platforms" refers to 10 customers, including custom accelerator developers and hyperscalers buying third-party accelerators for AI servers.
Mgmt stance: Bullish — diverse customer adoption of PCIe-like scale-up protocols, with UALink seen as additive and an upgrade path for higher-speed AI platforms.
Q7 — Quinn Bolton
Topic: Xscale Photonics acquisition and Scorpio X initial shipments
Key points:
Full optical solution requires three pieces: electrical IC, photonic chip, and packaging technology. Xscale acquisition solves two of three (packaging and photonics expertise).
Astera will build an internal photonics team but remains open to using third-party photonics solutions to meet customer requirements.
Scorpio X initial shipments in Q4 are production builds (post-qualification), not samples; big ramp expected in 2026.
Mgmt stance: Neutral-to-bullish — confident in electrical IC capability; optical solution will be built with a mix of internal and external components; Scorpio X production is on track.
Q8 — Sean O'Loughlin
Topic: PCIe-to-UALink transition and Gen 6 revenue clarification
Key points:
Scorpio X currently supports PCIe and PCIe-like protocols; UALink switch will be a new chip, but switching architecture and software features (COSMOS stack) are already ready for next generation.
UALink is an evolutionary step for PCIe customers: both are memory-semantic, lossless protocols; UALink adds higher data rates (200 Gbps and beyond) and AI-specific optimizations.
UALink is an open standard; IP ecosystem is mature; vendors targeting UALink switch deployments in 2026 with revenues in 2027.
The 20% Gen 6 revenue comment includes both Scorpio (all Gen 6 products) and Aries Gen 6 products.
Mgmt stance: Bullish — UALink transition leverages existing development; software stack is reusable; Gen 6 revenue is broad-based across product lines.
Q9 — Sujeeva De Silva
Topic: Optical vs. copper for scale-up and protocol roadmap
Key points:
Customers prefer copper for as long as possible due to reliability, lower power, and better TCO.
Optical will be needed when rack disaggregation (multiple racks) becomes necessary due to power limitations; optical deployments expected in 2028–2029 time frame (POC to revenue).
Astera is heavily engaged in scale-up (PCIe-like today); these engagements are expected to generate revenues into 2029.
Scorpio X-Series supports PCIe and can upgrade to UALink; if customers require alternate protocols (e.g., Ethernet), Astera is well-positioned due to focus on upper-layer data and management functions.
Mgmt stance: Neutral — copper remains primary; optical is a longer-term opportunity; protocol flexibility is built into the platform.
Q10 — Sebastien Cyrus Naji
Topic: China opportunity and NVIDIA design shifts
Key points:
In China, IP availability is limited (200 Gbps not readily available; PCIe Gen 5 and add-in card formats are most common), driving demand for PCIe-based scale-up.
More GPUs are needed in China to solve the same problems (e.g., 16–24 GPU clusters vs. 8 in the U.S.), which increases Astera’s revenue opportunity per cluster (switching, retiming, active cables).
For NVIDIA’s Vera Rubin platform, Astera’s opportunity comes when hyperscaler customers customize the platform for their own infrastructure (similar to Blackwell).
Choice of cable backplane vs. PCB backplane depends on GPU count; Astera does not comment on NVIDIA’s rationale.
Mgmt stance: Bullish on China — constraints create larger GPU counts and higher revenue per cluster; neutral on NVIDIA — opportunity remains in hyperscaler customization.