Thesis Revisit: Why We Do This
One discipline we hold ourselves to at GZC is the formal revisit. When we publish a research thesis, we commit to reviewing it publicly at a reasonable interval — not to celebrate what worked, but to honestly assess what we got right, what we missed, and what the forward setup looks like given updated information. This is that review for our optical interconnect thesis, originally published in March 2025 with specific focus on Astera Labs (ALAB) as a primary expression of the idea.
The Original Thesis: A Brief Recap
Our March 2025 optical interconnect research made three core arguments:
- AI compute scaling was creating bandwidth constraints between accelerators and memory that copper interconnect could not sustainably address at the densities required for next-generation AI systems.
- Optical interconnect solutions — specifically the active electrical cable (AEC) and connectivity subsystem products at the core of ALAB's product portfolio — were positioned to benefit from a multi-year adoption curve as AI infrastructure buildout accelerated.
- ALAB's competitive position, based on deep co-development relationships with hyperscaler customers and a product portfolio designed around AI-specific connectivity requirements, created a durable moat in a market with high switching costs.
What Has Played Out
Twelve months on, the core thesis has largely validated. Here is our honest accounting:
Demand durability: The bandwidth constraint argument has proven correct and has arguably become more visible, not less. Multiple hyperscalers have publicly acknowledged the role of high-speed interconnect as a critical bottleneck in their AI infrastructure buildout. ALAB's revenue trajectory has reflected strong execution against this demand environment.
Co-packaged optics adoption: Our original thesis anticipated that co-packaged optics (CPO) would begin moving from lab evaluation to production deployment within 18 to 24 months of publication. That timeline has largely held. Several major switch vendors have announced product roadmaps with CPO integration, and hyperscaler procurement for pilot deployments is underway. This remains a medium-term catalyst rather than an immediate revenue driver, but the technical validation is proceeding on schedule.
Competitive positioning: ALAB has maintained its position as a preferred supplier to the largest AI infrastructure buildouts. The co-development relationships that underpinned our original moat assessment have, if anything, deepened.
What We Underestimated
Our original analysis did not fully account for the pace at which the market would re-rate ALAB's valuation in response to AI infrastructure enthusiasm broadly. The stock's move in mid-2025 significantly outpaced what our fundamental models projected over the same timeframe, driven in part by multiple expansion as investors assigned a premium to AI infrastructure purity.
This created a position management challenge: we held conviction in the fundamental thesis but recognized that the valuation had moved ahead of even our optimistic estimates. We trimmed from peak weighting, capturing significant gains, but the position continued to appreciate further. The lesson, which we noted in our Q3 2025 investor letter, is that in strong secular demand environments, multiple expansion can persist longer than a fundamental framework would suggest.
The Forward Valuation Framework
As of February 2026, we are applying an updated framework to assess the forward return potential from current prices:
Revenue visibility: ALAB's backlog and customer concentration in hyperscalers provides strong near-term revenue visibility. We are modeling 2026 revenue in a range consistent with management guidance and our own channel checks.
Margin trajectory: As the product mix shifts toward higher-margin systems-level solutions, we expect gross margin expansion to continue. This is a meaningful driver of forward earnings power.
CPO optionality: We value co-packaged optics adoption as a call option on the thesis extending beyond the current product cycle. Our base case does not require CPO to drive returns over the next 12 months, but the option has meaningful value.
At current prices, we see a favorable but less compelling risk-reward than at our original entry. We remain positioned but have reduced to a size that reflects this updated view.
Key Catalysts to Monitor
Looking ahead, we are watching three specific developments:
Hyperscaler AI capex guidance: Any revision — up or down — to forward capex commitments from the major cloud providers directly affects the demand visibility that underpins ALAB's growth profile.
CPO product announcements: Formal product launches and customer adoption announcements from switch vendors moving CPO from roadmap to shipping product would be a significant positive catalyst.
Competitive dynamics: We continue to monitor whether any competing optical interconnect vendors are gaining meaningful design wins in hyperscaler AI infrastructure — a development that would require a reassessment of ALAB's competitive moat.
Conclusion
The ALAB thesis has been one of our more successful research-to-investment executions. The core insight — that bandwidth constraints in AI compute infrastructure would drive durable demand for high-speed optical interconnect solutions — proved correct, and the specific company we identified as the primary expression of that thesis has executed well. We remain invested, more selectively sized, and actively monitoring the catalysts that will determine the next chapter of this story.
