AEHR TEST SYSTEMS Q1 FY2027 earnings call

• Record Financial Start: Delivered record quarterly revenue of $33.0 million (+200% YoY, +75% QoQ) and non-GAAP net income of $9.4 million, demonstrating strong operating leverage as revenue scales.
• Expanding AI Demand: Demand for wafer-level and package-level burn-in is accelerating due to increasing semiconductor complexity, power requirements, and mission-critical applications in AI data centers.
• Wafer-Level Burn-in Progress: Significant progress with a major AI accelerator supplier who completed benchmark evaluations; installation of a fully configured XP system at an OSAT partner in Taiwan is planned for early calendar 2027. Additional production orders received for silicon photonics and improving demand in silicon carbide and gallium nitride.
• Package-Level Expansion: Shipping over a dozen Sonoma systems per month to support lead hyperscale customer’s AI processor ramp. Received follow-on orders for the second AI processor, which will reach higher volumes than the first.
• Diversified Customer Base: Expanded from a handful of key customers to more than a dozen potential sources of significant production orders across multiple markets, reducing dependence on any single program.
• Manufacturing Capacity: Increased manufacturing capabilities, including shipping Sonoma systems from Southeast Asia via contract manufacturers, adding capacity for 20+ additional systems per month without significant infrastructure spending.

The company reports a unified business segment focused on semiconductor test and burn-in solutions. First quarter revenue reached a record $33.0 million, representing 41% of total revenue from consumables (wafer pack contactors and Sonoma modules) and the remainder from system sales. The consumable segment grew significantly, contributing 41% of revenue compared to 24% in the prior year quarter, driven by expanding installed bases and new device introductions.

• Raised Fiscal 2027 Revenue Outlook: Increased full-year revenue guidance to $150 million (from previous range of $130–$150 million), reflecting 200% growth over fiscal 2026.
• Profitability Expectation: Expects non-GAAP net income of at least 20% of revenue for the full fiscal year.
• Backlog Visibility: Effective backlog reached $109.3 million as of November 1st, providing substantial visibility into future performance.
• Sustained Growth Trajectory: Management expressed confidence in continued growth beyond fiscal 2027, supported by expanding manufacturing capacity and a broad pipeline of qualification programs.

• Supply Chain Constraints: While currently not supply-limited, management acknowledges the need to shore up resilience for larger volumes, evidenced by pre-purchasing long-lead items and investing in supplier relationships.
• Execution Risk on New Programs: The success of the large benchmark opportunity with the major AI processor supplier depends on successful high-volume production validation in Taiwan, with decisions anticipated several months post-installation.
• Market Adoption Uncertainty: Some memory opportunities (flash) are progressing slowly due to limited incremental manufacturing capacity and timing of new architectures.
• Competitive Landscape: Competitors are attempting to enter the space using System Level Test (SLT) derivatives, though management believes these lack the thermal precision and cost-effectiveness of dedicated burn-in solutions.

Q: Analyst asked for clarity on the revenue potential of the 'largest opportunity ever pursued' with the benchmark AI customer, noting the timeline for high-volume decision-making. / A: CEO explained that complex multi-die packages (combining AI accelerators, HBM, and IOs) require precise thermal control at the wafer level, which is impossible at the package level. He quantified that for every million units shipped annually, 5-10 systems are needed per configuration, leading to potential revenues of $35-$70 million per million units. With accelerator market volumes around 25 million units in 2027, winning even one or two high-volume devices represents a massive deal.

Q: Analyst inquired about the expected revenue contribution from Silicon Carbide in FY27, given improved demand and preferred vendor status. / A: CEO noted that SiC represented less than 10% (~$10 million) of FY26 revenue, down from 90% historically. He confirmed that SiC revenue is expected to grow in FY27, citing recent orders exceeding $7 million and full utilization of installed base. The growth is driven by EV traction inverters, charging infrastructure, and solar, with emerging data center applications on the horizon.

Q: Analyst questioned the linearity of revenue recognition and supply chain resilience given the raised guidance to $150 million. / A: CEO indicated that demand is front-loaded, with customers requesting earlier shipments, suggesting Q4 might be the highest quarter. Regarding supply chain, he stated they have already invested in capacity, moving subassembly work to subcontractors to reduce final assembly time in Fremont by two-thirds, enabling 3x capacity. They are also securing long-lead components to buffer against disruptions.

Q: Analyst asked how competitors' new SLT-based burning capabilities compare to AIR's Fox and Sonoma platforms. / A: CEO distinguished SLT machines (which boot devices in native application mode) from true burn-in systems. He argued SLT cannot replicate the high-voltage/high-temperature strain required to screen infant mortality effectively without risking wear-out or failing to handle thermal runaway in multi-die packages. He emphasized that AIR's wafer-level solution offers superior parallelism and independent thermal control, making it the only viable option for next-gen complex packages.

Q: Analyst sought clarification on the pacing of the benchmark customer's Taiwan installation and the role of proprietary technology in the deal. / A: CEO confirmed the early 2027 timeline is consistent with expectations. He highlighted that the system includes patented features, such as independent die temperature control (DITC), which prevents thermal runaway and ensures optimal burn-in conditions for all dies in a multi-die package. This capability is critical for avoiding infant mortality failures in edge inference and data center applications where traditional methods fail.