Veeco (VECO) Q2 2026 earnings review

Massive AI Optics Inflection Drives Beat and Raise

Veeco definitively reversed a multi-quarter stagnation, delivering Q2 revenue of $193.5 million (+16% YoY) and comfortably beating its own guidance. The core narrative has shifted dramatically from traditional semiconductor cycles to a massive AI infrastructure build-out, specifically the transition from copper to Indium Phosphide (InP) optical interconnects. This inflection is powering an accelerating growth trajectory, driving a significant raise in full-year guidance to $780-$810 million. However, looming over this operational success is the unresolved Axcelis merger and the colossal execution risk of scaling manufacturing capacity 10x by 2027.

๐Ÿ‚ Bull Case

AI Optics Supercycle

The $250M+ order pipeline announced last quarter for InP laser manufacturing is translating into immediate revenue acceleration, with Q3 revenue guided up ~26% YoY at the midpoint.

Visibility into 2027

Unprecedented backlog depth across optical interconnects, HBM packaging, and a rebounding Data Storage (HAMR) market gives Veeco nearly fully booked capacity through the end of 2026.

๐Ÿป Bear Case

Severe Manufacturing Execution Risk

Capitalizing on the AI optics boom requires scaling Spector IBD capacity by 10x by early 2027. Any supply chain or operational stumble will immediately impact revenue realization.

Merger & Regulatory Limbo

The highly-touted all-stock merger with Axcelis remains hostage to Chinese anti-trust regulators, creating a strategic overhang while the core mature-node China business actively decays.

โš–๏ธ Verdict: ๐ŸŸข

Bullish. The financial reality is finally matching the technological narrative. The transition to optical interconnects in AI data centers provides a structural, multi-year tailwind that easily overrides legacy segment weakness.

Key Themes

DRIVER ๐ŸŸข๐ŸŸข

Indium Phosphide Lasers (Optical Interconnects)

The transition from copper to optical interconnects in AI data centers is the primary growth engine. Veeco's Spector IBD system (for laser facet coating) and Lumina MOCVD systems are critical to this $2B multi-year opportunity. After announcing a massive $250M+ backlog in Q1, Q2 results and Q3 guidance confirm this revenue is aggressively materializing.

DRIVER ๐ŸŸข

Advanced Packaging and LSA for AI Logic/Memory

Beyond photonics, core semiconductor strength is driven by Laser Spike Annealing (LSA) for High-Bandwidth Memory (HBM) and Gate-All-Around (GAA) logic. Wet processing systems for 2.5D/3D advanced AI packaging continue to see robust pull from top-tier OSATs and foundries.

DRIVER โšช

Data Storage Recovery Cycle (HAMR)

The hard disk drive (HDD) market is emerging from a severe slump, driven by AI data generation and the shift to Heat-Assisted Magnetic Recording (HAMR). Management previously noted this segment is fully booked for 2026, contributing to the heavily H2-weighted revenue ramp.

CONCERN ๐ŸŸข

Massive Capacity Ramp Execution Risk

Veeco is attempting to scale its Spector IBD manufacturing capacity by 10x from its base level by early 2027 to meet optical interconnect demand. Operating margins will be highly sensitive to any inefficiencies, yield issues, or supply chain bottlenecks during this aggressive expansion phase.

CONCERN โšช

Axcelis Merger & China Anti-Trust Overhang

The strategic combination with Axcelis, designed to create a $5B+ SAM powerhouse, remains unclosed pending Chinese regulatory approval. Given current geopolitical tensions, this represents a significant structural uncertainty.

CONCERN โšช

China Mature Node Decay vs AI Boom

While overall revenue is surging, there is a stark contradiction beneath the surface: the legacy China mature-node business is aggressively decaying. In 2024, China was routinely ~40% of revenue; by 25Q4 and 26Q1, it had dropped to 13-20%. The AI boom is currently masking this geographic demand destruction and the continued threat of BIS export licensing friction.

Other KPIs

Non-GAAP Gross Margin 39.5%

Reversing the dip seen in Q1 (36.2%). The margin recovered nicely into the high-30s, reflecting a normalization of product mix and absorption of lost high-margin China LSA shipments from prior quarters. Guidance for Q3 projects further expansion to 41-42%.

Operating Income (Non-GAAP) $23.1 million

Accelerating significantly from $8.6 million in Q1 2026. Operating leverage is returning as top-line volume scales against fixed costs, generating an 11.9% Non-GAAP operating margin.

Guidance

26Q3 Revenue $200 - $220 million

Accelerating sharply. At the $210M midpoint, this implies 26.6% YoY growth (compared to $165.9M in 25Q3) and 8.5% sequential growth. This confirms the start of the heavy H2 delivery ramp for recent AI-related bookings.

26FY Revenue $780 - $810 million

Accelerating. Raised from the prior range of $740 - $800 million. With H1 actuals at $351.8M, the midpoint of the new guidance ($795M) implies an H2 revenue run-rate of $443.2M, a massive 26% step-up compared to H1.

26Q3 Non-GAAP Diluted EPS $0.35 - $0.49

Accelerating. The $0.42 midpoint represents sequential growth from Q2's $0.33, driven by the expanding gross margins (guided at 41-42%) and higher top-line volume.

26FY Non-GAAP Diluted EPS $1.36 - $1.61

Stable to slightly adjusted. This reflects tight cost controls amid the aggressive capacity scale-up, projecting significant earnings power in the back half of the year.

Key Questions

Axcelis Merger Contingency

With the Axcelis merger still pending Chinese anti-trust approval, is there a drop-dead date for the transaction, and what is the standalone strategic pivot if approval is blocked?

Margin Profile of Spector Ramp

As Spector IBD capacity scales 10x by 2027 to fulfill the $250M+ order pipeline, how will the margin profile of these tools compare to the corporate average once volume discounts and initial ramp inefficiencies are factored in?

Memory Evaluations Update

Evaluations for IBD300 systems at leading DRAM customers were previously extended through the end of 2026. Has there been any acceleration in customer timelines or resolution of the particle performance engineering work?