VECO

Resolved PARTIAL on 2026-08-07. The core leg asked for an explicitly dollar-quantified indium-phosphide laser-manufacturing equipment order or backlog figure above the $250M level disclosed at the first-quarter print. It failed on absence: the second-quarter release reports net sales of $193.5M, a raised full-year guide of $780M to $810M and demand attributed to AI, with no InP-specific dollar figure anywhere in it. The confirm leg passed at $193.5M against a frozen baseline of $180.083M, 7.45% above it.

The resolution is more interesting than the verdict. On the same day as the results release, Veeco separately announced that a global leader in optical and photonic technologies had selected its LUMINA+ system to fabricate indium-phosphide lasers for optical-transceiver manufacturing at scale. The substantive fact the leg was reaching for, that InP laser-fabrication capacity is being tooled at volume, was therefore disclosed on the resolution date itself, in a document the leg could not score because it carries no dollars and names no customer. The leg resolved correctly: scoring is done against what the call said, and the call said dollar-quantified. But the record should show that the underlying claim fared better than its measurement did.

Three dollar figures were disclosed on the print and call, and none of them is what the leg asked for. A $700M served-available-market projection for 2030 covering Veeco's role in InP laser manufacturing is a market size, not an order book. An "at least $2 billion over the coming years" silicon-photonics framing from chief executive William Miller is an opportunity assessment drawn from customer engagements, not a commitment. And $200M of orders improving 2027 visibility belongs to advanced packaging, a different product line. Refusing to credit any of the three is the point rather than a technicality: a call that names an InP-specific order book is not satisfied by a market projection, an opportunity framing, or a cross-product order total, however plausible each looks beside it.

Underneath the unquantified metric the business is ramping. Compound-semiconductor revenue was $21M in the quarter, 11% of the total and up 9% sequentially, with full-year compound-semiconductor revenue guided to roughly double the prior year on the strength of the same MOCVD and ion-beam-deposition tools that serve silicon photonics. So the shape here is a business visibly accelerating while the one number the call named went unstated.

For the thesis nothing is contradicted. The claim that the optical architecture is transitioning was reinforced this cycle on fab-layer evidence, and this tool order extends the same pattern of capacity being funded and built one layer further up. The narrower claim about indium-phosphide substrate supply is untouched: it had already been narrowed on 2026-07-31 from a scarcity claim to a time-bounded pricing-power claim, and an unquantified tool order neither supports nor undermines a time bound.

The authoring lesson is that demanding a dollar-quantified figure is simultaneously the strongest verifiability guarantee available and a strong failure risk. Where the analytically important fact is that an order exists rather than how large it is, the call should have tested whether the order was disclosed at all. That would have captured the same signal without conditioning the verdict on a number the company had no particular reason to publish.