Open in interactive viewer → charts, metric popovers & call review

Oxford Instruments: The AI Datacom Order Book Hiding Inside a Scientific Instrument Maker

A 44% order surge and a doubled Advanced Technologies backlog have turned a compound-semiconductor niche into multi-year visibility — but the margin story is still mostly ahead.
OXIG.L · Earnings Call · 2026-11-10

The beat is real, but the order book is the story

Oxford Instruments did not just report a good half. It reported a change in the shape of its business. Group order intake rose 44%, revenue 9% at constant currency, adjusted operating profit 28%, and adjusted margin climbed 160bps to 14.9% despite a £1.7m currency headwind. “Order intake was up 44% with growth across both divisions” — Paul Fry, Chief Financial Officer · 2026-11-10 Cash conversion jumped to 85% from 41%, free cash flow improved by nearly £19m, and ROCE reached 30.1%. Those are not cosmetic. But the number that matters is £132m of Advanced Technologies orders, more than double last year, with the divisional order book almost 90% higher. Management now says FY27 revenue is effectively covered and bookings extend into FY28 and FY29. “we now have revenue expectations fully covered for FY '27 with bookings extending well into next financial year” — Richard Tyson, Chief Executive Officer · 2026-11-10 That is why the FY28 Advanced Technologies outlook was raised to roughly 20% revenue growth, with the 12%-15% medium-term margin range now expected a year earlier. The old Oxford Instruments was a diversified scientific instruments company with academic exposure and lumpy capital orders. The new one has a Advanced Technologies division selling into volume manufacturing, an order book that increasingly looks like a semiconductor capital equipment backlog, and a demand driver — Datacom — that sits directly in the AI infrastructure build-out.

The bottleneck is optical, and the customer is booked out

The clearest explanation came from CEO Richard Tyson:

The enormous amount of data that needs to move around an AI data center is driving the need for new technology and an increased volume of indium phosphide datacom chips.

Richard Tyson, Chief Executive Officer · 2026-11-10
Copper cannot carry the bandwidth, reach and power efficiency at higher data rates; optical connectivity can. Oxford makes the front-end wafer processing tools that help produce those optical devices, especially with indium phosphide, a fragile material where precise surface control matters. This is a genuinely company-specific angle. The global tape and recent reporters confirm the datacom optical supply chain is hot: AAOI talked about terabit transceiver products and capacity expansion; 1347.HK cited Silicon photonics; LWLG discussed an optical engine and G per lane roadmaps. Oxford is not a transceiver maker, but it sells the process tooling that sits upstream of that demand. That is why its compound semiconductors order intake grew 173% and why commercial customers are now around 93% of Advanced Technologies orders, with roughly 75% for datacom. The tension is timing. The global tape has been ugly for the AI data center complex over the last 30 days: AI data center is a 30-day decliner with a long list of negative names, and co-packaged optics and high bandwidth memory have sold off. Oxford’s orders are not spot prices; they are multi-year capacity commitments. Tyson said customers are “all booked out to talking about ’29 to ’30” when discussing volumes of chips. “they're all booked out to talking about '29 to '30 in that sort of range where they're talking about giving volumes of chips right now” — Richard Tyson, Chief Executive Officer · 2026-11-10 That does not make the stock immune, but it does explain why management is willing to raise FY28 before FY27 has finished.

Margins and capital allocation are the honest caveats

A demand surge without margin conversion is just revenue. Oxford’s Advanced Technologies margin rose to 7.5% from just over 2%, and Paul Fry said roughly 30% of incremental revenue is dropping through to profit. “if you continue to think that revenue growth should continue to drop through around 30%, and you can work through the margin as a result of that” — Paul Fry, Chief Financial Officer · 2026-11-10 That is respectable, but it explains why FY27 Advanced Technologies margin is still guided to high single digits, not double digits. The offsets are real: group mix is a headwind while Advanced Technologies remains below group margin, currency took about 80bps off the half, and Severn Beach carries roughly £2m of incremental depreciation and running costs in H1. There is also a pricing discipline point that matters. When asked whether tight capacity could produce super-normal pricing, Tyson pushed back: “we're basically pricing ahead of any inflation dynamics and managing to pass that on fine” — Richard Tyson, Chief Executive Officer · 2026-11-10 He added that Oxford is trying to win new volume customers, so it is balancing value capture against entry pricing. In other words, this is not a commodity squeeze. It is a capacity and execution story. The balance sheet supports the investment case. Oxford ended the half with nearly £75m net cash after £21m of buybacks and a £10m dividend, completed £83m of its £100m buyback, and added a further £25m program through mid-2027. The interim dividend rose 7.4%. Management also admitted it got close to an acquisition in H1 but walked away on valuation — a useful sign of discipline rather than empire-building.

What to watch

The first test is Imaging & Analysis execution. Orders grew nearly 13% at constant currency, the closing order book reached £144m, and management expects H2 revenue around 25% higher than H1 to deliver low single-digit full-year growth. Commercial orders rose over 20%, while academic orders grew nearly 3% despite U.S. weakness in Life Science and healthcare. Materials analysis remains the largest I&A market and grew just over 10%. That diversification is the ballast. The second test is Advanced Technologies capacity. Management is hiring on a six-month training cycle and thinking 12-18 months ahead, trying not to overbuild if datacom growth normalizes to 10%-15%. The third test is the FY28 margin bridge: 20% revenue growth at a 30% drop-through should move Advanced Technologies into its 12%-15% range, but supply chain, clean-room throughput and customer ramp schedules all have to cooperate. The stock tape is not available in the context, so the report cannot be paired with a price reaction. That is a limitation. But the fundamental signal is clear enough: Oxford Instruments has moved from a “nice scientific instruments company” narrative to a genuine AI datacom capacity supplier with multi-year order cover. The market may still classify it as a diversified instrument maker; the order book is starting to classify it as something more cyclical, more concentrated and potentially more profitable. The next two quarters will decide whether that reclassification is deserved.