A Texas wafer house is now selling strontium titanate to quantum labs. Its launch contains no measured performance data
La Luce Cristallina, a small materials company based in Austin, Texas, announced on Wednesday, Sept. 9, the launch of what it calls a quantum paraelectric strontium titanate on insulator wafer platform — QP-STOI in the company's shorthand — aimed at researchers building cryogenic and quantum devices. The release went out over PR Newswire and was picked up by The Quantum Insider the same day.
The product, as described by the company, is a 0.5-micrometre strontium titanate layer on a thick silicon dioxide layer on silicon. The company's distinguishing claim is thickness: it describes the STO layer as bulk-like and highly insulating, and specifically as 100 times thicker than conventional STO buffer layers used for epitaxial growth on silicon. It says the platform is CMOS- and foundry-compatible, that it is available now in 2-inch wafers, and that it expects to scale to 200-millimetre format in early 2027.
The intended applications listed are specific and credible: parametric amplifiers for qubit readout, tunable microwave components, future electro-optic components, electrically controlled actuators, precision sensors and advanced electromechanical devices. Chief executive Alex Demrov said the wafer provides a highly tunable materials platform for researchers developing quantum and cryogenic devices, and that it helps support increased market interest in emerging use cases. Pricing was not disclosed.
What the announcement does not contain is a single measured number. On a reading of the release itself, the only figures in it are the 0.5-micrometre layer thickness, the 100-times-thicker comparison, the 2-inch and 200-millimetre wafer formats, and a contextual reference to US Department of Commerce quantum investment. There is no dielectric constant, no loss tangent, no tuning range, no temperature at which any of those were characterised, and no third-party verification; The Quantum Insider's write-up adds none either. For a materials product whose entire value proposition is behaviour at cryogenic temperatures, that is a conspicuous absence, and it is the reason this launch should be read as a sampling announcement rather than a performance claim.
The absence is worth dwelling on because strontium titanate has had an extraordinary 12 months in the peer-reviewed literature, and it would be easy to let that halo attach to a wafer that has not yet demonstrated anything publicly.
In October 2025, a team at imec with collaborators at KU Leuven and Ghent University — first author Anja Ulrich, with Christian Haffner as senior author — published in Science: "Engineering high Pockels coefficients in thin-film strontium titanate for cryogenic quantum electro-optic applications," volume 390, issue 6771, pages 390 to 393, DOI 10.1126/science.adx3741, published Oct. 23, 2025. This is the published journal article, not a preprint, and the distinction matters for the figures.
The published abstract reports that strontium titanate can be engineered to exhibit a Pockels coefficient of 345 picometres per volt at 20 hertz at cryogenic temperatures, a value it describes as twice as high as any other thin-film electro-optic material, with what it calls limited optical losses of decibels per centimetre. Imec's own press release describes the result as close to 350 picometres per volt at 4 kelvin, the highest reported for any thin-film electro-optic material at that temperature. Those are the same experiment summarised two ways — the abstract specifies the modulation frequency, the press release specifies the temperature — and a reader who takes only one of them away will be missing half the measurement condition. Haffner said in the imec release that by converting a quantum-paraelectric into a cryo-ferroelectric thin film, the team revealed a powerful Pockels effect where none was expected.
That sentence contains the technical fork that separates imec's result from La Luce Cristallina's product. The Science abstract is explicit about the mechanism: by adjusting the stoichiometry the authors increased the Curie temperature and realised a ferroelectric phase yielding the high Pockels coefficient. In other words, imec got its record by pushing strontium titanate out of the quantum-paraelectric phase and into a ferroelectric one. La Luce Cristallina is selling quantum-paraelectric strontium titanate — the phase imec engineered its way out of — for a different purpose, namely voltage-tunable dielectric behaviour at low temperature, which is what a parametric amplifier or a tunable microwave resonator wants.
These are not competing versions of the same claim, and nothing in the imec work validates the La Luce wafer or vice versa. Anyone reading this week's launch as commercialisation of the Science result has read it wrong. A separate Stanford-led study of the same compound at 4 to 5 kelvin, referenced in imec's release, adds to the material's standing without transferring to any particular vendor's substrate either.
Where La Luce Cristallina does have a demonstrated record is in a different oxide. The company's core business is barium titanate on silicon for electro-optic modulators used in optical interconnects and datacentre links. It launched a 200-millimetre barium titanate substrate in September 2025 and released a beta version of the 200-millimetre BaTiO3 wafer for customer evaluation on March 11, 2026, according to Semiconductor Today, which reported it exhibiting 200mm and 50mm wafers at OFC 2026 at the Los Angeles Convention Center on March 17 to 19. Chief technology officer and co-founder Agham Posadas said the beta version marked another step toward advancing co-packaged optics amid the rise of AI, quantum photonics and other high-capacity applications.
One figure from that barium titanate line illustrates why measurement conditions have to travel with numbers. Semiconductor Today records a Pockels coefficient of about 1,300 picometres per volt — a figure it states explicitly is for bulk single crystal. That is a property of the crystal, not of a thin film on a wafer, and thin-film values for these oxides are routinely far lower. Quoting the bulk figure as a device number would be a category error.
The barium titanate history does give a useful yardstick for the STO roadmap, though. La Luce Cristallina took from a September 2025 substrate launch to a March 2026 beta release to get customer-evaluable 200mm barium titanate out of its Austin fab. Against that, a 2-inch-to-200mm transition for a new oxide platform between now and early next year is aggressive but not obviously fantastical for the same team and the same fab. The Quantum Insider's profile lists the company at 1 to 10 full-time employees, which is the constraint to watch.
For the quantum hardware supply chain the significance is modest but real. Parametric amplifiers for qubit readout are a genuine bottleneck component, and a foundry-compatible tunable dielectric on silicon would matter to anyone trying to move readout electronics closer to the qubits. Whether this particular wafer does that is a question the company has not yet answered with data, and the honest position today is that nobody outside La Luce Cristallina knows.
The company is private and has no listed equity, so there is no direct market read. On the broader tape, the last completed session was Tuesday, Sept. 8, when the Nasdaq Composite closed at 26,421, down 0.32%, and the Russell 2000 at 2,960, down 0.52%, according to Investrade's market review, which also noted mixed semiconductor performance with global chip sales down 9.8% in July from June and memory sales down 16.3% month on month. As of about 11:15 a.m. ET on Wednesday the US session was still open and no closing figure for the day existed.
