Startup Insights – SweGan

SweGaN: $14M Series B, 7 patent families, and a thin-but-real moat

The most interesting number in SweGaN’s file is a date. The last public filing is July 20, 2021, which is 61 months before this round. Five years of work sits behind the publication curtain, because anything filed since then has not yet published. So the seven families on the public record describe what SweGaN was protecting halfway through the last decade, not what it is protecting now. For a company raising a Series B in a field this busy, that gap is the thing to hold in mind: the visible portfolio is real, granted in part, and almost certainly incomplete.

The Round

SweGaN, founded in Sweden in 2014, raised $14M in a Series B from Atlantic Bridge, Almi Invest and MediaTek. The company provides GaN-on-SiC epiwafers based on epitaxial growth technology for manufacturers of RF components and devices.

The Portfolio

Seven families since 2015. Four are granted and three are published and pending, and the first of those pending cases could grant at any time. Filings run across EP, TW, US and WO, so the coverage is European, Taiwanese and American with the PCT route used to keep options open. Velocity was front-loaded: two families in 2015, three in 2017, then one in 2019 and one in 2021. Claims cluster in semiconductor devices, in the H10D and H10P groups. The latest published titles point at the same technical center: a heterostructure for a high electron mobility transistor and a method of producing the same, plus two on semiconductor device structures with recessed ohmic contacts. Device structure and the process that makes it, claimed together. Four of the seven families are granted and are enforceable today.

The Company They Keep

The same CPC groups drew TSMC with 4,969 families over the last five years, Samsung with 2,136, Semiconductor Energy Lab with 1,010, Intel with 970, IBM with 827 and Toshiba with 785. That is a crowded space, and in a crowded space count means very little while scope means a great deal. Seven families next to five thousand is not a fair fight on volume, and it does not need to be. What matters is whether the recessed ohmic contact and heterostructure claims read narrowly on a specific epitaxial method or broadly on a class of GaN-on-SiC structures. On the citation side, SweGaN is already cited by Globalwafers with four citing publications, Aixtron Se with four, Huawei with three, Infineon with three, and National Yang Ming Chiao Tung Univ with three. Equipment makers, wafer suppliers and device houses all reaching the same disclosures is a reasonable sign the work sits where the field is building.

The People

Twelve inventors are named across seven families, but Jr-Tai Chen appears on six of the seven. Erik Janzen is on three. That is a concentrated inventive core, which usually means the portfolio reflects one coherent technical program rather than several parallel efforts.

What to watch

Three things. First, grants: the first of the three pending families could issue at any time, which would take the portfolio to five granted out of seven. Second, the publication window: filings made after July 2021 remain unpublished, so expect the shape of the portfolio to change as those surface. Third, the hiring mix. Eighteen open roles break down as five software and AI, five other, four go-to-market, two hardware and two bio and clinical, with sample postings including a SiC PVT Development Engineer and a Senior R&D epi engineer. The filings are hardware. The hiring leans software. Value may be moving toward software that the current published claims do not cover.

Scintillation Research does essentiality, landscape, and infringement research for patent holders.