Devialet Patent Portfolio Analysis

One pulse, seventeen years after the first filing

September 2022. Fifty point seven million dollars, announced in a single round. On a timeline it is one clean spike, and spikes invite a familiar reading: fresh capital, fresh engineering, a portfolio that starts here. That reading is wrong in this case, and the filing record says so without argument. Devialet's first patent family carries a first filing date of 25 July 2005. The company had been filing for seventeen years before this pulse.

That matters because the usual sequence runs the other way. Money comes in, headcount follows, and the first filings appear eighteen months later with the ink still wet. Here the sequence is inverted. The invention record is the older document. The round is the newer one. Everything that follows in this piece is an attempt to read the older document properly, starting with how much of it survived examination.

One hundred and four against nine

One hundred and thirteen families in total. One hundred and four of them granted. Nine published and still pending. That is a ninety-two percent grant rate, and the two ribbons are wildly unequal in a way that tells you which end of the lifecycle this company sits at. A portfolio built in the last three years cannot look like this. Examination takes years, and grant counts lag filings by a wide margin everywhere.

The small pending group is the interesting part precisely because it is small. It means the record is not a pile of hopeful applications waiting for a first office action. It is a body of work that has already been argued with examiners, narrowed, and allowed. Whatever else the 2022 round funded, it did not fund the building of this. The building was done and mostly certified.

Ninety-two percent granted is not a young company's number.

The busiest year was 2018, not 2023

The bars tell a story in three movements. A single family in 2005, then nothing until a cluster of six in 2010. A first surge across 2013 and 2014, fifteen families in 2014 alone. Then the real climb: ten in 2017, nineteen in 2018, the highest year on record. Three in 2019, then fifteen in 2021 and thirteen in 2022. The rhythm is of a team that files in waves rather than continuously.

The funding marker lands in 2022, well to the right of the 2018 peak. The four families in 2023 and four in 2024 sit inside the shaded region, where publication delay means the count is provisional and will rise. Read honestly, this chart shows invention running hard for a decade before the round, with the post-round years still incomplete rather than conclusively thin.

One hundred and eighty-four citations from a single source

SONOS INC cites these filings 184 times. That is not a passing reference; it is a sustained pattern of one organisation's examiners and drafters repeatedly landing on the same French record. Behind it: STEELCASE INC at 26, SAMSUNG ELECTRONICS CO LTD at 16, GOOGLE LLC at 13, HARMAN INT IND at 9. Five names, one of them accounting for more citations than the other four combined by a factor of three.

Citation is a crude signal and worth stating plainly for what it is. It shows that the prior art in this space routes through Devialet's filings, and that large, well-resourced patent operations keep encountering them. It says nothing about products or markets. But a portfolio nobody cites and a portfolio cited 184 times by one organisation are different objects, and only one of them is load-bearing.

Prior art in this space routes through a French filing record.

Fourteen offices including Denmark and the Philippines

US, EP, CN, JP, KR, DE, FR, CA, AU, BR, TW, DK, PH, ZA. Fourteen national and regional offices, with PCT set aside. The first eight are the standard route for anyone with export ambition. The last six are the ones worth pausing on. Brazil, Taiwan, Denmark, the Philippines and South Africa are deliberate choices, each one a separate translation, a separate agent, a separate annuity bill for twenty years.

Nobody files in the Philippines by accident or by template. Each of those boxes represents a decision taken years before the money arrived, by a company that was choosing where a right would be worth paying to keep. Combined with the grant rate, the picture is of long-horizon prosecution discipline rather than opportunistic coverage. That discipline is the asset. The 2022 round is what a company raises when it already owns one.

What the cheque could not have bought

A seventeen-year filing history. One hundred and four grants. Fourteen offices. One hundred and twenty named inventors, with CALMEL PIERRE-EMMANUEL on 36 families and NARDIN EMMANUEL on 22. None of that can be purchased in 2022 and delivered by 2024. Examination, opposition windows and annuity decisions run on their own clock, and that clock started in 2005. The round bought time and scale. It could not have bought priority.

Which leaves the question that the whole record has been building toward. If the invention peaked in 2018 and the capital arrived in 2022, the capital was not financing discovery. It was financing whatever comes after discovery. The filings that matter for the next decade are the nine still pending and the eight from 2023 and 2024 that the shading has not yet released.