Socure Patent Portfolio Analysis

One pulse, and what sits under it

A $156M Series E from Accel, Founder Collective and Bain Capital Ventures. That is the event, and it is the only part of this story that made headlines. Socure describes itself as a predictive analytics platform for digital identity verification of consumers, which is a sentence about a market, not about property. The property is 21 patent families, filed since 2014, and the dates on them are the part worth reading.

A single pulse tells you nothing about sequence. Money either follows a technical position or pays for one to be built. Both happen. The difference shows up in filing dates, in who is named as inventor, and in whether anyone else has cited the work. All three are on the record here. Start with the shape of the portfolio itself, then work back to when each piece of it was laid down.

Eleven solid ticks, ten still thin

Twenty-one families. Eleven granted, which means enforceable today. Ten published and pending, and the first of those could grant at any time. Jurisdictions run across Canada, Europe, Hong Kong, Israel and the United States, which is a deliberate spread rather than a domestic default. For a company founded in 2012, twenty-one families is not a thicket. It is a set of positions, and the question becomes where they sit in time.

Two filings in 2014. Then the record picks up again in 2020 with three, one in 2021, three in 2022, three in 2023, and seven in 2024. Two more in 2025 so far. The bulge is recent and it is sharp: a third of everything filed in a single year. The 2014 pair are the founding claim, staked six years before the portfolio began to grow around it.

A third of the portfolio was filed in one year, 2024.

Sixty names, one on fifteen of them

Sixty inventors are named across the portfolio, which is a lot of people for twenty-one families. But Pablo Ysrrael Abreu appears on fifteen of the twenty-one. Ali Haddad is on nine, Feng Xiao on six. The rest thin out fast. That is not a research organisation filing broadly. It is one continuous line of authorship with teams assembling around it and dispersing again, family by family.

This matters for reading the 2024 bulge. A sudden jump in filings can mean a new group hired in to produce them. Here the dominant name spans from the early work to the recent titles, which include stateless account authentication provable by cryptographic biometric key and detection of image injection attacks in mobile document capture using accelerometer vibration analysis. Same hand, harder problems. The volume changed; the direction of the thinking did not.

The small circle in a very large room

The claims cluster in digital communication and business data processing, H04L and G06Q. Over the last five years those same groups drew 995 families from Tencent Shenzhen, 900 from Industrial & Commercial Bank Of China, 865 from Bank Of America, 773 from State Grid China, 762 from China Mobile Communications and 729 from Bank Of China. Against those numbers, twenty-one is a rounding error, and the picture says so honestly.

Crowded classifications are normal for anything touching payments and network identity. What they mean practically is that count buys nothing. Nobody wins H04L. What survives in a space that dense is narrow, specific claim scope: a defined mechanism rather than a description of an outcome. The recent titles read that way, naming an accelerometer, a cryptographic biometric key, synthetic training data for detecting invalid identification documents. Specificity is the only usable position at this scale.

Nobody wins H04L. Scope is the only thing that travels.

Who reads a twenty-one family portfolio

Bank Of America has cited this work in 43 publications. IBM in 37, Wells Fargo Bank Na in 36, Consumerinfo Com in 35, and the Government of the US Secretary of Homeland Security in 27. Bank Of America appears on both lists: one of the largest filers in the same classifications, and one of the heaviest citers of a portfolio a fortieth its size. That is the clearest signal in the file.

Citations from that direction are examiners and drafters marking prior art that sits close to their own path. They do not describe a market position, but they do describe technical relevance, and they come from institutions that would otherwise have no reason to read a twenty-one family portfolio. Eleven of those families are granted. The recent ten are still moving through. The work has been found, by the parties best equipped to find it.

A hundred roles, and where they point

Now the hiring. A hundred open roles, of which 38 are go-to-market and 26 are software or AI. Twenty in operations and finance. The sample titles are an account executive director for gaming and a strategic operations lead for Tier 1 banks. That is a commercialisation phase, and it follows the 2024 filing bulge rather than preceding it. The invention came first; the sales organisation is being built on top of it.

So the plain version. The $156M did not buy a patent position, it arrived after one had been assembled: a founding pair of filings in 2014, a decade of continuity under one inventor, eleven grants live now, and citation traffic from the largest institutions in the field. The portfolio is small by count and narrow by design, which in H04L and G06Q is the only shape that holds.