Genomatica Patent Portfolio Analysis

One pulse, and everything behind it

The $118 million Series C was announced on 1 July 2021. That is the single event on screen, and it is the most recent thing in this story by a wide margin. The record underneath it starts on 17 January 2005, when Genomatica filed for the first time. Sixteen years and six months separate those two dates. Almost everything the company has ever committed to paper sits in the gap.

A raise is a moment. A portfolio is an accumulation. The interesting question with Genomatica is not what the money bought, because the filings show that most of the buying had already happened. By mid-2021 there were 184 families with a first filing date of 2020 or earlier. The capital landed on top of a body of work that had already been built, paid for, and in many cases already granted.

Two hundred and twenty-six ticks, mostly solid

226 families. 125 of them granted, 101 still published and pending. That is a 55 percent grant rate, and it means the solid marks outnumber the thin ones. For a portfolio whose oldest filing dates from January 2005, that ratio is what you would expect from a house that has been prosecuting patiently for two decades rather than filing in a hurry.

The thin ticks matter too. The most recent published filing is dated 8 October 2025, which tells you the pending half is not just residue from old campaigns. Recent titles include carboxylic acid reductase and alcohol dehydrogenase variants and methods of use, desaturase variants and methods of use, and a method for separating hexamethylene diamine from an aqueous solution obtained from a microbial fermentation process.

Most of the marks on this arc were made before anyone wrote the cheque.

The peak sits eight years before the cheque

The bars climb early. Two families in 2005, three in 2006, three in 2007, then seven, then seventeen in 2009. A dip, then the real surge: 27 in 2012 and 37 in 2013, the peak year. Then the decline is fast and deep. Eighteen in 2014, ten in 2015, seven in 2016, two in 2017. By the time the 2021 marker appears on the axis, the company had already filed and abandoned its most productive posture.

What happens after the marker is not a second peak. It is a steady floor: ten families in 2020, ten in 2021, twelve in 2022, eleven in 2023, seven in 2024. The shaded window at the right covers filings that may exist but have not published yet, so the last two years are provisional. Read the shape and the funding year is not the inflection point. 2013 was.

The citations come back to four names

The nodes on this diagram are unusual. The entities citing these filings most heavily are not corporate names. They are individuals: Anthony P Burgard with 92 citations, Mark J Burk with 69, Robin E Osterhout with 61, Priti Pharkya with 57. Three of those four are also the most frequently named inventors on the portfolio itself. Burgard appears on 86 families, Osterhout on 76, Pharkya on 69, out of 226 total.

So the densest citation traffic into this record runs back to the people who wrote it. That is what a long-tenured technical core looks like in the public file: the same names generating prior art and then building on it, across filings separated by a decade. 468 inventors are named across the portfolio in total. A handful of them account for the through-line.

Twenty-eight offices, and some unexpected ones

Excluding PCT, the families appear before 28 patent offices. The expected set is there: US, EP, CN, JP, KR, IN, CA, AU, BR. So is a second tier that costs real money to maintain and is rarely chosen casually: TW, SG, MX, ES, IT, AR, CO, DK, MY, NZ, PH, TR, ZA. And then the outliers, Hungary, Lithuania, Poland, Slovenia, Morocco, and the Eurasian office.

National filings in Lithuania and Slovenia are a decision, not a default. Somebody paid translation and prosecution costs in each of those jurisdictions and has kept paying renewals. Set that against the timeline and the picture sharpens: a company that filed hardest in 2012 and 2013 also chose to carry that work into places most portfolios never reach.

Geographic breadth is the most expensive form of conviction a patent file can record.

What the file says the company believed

Here is the plain version. Genomatica's patent record is the record of a bet placed between 2008 and 2014, prosecuted steadily since, and extended into 28 offices at real cost. The 2021 raise did not start the filing programme and did not restart it. It arrived when the portfolio was already 184 families deep and more than half of what would eventually be granted was already in prosecution.

The second thing the file says is about people. Three inventors carry a third or more of the families each, and the heaviest citing names are the same names. That is continuity, and continuity is difficult to acquire. Whatever else the money bought, it did not buy those two decades of accumulated filing dates, because they were already there.