Disclaimer: this is all personal opinion. I don’t represent the views of my employer, Azavea. I disclosed publicly, and regularly, that I have also worked directly with CosmiQ Works, a research lab under the In-Q-Tel umbrella, but that is the extent of my personal experience with the organization.
It’s hard to imagine today, but in 2002, technology startups were decidedly not cool. That year, the NASDAQ-100 index had bottomed-out at 78% below its March, 2000 peak. The infamous “dot-com bubble” had burst, and the venture capital industry was soaked. At that time, no self-respecting investor would have been caught dead speculating on internet startups; hadn’t public markets just illustrated their worthlessness? No one would feel sorry for the incorrigible VC firms who hadn’t learned their lesson and were already back to spraying money at cash-guzzling internet hucksters.
The internet was “out.” Pants with words like JUICY embroidered across the butt were “in.”
In November of that year, just one month after the NASDAQ-100 reached rock bottom, a young entrepreneur by the name of John Hanke pitched investors on an exciting new company called Keyhole, Inc. The recent orphan of a failed video game startup, John and his small team painted a fantastical picture: the web would enable global-scale, interactive, 3D mapping for the masses and recycled video game graphics would power it. I imagine there were more than a few skeptics. After all, who would pay for this fancy 3D globe thing? What would they use it for? How would Keyhole acquire global datasets to populate the map, and how many bajillions of dollars would that cost? It’s hard to imagine a less attractive pitch at that moment in history…the days of investing in a demo and a prayer were all but over.¹
The one institutional investor who stepped up to help was a little-known venture capital firm, themselves only a few years old, called In-Q-Tel (IQT). They funded Keyhole alongside Sony for a whopping $527,000 according to Crunchbase. Keyhole would go on to be acquired shortly thereafter in 2004 by a buzzy search startup, and would launch their flagship product soon after that in 2005: Google Earth. John would remain at Google and lead their geospatial group with a few other mildly successful product launches including Google Maps.² IQT became partial owners of Google through the acquisition, though they promptly sold their shares worth about $2.2 million at the time.³ That’s something in the ballpark of a 10x return over two years…not bad. Maybe this “internet” thing still had legs.
Who the Hell is In-Q-Tel?
The firm’s official history dates back to September, 1999 when it was first introduced to the world as “In-Q-It.”⁴ The terrible name lasted only a few months before being superseded by the equally terrible name, “In-Q-Tel.” I like to imagine spirited debates about the early choice to rebrand ultimately resulting in an unholy compromise agreed to by all and enjoyed by none. Maybe “In-Q-It” just sounded too much like “Intuit,” and they were worried about getting sued. Or maybe the CIA felt there wasn’t enough emphasis put on the in-tel-ligence part of the firm’s mission. I‘m deeply curious about the story behind the (re)naming, so please comment below or write to me about it if you have the receipts.
IQT is easy to caricature. Typical stories about them emphasize their affiliation with the CIA in order to get clicks.⁵ I’m not going to spend too much time rehashing their origins or their affiliation with the U.S. Intelligence Community. I think a fair summary is that they’re a government-funded investment firm measured on their ability to back companies who have created noteworthy technology relevant to national defense interests and subsequently to encourage adoption of that technology across the federal government. Curiously, they’re incorporated as a non-profit, which subjects them to fairly rigorous oversight and means they have to file a summary of their finances annually. The last complete filing, from 2018, showed their net assets at almost $450M with over $116M in fresh money between investment income and federal grant revenue (the latter far exceeding the former). 990 filings are fun, because you can see how much executives are compensated and IQT is no different…Mom always warned me that the non-profit world was a tough way to make a living, but I think she was just working at the wrong non-profits.
A non-profit venture capital firm…they are the first I’ve ever come across. Those words don’t go together — it’s quite nearly an oxymoron. But rather than think of it like a traditional VC firm, I think of it more like DARPA with a side of equity. I actually love the model, at least intellectually, as a way of aligning incentives with the companies IQT chooses to back. If you’ve ever competed for federal grant funding, you’ve probably thought to yourself, “I would like some of that free money, please.” After all, there are few phrases as sexy as non-dilutive capital.⁶ But in reality, most of the agencies that give out grants to commercial entities for commercial purposes, like through the SBIR and STTR programs, can hardly be bothered to raise a finger to help you actually commercialize the work they’ve funded (not even within their own agency, let alone across the whole of government). As long as you turn in the reports on time, they’re happy. It’s a box-checking exercise, at least from the grantor’s perspective. With IQT’s model, the company gets a genuine partner — literally an equity partner — who is a government insider directly measured on their ability to get you paying customers. And the company has real skin in the game (as anyone who has given a chunk of ownership over to an investor well knows). I’m pretty sure there is a cottage industry of thinly-veiled government SBIR/SBTTR moochers who have figured out a way to eek out a living on “R&D investment” that they book as revenue with no real customers to speak of, although I’ve never seen anything written about this phenomenon and don’t have a ton of evidence to back up the claim (besides years of looking up awardees after announcements). I’m pretty sure it’s possible game the grants system. But I’m positive there is no way to game IQT, at least not indefinitely.
The Sultans of Spatial
Mapping is a strange discipline —it reminds me of a hotdog place I used to eat at all the time in Charlotte, North Carolina called Green’s Lunch. All walks of life enjoyed its fine greasy fare: tattooed tree huggers in socks and Birkenstocks, safety-orange-clad construction workers, pant-suit-rocking bankers, politicians in cheap suits that always seemed several sizes too large, and me. Green’s Lunch is a decent analogy for the field of “GIS.” Both are what Kurt Vonnegut would call a granfalloon: a jumbled-up assortment of otherwise unaffiliated people tied together by one common circumstance.
I’ve only ever been able to find a single thread that ties the wide-ranging field of mapping technology together, and that is: military subsidy. The defense industry is like the kitchen in the back of Green’s Lunch — virtually unseen but integral to the whole operation. You think chili cheese dogs make themselves?
Whether you like to admit it or not, almost everything about modern mapping — from the GPS constellation to the internet itself — has been aided by defense-related R&D expenditure. Many of the most beloved open source geospatial software packages, from QGIS, to GeoServer, to PostGIS and beyond have been subsidized by DoD dollars. There are purists who feel all military work is unethical and that any defense dollars are tainted, money no matter the context. I used to take this stance. But, at some point I came to the conclusion that without defense, there is no GIS/mapping field (at least not at the scale and importance that it enjoys today). IQT is just one small example of how the U.S. military and intelligence machine exerts influence over the most prominent GIS/mapping companies in existence today.
At least they aren’t bashful or secretive about their mission— on the front page of their website, they proudly advertise:
IQT accelerates the introduction of cutting-edge technologies that protect and preserve national security.
Their geospatial track record is incredible. Keyhole turned out to be a sign of things to come. They’re also investors in Mapbox, Orbital Insight, OmniSci, and Cape Analytics which have collectively raised over $440M in recent years.⁷ They were prominent investors in Boundless Spatial, the purveyor of an open source suite of GIS tools that was ultimately acquired by Planet in early 2019. They’re also investors in Palantir, which keeps threatening to go public but never seems to and whose Gotham software prominently features mapping features.
And there are plenty of lesser-known investments in the space that have spawned a network of serial founders and prominent geospatial technologists. GeoIQ is a former portfolio company that wound up getting acquired by Esri in the summer of 2012 and whose founder, Sean Gorman, went on to found another successful startup backed by IQT, Timbr.io, which was quickly snapped up by Maxar (née DigitalGlobe).⁸ Other bets include Geosemble, a short-lived software firm that was acquired by TerraGo (itself an IQT portfolio company) and ImageTree, a forestry-focused remote sensing software company. IkeGPS raised money from IQT all the way back in 2012 and is still around today making mobile tools for field inventorying and surveying. CyPhy Works was a high-profile recipient of IQT funds that made compact drones and related software and turned out to be one of the more dramatic failures within the commercial drone industry after raising $39M in total. The list of geospatially-oriented startups is longer than I have time to cover here, but there is no doubt IQT has stuck to a mappy thesis consistently for more than two decades.⁹
More Than a Check Book
Perhaps what I find most interesting about IQT is its commitment to advancing the state of the art in the various strategic technological fields it invests in (above and beyond simply throwing money at startups). The firm also spins out research groups, known collectively under the umbrella of “IQT Labs.” So far there are four: Lab41 (machine learning), Cyber Reboot (cyber defense), B.Next (biology), and CosmiQ Works (commercial space). These groups publish original research, write open source software libraries, and fund research competitions. They are active participants in the same markets IQT hopes to invest in. I’m personally a huge fan of CosmiQ Works, whose blog is one of the most underrated sources of technical insight I’ve come across and an easy way to forecast where mapping technology is heading in the near term.¹⁰
I’m just as squeamish as the next person about the militarization of the United States. I don’t think history will look back kindly on drone strikes and foreign nation building campaigns, nor will it wax eloquent about the flimsy justifications for ICE and domestic surveillance programs. But I do think IQT is an example of the good that the military and intelligence communities can do — who else in the government has the budget or purview to throw tens of millions of dollars a year at early-stage, unproven companies, many of whom give their technology away for free? I don’t have any data to back this hunch up, but I would bet that the median outcome of IQT’s portfolio vs. the median outcome of any government-backed commercial research grant program is not even close, both in terms of societal impact and eventual return on tax dollars invested.
Who Else Comes Close?
Does any other VC challenge IQT’s dominance when it comes to geospatial company building? That’s not a rhetorical question — lay some knowledge on me if you know better. In recent years, I’ve closely followed DCVC and Lux Capital who have quickly amassed an impressive (and substantially overlapping) portfolio leaning heavily toward space-and-drone startups including Airmap, Cape Analytics, Capella Space, Descartes Labs, DroneDeploy, Hangar, Orbital Insight, and SailDrone. DFJ holds positions in Planet and Mapbox, and was one of the catalysts for the “new space” race that defined the majority of VC investment in the industry over the past decade. But no one that I know of matches IQT’s longevity and consistency in expressing a thesis around the value of mapping.
¹ Today, you don’t even need the demo part. It seems to me that a “seed round” can customarily fetch $2M of funding on a $8M “pre-money” valuation for a pair of well-pedigreed founders even if they haven’t landed a single customer yet. Someday, I’ll write a whole blog post about the benefits of raising less money, but for now I’ll just point out that Keyhole’s start was quite a bit slower and they did alright. For a wonderful profile on the history and importance of Keyhole, I recommend this story: https://techcrunch.com/2019/03/29/how-a-google-side-project-evolved-into-a-4b-company/. Thanks also to Kieth Masback for sharing this history of Keyhole, which I hadn’t seen before: https://trajectorymagazine.com/genesis-google-earth/.
² He’s now the CEO of Niantic, Inc. — makers of the modestly popular game Pokémon Go. I guess some people are just destined to live quiet, unassuming lives.
³ Had they held on, those shares would be worth north of $17 million today. But IQT isn’t like other institutional investors — they aren’t measured on how much money they can return to their funds as far as I can tell. No one is crying over spilt milk on the Keyhole deal — in fact, they’re probably thrilled with how it turned out. To understand why, you have to understand what IQT is.
⁴ There’s a legitimate history of the firm available from the CIA itself available here. It’s fascinating and gives credit where it’s due to Dr. Ruth David, the visionary inside of the CIA who saw the role the U.S. Intelligence Community could play in advancing technologies relevant to national security.
⁶ Romance has never come naturally to me.
⁷ Oddly, I can’t find anything on IQT’s site suggesting they invested in Mapbox. There’s an exposé-style report that claims they have, and a business intelligence site that lists IQT as a funder…but I’m not 100% positive about the accuracy of these reports. It seems maybe not all of IQT’s portfolio is listed on their portfolio page, which is usually something VCs do when startups fail and they want to wipe them off the pages of history. But Mapbox is one of the most successful and prominent mapping startups of all time, so it’s a little baffling to me. Maybe it’s just a clerical error. There are a ton of companies in that portfolio, after all.
⁸ Sean is now back in the game with an augmented reality mapping startup, Pixel8. Some people just can’t stop. He’s a good follow on Twitter if you’re into that kinda thing… https://twitter.com/SeanGorman
⁹ In a related longterm commitment to a thesis that has panned out even better than the mapping one, IQT seems to have an excellent eye for commercial open source companies. IQT is somehow invested in many of the most successful “open core” companies of all time including Anaconda, Cloudera, Databricks, Gitlab, Mesosphere (now D2iQ), and MongoDB, who are collectively valued at over $25 billion today. Traditional VC’s willingness to pour resources into openly licensed software is spotty at best; there is only one RedHat, after all. But IQT, with its emphasis on government adoption rather than traditional cash returns, is able to see the value of hybrid business models that grow the overall market just as fast as they grow their paying customer base.
¹⁰ Full disclosure: in my day job, I got to work on a project with them to make some of their open data available on our labeling tool, GroundWork, and I also appeared on their podcast once. So take my praise of them with a grain of salt (or just trust me — they’re awesome).