The In-Q-Tel Fusion Trail

2 A Billion Dollars With No Budget Line

In-Q-Tel, Inc. is a 501(c)(3) non-profit corporation registered under Employer Identification Number 52-2149962. Its financial year ends in March. Its most recent public filing, for the year to March 2025, reports total revenue of $132.1 million, total expenses of $134.1 million, and total assets of $980.6 million. Assets have sat between $928 million and $1.03 billion every year since 2021.

Those numbers come from a tax form. Anyone can pull them in about ninety seconds, and the same form lists what the officers are paid. Steve Bowsher, the president and chief executive who supplied the Kyoto Fusioneering quote, was compensated $2,436,535 in the year to March 2025. Four other executives cleared $800,000.

In-Q-Tel is emphatic about what it is not. Its own website answers the question directly: "IQT is a not-for-profit with 501(c)3 designation in the United States. It's not a government entity nor a part of the CIA." That distinction is real and it is load-bearing, because it's the thing that lets the organisation behave like a venture firm. It also names its customers, and the list is longer than the CIA: NSA, DIA, FBI, NGA, NRO, DHS, US Cyber Command, the Army, the Navy, the US Space Force, and the Department of Energy's National Nuclear Security Administration, plus partner intelligence communities in the United Kingdom and Australia. Hold on to that last one. The agency that runs America's nuclear weapons complex is a named partner of the organisation in section 5, buying into fusion companies.

The cheque sizes are the part that reframes everything else on this page. In-Q-Tel runs two instruments: technology development agreements, typically $500,000 to $3 million, and equity-only investments, typically $250,000 to $500,000. That is small money. A $400,000 equity position doesn't buy control of a company, and it doesn't buy a secret reactor. What it buys is a seat, early sight of the technology, and a signal to other investors that the intelligence community wants this to exist.

So when Forbes says of In-Q-Tel, "that's where the money is, that's where they're hiding the technology," the first half understates it and the second half misdescribes the mechanism. A billion dollars in assets really is where the money is, and it genuinely doesn't appear as a line in any public defence budget. But the way that money touches a company is a quarter-million-dollar stake and a customer introduction, which shapes a roadmap rather than hiding a device. That's the more interesting version of the claim, and it survives contact with the filings.

3 The Directorate That Chartered It, and the Date That Matters

The CIA published its own account of where In-Q-Tel came from, and it names a directorate. Rick E. Yannuzzi, who sat on the In-Q-Tel start-up team, wrote it up for the Agency: "The origins of the concept that has become In-Q-Tel are traceable to Dr. Ruth David, a former CIA Deputy Director for Science and Technology. She and her Deputy, Joanne Isham, were the first senior Agency officials to understand that the information revolution required the CIA to forge new partnerships with the private sector."

So the Science and Technology connection isn't an inference anyone had to make. The idea came out of that directorate, from the officer running it, and it reached the private sector through Director of Central Intelligence George Tenet's Strategic Direction review. That's the institutional address Forbes has been circling for months, and on the origin question he's pointing at the right one. On 6 May 2026 he walked his audience through the CIA's structure and landed on Science and Technology as the branch where any exotic physics work would live.

One correction to his version of the relationship, since precision matters more than the shape of the claim: In-Q-Tel is not an office inside that directorate. It's a separate corporation, and the CIA brokers the relationship through a dedicated unit, the In-Q-Tel Interface Center, which translates what the agencies need into problems In-Q-Tel can shop for.

The date needs correcting, though, and it isn't a pedantic correction. In the 29 April stream the founding is placed after the September 11 attacks: "Inqel started after 9/11, 2001," and then, "2001 is when In-Q-Tel started, and this perfectly aligns with the development of the dense plasma fusion reactors. That paper by George Miley, by Frank [Mead], by David Fro[n]ing. It's from 2005." (The brackets are ours: the auto-captions render those two surnames as "Frank me" and "Frooning". The co-author is Franklin Mead of the Air Force Research Laboratory, and the site's dense plasma focus guide dates that paper to STAIF-2006.) The alignment is the argument. Yannuzzi gives the actual chronology: the leadership decision came in early 1998, the corporation became a legal entity in February 1999, it took its first Agency contract that March, it went public on 29 September 1999 under the name In-Q-It with Gilman Louie as chief executive, and it was renamed In-Q-Tel in January 2000. Every one of those dates is before September 2001, so the founding can't be a response to 9/11 and it doesn't sit in a 2001-to-2005 window with the Miley and Froning work.

Losing that particular coincidence costs the broader case very little, because the coincidence was never carrying much weight. A venture fund founded in 1999 was still funding companies in 2005, and the interesting question was always what it bought rather than what year it opened. Getting the date right just means the page rests on the portfolio instead of on a timeline that doesn't hold.

4 Glenn Gaffney Lists It Himself

Ruth David's directorate supplied the idea in 1998. Twenty years later the traffic was still running, and you can watch one person make the trip in a conference speaker biography. AFCEA, a defence and intelligence trade association, hosts a PDF that introduces Glenn Gaffney under the title "Executive Vice President, Director of IQT Labs." The next sentence reads: "Mr. Gaffney's 31 years of government service include senior positions as the Director of Science and Technology for the Central Intelligence Agency, the Deputy Director of National Intelligence for Collection, and the Associate Director of CIA for Talent."

One man, one page: he ran the CIA's science and technology directorate, then ran the labs at the CIA's venture firm. He held the In-Q-Tel role from 2017 to 2022 and went on to be chief strategy officer at the NobleReach Foundation, which was built around Emerge, a platform spun out of In-Q-Tel. The AFCEA bio itself carries only the titles, so the dates and the NobleReach move are sourced separately below.

No FOIA request produced that. Nobody leaked it. It's a speaker bio, posted so that conference attendees know who is about to talk to them, and it does more evidentiary work than most of what gets argued about in this field. Forbes built a whole stream around Gaffney on 19 January 2026, and the reason that stream stands up is that its central subject has published his own career path.

5 The Fusion Trail

In-Q-Tel sorts its portfolio into eight categories. Thirty-five companies sit under Energy. Five of those have a nuclear or fusion connection, and they're worth naming individually rather than counting. Three of the five are fusion, and between them they cover the two main routes to it: magnetic confinement, laser inertial confinement, and the fuel cycle that either one would need.

Kyoto Fusioneering is the solid one. Its own release names In-Q-Tel as a lead, quotes In-Q-Tel's CEO, and dates the close to July 2024. The company doesn't build a reactor. It builds the parts that any deuterium-tritium plant needs and almost nobody makes: gyrotrons, which are high-power microwave tubes used to heat plasma; tritium fuel cycle systems; and breeding blankets, the lithium-bearing wall that regenerates tritium from the neutrons the reaction throws off. Its UNITY-1 and UNITY-2 facilities test that hardware.

That is a picks-and-shovels investment, and it tells you something about intent. Backing the only serious supplier of tritium-handling equipment gives you visibility across every fusion programme that eventually needs to buy it. Bowsher's own stated reason is commercial: "IQT continues to look for investment opportunities in the power and energy domain as global efforts to achieve sustainable energy intensifies." Take that at face value and it's still an intelligence organisation buying a seat in the supply chain of a technology it expects to matter.

Tokamak Energy is where the record thins out, and this page is going to be blunt about that rather than paper over it. In-Q-Tel lists the company in its portfolio, under Energy, marked active, linked to tokamakenergy.co.uk. The company has never announced it. Its November 2024 round of $125 million names East X Ventures and Lingotto as co-leads and lists Furukawa Electric, British Patient Capital, BW Group and Sabanci Climate Ventures as new investors, with no mention of In-Q-Tel. The funding databases that track its cap table don't carry In-Q-Tel either.

The obvious explanation doesn't survive contact with In-Q-Tel's own FAQ. You'd expect the answer to be that a technology development agreement buys work rather than shares, leaving a portfolio entry and no cap-table line. In-Q-Tel says otherwise: it has "two primary tools for building a relationship: a technology development agreement or an equity-only investment, both of which where IQT receives a small equity stake and status as a board observer." Both routes end in equity. That makes the silence around Tokamak Energy harder to explain, not easier.

So the honest statement of the Tokamak Energy relationship is a short one. In-Q-Tel says there is one, the company has never said what it is, and the date, the size and the instrument are all outside the public record. It's the weakest item on this page and it's flagged as such in the assessment table.

Inertia is the one that most surprises people who assume the intelligence community only backs the safe option. It's a laser inertial-confinement company, built on the December 2022 ignition shot at the National Ignition Facility, and its plan is a modular array of a thousand lasers firing at 4.5 millimetre fuel targets ten times a second, with the targets mass-produced for under a dollar each. Nobody announced this investment either. It sits in the portfolio and nowhere else.

What Tokamak Energy actually builds bears on the next section. The ST40 is a spherical tokamak, a squashed variant of the doughnut that reaches a higher plasma beta, meaning more plasma pressure for a given magnetic field. That's magnetic confinement physics, published in the open literature, with a company that spun out of the UK Atomic Energy Authority in 2009. Nothing about it is exotic.

6 Finding Your Way When the GPS Dies

Forty-nine companies sit under Microelectronics and Quantum, which makes it the biggest technical category after AI, hardware, enterprise software and cyber. Read the names and a specific appetite comes through: Infleqtion, Q-CTRL, SandboxAQ, Xanadu, IonQ, Rigetti, Quantum Brilliance, Universal Quantum, Diraq, QuantWare, Maybell Quantum, Quantum Circuits.

Some of that is quantum computing, which is a different bet. The thread that matters here is quantum sensing, and Infleqtion is the clearest case. Formerly ColdQuanta, it builds cold-atom inertial sensors, gravimeters and optical clocks. The application is navigation when satellite positioning is unavailable or being jammed. In October 2025 its Tiqker optical clock flew on a Royal Navy autonomous underwater vehicle, which the company describes as the first quantum optical clock deployed that way. In April 2026 it took an $11 million Department of Defense APFIT award to push quantum timing into service. It completed a merger with Churchill Capital Corp X in February 2026 and now trades on the NYSE, and the portfolio marks it exited.

Why does that belong on a 4Orbs page? Because it's the same instrument family the site keeps running into from the other direction. Atomic magnetometers and gravimeters are the technology the Ghost Murmur claim invokes, and the Iran rescue reporting put a quantum magnetometer at the centre of a public argument about what the CIA can detect. What In-Q-Tel's portfolio adds is the funding record underneath: the intelligence community has been buying into precision atomic sensing for years, in the open, at commercial scale.

That doesn't validate the heartbeat-detection claim, which rests on one newspaper report and gets weighed on its own page. It does mean the underlying capability class isn't speculative. Somebody with an intelligence customer has been paying for it since before most people had heard the phrase.

7 The Propulsion Holdings, Described Accurately

Thirty-five companies sit under Space, and three of them build propulsion: Morpheus Space, which makes electric thrusters for satellite manoeuvring; Digital Solid State Propulsion, which builds electrically controlled solid propellants; and Impulse Space, which builds orbital transfer vehicles. Alongside them sit the launch and servicing names, ABL Space Systems, Rocket Lab, Stoke Space, Starfish Space, ClearSpace, D-Orbit.

This is conventional spaceflight. Electric propulsion covers ion, Hall-effect and field-emission thrusters, which is what Morpheus Space builds. All of it is mature technology that moves satellites slowly and efficiently on solar power. It has nothing to do with field propulsion or mass reduction, and describing it as though it did would be the kind of overreach this site exists to avoid.

Say it plainly: there is no exotic propulsion company in In-Q-Tel's published portfolio. That's a check run across all 459 names rather than just the Space category, which also catches Hermeus, a hypersonics company filed under Hardware whose engine is a turbine-based combined cycle. Conventional too. What's there is the industrial base for putting sensors in orbit and keeping them pointed.

8 The Companies That Aren't There

Here's the finding. Every compact-fusion company whose physics the orb thesis actually depends on is absent from In-Q-Tel's list of 459.

Searched by name against the full snapshot: Helion Energy, absent. TAE Technologies, absent. General Fusion, absent. Commonwealth Fusion Systems, absent. Zap Energy, absent. The field-reversed configuration is the confinement geometry this site has spent years documenting, because a self-organised plasmoid that holds its own shape is the closest published physics to what the MH370 footage appears to show. In-Q-Tel's fusion money went to a spherical tokamak, a laser inertial-confinement startup and a tritium fuel-cycle supplier instead. Two confinement routes and the plumbing, and not one plasmoid.

Now the caveat, and it comes from In-Q-Tel rather than from us. The portfolio page carries a sentence directly under the counter: "800+ Total Investments. The following is a snapshot of our investments." Be careful with that arithmetic, because the two numbers count different things: 800+ investments against 459 companies, and a venture firm books a follow-on round into a company it already owns as another investment. The gap could be nothing but second and third cheques into names already on the page. It could also be companies left off it. In-Q-Tel says only that the list is a snapshot, and doesn't say which of those two the difference is. Absence from a list its owner calls partial is weak evidence.

Weak evidence still has a direction, though, and this is where fence-sitting would be dishonest. Weigh what's in front of us. The one fusion investment that is fully corroborated, by a company press release with In-Q-Tel's CEO quoted in it, is in mainstream deuterium-tritium infrastructure. The stated rationale is commercial energy. The propulsion holdings are satellite thrusters. The sensing holdings are precision atomic instruments sold to militaries in public. Four separate strands, and every one of them points at an organisation buying published technology early rather than one banking exotic physics.

The lean this page lands on: on the open record, In-Q-Tel is not where the FRC trail runs. That's a specific and limited claim. It doesn't say the CIA has no interest in compact fusion, and it doesn't say that whatever the snapshot leaves out is innocent. It says the instruction to go and look at In-Q-Tel's companies was a good instruction, we followed it, and the companies it leads to are the ones you can already read about in Nature. If the exotic work is being funded, the public half of the portfolio isn't the route to it, and anyone who says otherwise now has 459 names to argue with.

9 What the Track Record Actually Predicts

In February 2003, In-Q-Tel invested in Keyhole, Inc., a small company with a product called EarthViewer that streamed satellite imagery over a 3D globe. It announced the deal that June, and said in the announcement that this was its "first engagement with a company on behalf of the National Imagery and Mapping Agency", the outfit now called the National Geospatial-Intelligence Agency. Google bought Keyhole on 27 October 2004 and renamed the product Google Earth.

Palantir is the other name everyone cites, and it's in the portfolio too, marked exited. Between them they explain why press coverage of In-Q-Tel almost always reaches for the same phrase about the fund that helped launch Palantir and Google Earth.

Forbes draws a line from that history to the Gorgon Stare footage: "You want to know why the Gorgon Stare video looks like Google Earth? Because the same people produced it." The literal reading of that sentence is too strong, since Gorgon Stare is a Sierra Nevada and BAE Systems airborne sensor and Keyhole was a startup Google absorbed. The underlying observation is sound and better than the phrasing suggests. Consumer geospatial visualisation and intelligence geospatial visualisation share an ancestor, share a customer in the NGA, and share design conventions for the same reason any two products from one lineage do.

Which points at the honest conclusion about this organisation. In-Q-Tel's demonstrated competence is sensing, geospatial and analytics. That's where its famous wins are, that's where the bulk of the 459 sit, and it's a reasonable predictor of where the next win comes from. For a site whose central evidence is satellite footage of objects nobody can explain, the interesting In-Q-Tel story was never a hidden reactor. It's that the pipeline which built the world's imagery infrastructure was capitalised by the same agency that later declined to say what its cameras saw.

10 The Transfer That Hasn't Happened

In March 2026, Politico reported that Director of National Intelligence Tulsi Gabbard was finalising a plan to move In-Q-Tel out from under the CIA and into her own office. Forbes covered it across two streams and read it as done: "DNI is taking over Inqel" on 29 April, and on 6 May, "they just took this ... weapon away from that exact branch of the CIA."

The record doesn't support the past tense yet. Follow-up coverage on 27 March 2026 reported the proposal as contested and unannounced, with supporters arguing that ODNI oversight better reflects In-Q-Tel's community-wide role and opponents warning about added coordination layers. More to the point, ODNI spokesperson Olivia Coleman went on the record against the story itself: "This is false and I told Politico that and their response was that they'd update the story with my comment."

As of 15 August 2026 no transfer has been announced, and searches for any update since March return nothing. So the status is one paywalled report, an on-the-record denial, and five months of silence. That's genuinely worth watching, because if it does happen it moves a billion-dollar technology-buying operation from an agency to the office that coordinates all of them. It just hasn't happened.

11 What Would Settle It

Four things would move this page off its current position, and all four are obtainable by someone willing to do the paperwork.

1. Investments against companies. In-Q-Tel claims 800+ investments and names 459 companies. Establishing whether that gap is follow-on rounds into names already listed, or separate companies withheld, would decide how much the absence in section 8 is worth. The Form 990 Schedule D and the audited financial statements are the place to look.
2. The Tokamak Energy shareholding. In-Q-Tel says both its instruments leave it holding a small equity stake, so a UK Companies House filing for Tokamak Energy Ltd should show it on the share register whichever route was used. Whether it does is the single cheapest check on this page, and it settles section 5 either way.
3. Dates on the uncorroborated holdings. Radiant and General Matter appear with no announcement from either side. SEC Form D filings for their raises would at least bound when In-Q-Tel came in.
4. The transfer. An ODNI or CIA statement either way ends section 10.

None of that requires a whistleblower. It requires a database search and a filing fee, which is the whole argument for treating In-Q-Tel as an open-source problem instead of a mystery. The organisation publishes its portfolio, files its taxes, and quotes its own CEO in other companies' press releases. Most of what people assume is hidden about the CIA's venture arm is sitting in public documents that nobody has bothered to read end to end.

Until then, the trail leads where it leads. The CIA's venture firm funds fusion, and it funded the fusion everyone already knows about.