Stranded IP: Why a Carve-Out Isn't a Transaction

Eleven spin-offs, together worth twice as much as the company that let them go. The company is Xerox. Between 1979 and 1998, 35 projects leave its research centres, above all the famous PARC, after internal funding runs out. Henry Chesbrough later traced what became of them. 24 came to little. The other eleven ended up with a combined market value of twice that of Xerox itself.

In many corporate labs today it looks no different. A technology sits at readiness level three to four, the results are good, the patents filed. The internal review concludes: it fits no business unit. "Strategically interesting, but not our core business." The budget gets reallocated, the lead scientist leaves. Five years later the M&A team presents the board with a startup working in exactly that field. The valuation is ambitious.

Same technology. Once in the drawer, once outside. In the end, the difference is paid by whoever left it lying there.

The numbers from the drawer

That companies sit on unused technology isn't pub wisdom. It's well measured. The PatVal survey of thousands of European inventors, published in Research Policy, finds that around 36 percent of European patents are used neither industrially nor commercially. At large companies it's around 40 percent, at small ones only 18. The bigger the house, the fuller the drawer.

To be honest, part of the picture is this: companies hold roughly half of these unused patents deliberately, to block competitors. The rest is simply asleep. A US study by Walsh, Lee and Jung shows the same pattern from another angle: only 55 percent of the patents it examined are commercialised, and non-use is more often failure than strategy.

At universities and research institutes the picture is starker still. According to the European Patent Office, only about a third of their patented inventions are commercially exploited. For a fifth there isn't even a plan.

All of this against a backdrop of enormous investment. The 322 EU-based companies among the world's 2,000 biggest R&D investors alone put €234 billion into research and development in 2023.

You pay for the research. You pay for the invention. You pay for the patent, and year after year for its maintenance. And then you put it in a drawer.

Why good technology gets stranded

The first reflex is to assume waste or stupidity. Both are rarely the case. Technology gets stranded out of logic, not negligence.

It's created in an environment optimised for scientific excellence, not commercial scale. It's often excellent. But it doesn't fit the core strategy. It would need its own market, its own team, its own sales and its own appetite for risk. And no one in the building has the mandate, the time and the incentive for that all at once.

The Draghi report offers the deeper explanation: European companies are specialised in mature technologies where the potential for breakthroughs is limited. Draghi calls it the "middle technology trap". And he states the problem without hedging: Europe doesn't lack ideas; it fails to translate innovation into commercialisation.

Add to that a simple asymmetry. Inside the corporate, the cost of trying is visible and the upside is uncertain. And the people best placed to take a technology to market rarely want to do it inside a corporate department. They want a venture, founder equity and a clean stage.

That's the definition of stranded IP: valuable technology with no organisational path to market. Too good to throw away. Too foreign to keep.

The misconception: the carve-out as a transaction

When corporates talk about carve-outs, they mostly mean a deal. A unit gets split off, sold, taken public. Balance sheet cleaned up, focus sharpened, press release out.

The numbers on that model are sobering. In 2026 McKinsey examined large spin-offs: in 55 percent of cases, parent and spin-off together show a negative shareholder return three years after the separation. Bain analysed private-equity carve-out returns: before 2012 the average multiple was around three times invested capital; since then it has been 1.5.

That's not an argument against carve-outs. It's an argument against carve-outs as transactions. Bain draws the conclusion itself: value no longer comes from financial engineering, it comes from building growth. Re-hang a unit and you shift costs. Build a company and you create value.

The insight: a carve-out is a company build

Stranded technology doesn't need a new label. It needs a vessel: an independent company with its own team, its own market and its own capital. And that company has to be built first.

It follows a clear order. First, assess the maturity of the IP and the technology. Then map the commercial path: which first market, what willingness to pay, what minimum viable technical capability. Then find the founding team. Then design the licensing, ownership and incentive structure. And only then incorporate the company, transfer or license the IP, fund it and launch.

The decisive thing about this construction: the owner doesn't have to give up everything to gain something. The technology source keeps a minority stake, upside without operational burden. External investors fund the growth, not your own balance sheet. A book value that only generates fees becomes a share in a growing company.

The structure matters for tax too. In Germany, a straight IP sale is taxed in full. With a licence-plus-equity construction, a later exit through the shares is largely tax-exempt for corporations under § 8b of the German Corporate Income Tax Act. Which structure fits a given case belongs on a tax advisor's desk. That the structure matters is beyond doubt.

Build the M&A target you'd otherwise have to buy back

Now the part boards rarely talk about out loud. The technology you don't commercialise today doesn't disappear. Someone else builds it. And maybe you buy it back later, at a multiple.

Cisco turned this into a model and showed what it costs along the way. Cisco funded the networking specialist Nuova, held 80 percent and secured a call option on the rest. For the remaining 20 percent, up to $678 million eventually came due. At Insieme, Cisco held around 84 percent and a put/call option; the other shareholders were eligible for up to $863 million in milestone payments.

That isn't cheap. But those were terms Cisco defined itself, not the price an open-market bidding war dictates. That's the point: a carve-out today means a contractually agreed way back tomorrow. Licence, partnership or acquisition, on terms you know.

A guardrail: not every patent is a venture

Before anyone starts combing their patent database for spin-outs: most unused patents are not companies. An expert group at the European Commission estimated that only 8 to 24 percent of European patents could be valorised at all. And value is extremely concentrated: a very small share of patents carries most of it.

At Xerox, too, 24 of the 35 projects came to nothing. Eleven more than made up for the rest.

So a carve-out is a selection game. The work isn't getting as much as possible out of the drawer. It's finding the right thing, checking it properly and then building it with conviction. Volume isn't the goal. Hits are.

What changes inside the parent company

A carve-out doesn't only work outwards. It gives management back the quarters it would otherwise have spent defending a project that fits no structure. And it sends a signal inwards: this company brings research to market, not just to slide decks. That attracts exactly the people you want to keep in research, strategy and innovation.

There's a difference from the classic divestment of a division, too. Split off a running unit and you often struggle afterwards with orphaned fixed costs and falling margins. Spin out technology that generates no revenue today and you barely have that problem. No running business gets cut apart. One gets built.

The zoom-out: part of the solution has already been paid for

Europe has spent years debating its innovation weakness and usually reaches for the most expensive remedies: more grants, more programmes, more research. The obvious gets overlooked. A substantial part of the solution doesn't need to be invented. It's already researched, patented and paid for. It just needs to be carved out.

The question isn't primarily whether Europe invents enough. The question is whether it ever brings what it has already invented to market.

The principle: carve out in order to build

What would a system look like that actually turns stranded IP into companies? It would have four properties.

First: selection before volume. Every technology gets assessed before anyone incorporates anything: technological differentiation, real demand, viable venture economics, the right timing.

Second: build, don't split. Market first, then team, then structure, then capital. A carve-out doesn't end with the signature. That's where it starts.

Third: equity instead of a sale. Licence and minority stake instead of a one-off proceed, with a contractually agreed way back: licence, partnership or acquisition on terms defined up front.

Fourth: the link to the lab stays. The inventors stay in the company's orbit, as advisors, development partners or pilot customers. The science doesn't leave the house. It gets a structure that fits it.

Coming full circle

Xerox invented some of the most important technologies of its time and watched others make the money. Today's corporate lab is well on its way there. Both had the technology first.

A patent in a drawer is a bet you've paid for but never played.

The most expensive technology isn't the one that fails. It's the one that never leaves the drawer.

Sources

• Giuri, Mariani et al., "Inventors and Invention Processes in Europe: Results from the PatVal-EU Survey", Research Policy 36 (2007), Tables 6–7: around 36% of European patents unused, large firms 40%, small firms 18% — doi.org
• Torrisi, Gambardella, Giuri, Harhoff, Hoisl, Mariani, "Used, Blocking and Sleeping Patents: Empirical Evidence from a Large-Scale Inventor Survey", Research Policy 45 (2016) — doi.org
• Walsh, Lee, Jung, "Win, Lose or Draw? The Fate of Patented Inventions", Research Policy 45 (2016): 55% commercialised, failure more common than preemption — doi.org
• European Patent Office, "Valorisation of Scientific Results: Patent Commercialisation Scoreboard" (2020), p. 27: 36% exploited, 42% planned, 21% no plan — epo.org
• European Commission, "Towards Enhanced Patent Valorisation for Growth and Jobs", SWD(2012) 458 (2012), pp. 5–6: 8–24% of patents valorisable — ec.europa.eu
• European Commission JRC, "The 2024 EU Industrial R&D Investment Scoreboard" (2024): 322 EU companies, €234bn R&D in 2023 — jrc.ec.europa.eu
• Mario Draghi / European Commission, "The Future of European Competitiveness, Part A" (2024), pp. 6, 28, 29: mature technologies, "middle technology trap", failure to commercialise — commission.europa.eu
• Henry W. Chesbrough, "Why Companies Should Have Open Business Models", MIT Sloan Management Review 48(2) (2007), p. 23: informal estimate, 5–25% of patents used — sloanreview.mit.edu
• Henry W. Chesbrough, "Graceful Exits and Missed Opportunities: Xerox's Management of Its Technology Spin-off Organizations", Business History Review 76(4) (2002), and "Open Business Models", ch. 2 (2006): 11 of 35 spin-offs worth twice Xerox's market value — berkeley.edu
• Cisco Systems, "Cisco Completes Acquisition of Nuova Systems" (2008), and Form 10-K, fiscal 2013: spin-in with call option, up to $678M and $863M — cisco.com · sec.gov
• McKinsey & Company, "Beating the Odds: What Really Matters for Successful Spin-offs" (2026): 55% of large spin-offs with negative combined return after three years — mckinsey.com
• Bain & Company, "PE-Backed Carve-Outs Used to Be Reliable Winners. So What Happened?" (2025): multiple 3.0x before 2012, 1.5x since — bain.com
• Deloitte UK, "Unlocking Value: Approaching Stranded Costs in M&A" (2025): margin dip after divestments — deloitte.com
• German Corporate Income Tax Act § 8b, Income Tax Act § 6(6) — gesetze-im-internet.de