Draft FrameworkJuly 2026

The evidence base

Overview of the research behind the Berlin Model.

Research summarized from JUNI's report on the topic · Full report coming soon

Europe produces world-class research but struggles to commercialize it. In Germany, tech transfer is slow (takes an average of 18.4 months), terms vary widely between institutions, and investors consistently report that common term structures put spinouts outside what they can finance.

The Berlin Model is built based on the evidence laid out below: what the world's leading spinout ecosystems actually do, what the empirical record shows works, and what investors will fund. The goal is full alignment with international best practice and a true step forward for the ecosystem.

01 · International benchmarks

Top ecosystems converge on a single-digit range

RegionInstitutionUniversity equityNotes
USStanford≤ 5%
USMIT≤ 5%
USUCLA≤ 5%
USDuke2–5%
EuropeETH Zürich2–4%highest spinout founder satisfaction in Europe
EuropeTU Munich≤ 7%virtual shares
EuropeEPFL≤ 10%
EuropeDenmark≤ 5%capped by national law (2026), mandatory
UKUK uni median · 202511%stakes still falling
UKUK uni avg · 202416%international outlier; stakes continue to fall fast
UKUK uni avg · 202322%down from ~50% a few years ago

Leading ecosystems converge on a single-digit range for equity, regardless of legal system. Ranges are inclusive of all sectors. Upper bounds typically reflect life sciences; lower bounds typically reflect software, where deals are often equity-only. Where royalties apply, they are typically up to 3%, occasionally up to 5%, and often lower or absent (particularly for software).

Where stakes have been higher (UK, parts of continental Europe), spinouts struggle to raise, cap tables get retroactively restructured, and the terms are correcting. The trend worldwide is toward single-digit university equity in spinouts.

Show full benchmark table
Can include life sciences Deep tech (non-life-sci) Software
InstitutionEquityDilution protection≈ Equiv. fully-dilutableRoyalties
Stanford≤ 5%Negotiated case-by-case
MIT≤ 5%Sometimes
UCLA≤ 5%None≤ 5%~ ≤3%
Duke2–5%
UCSD≤ 5%None≤ 5%None (standard; sometimes optional)
Northeastern3%2%
Harvard (software)2%None2%None
Cornell (life sci)4%$4M~ 5.5%Yes
Cornell (hardware / deep tech)4%$2M~ 4.5%Yes
Cornell (software)4%$2M~ 4.5%None
Carnegie Mellon (patent)5–6%$2M~ 5.5–7%2%
Carnegie Mellon (non-patent / software)3%$2M~ 3.5%None
Yale (software)3%None3%None
Yale (other, non-life-sci)5%None5%1–2%
Wisconsin–Madison (WARF)5%$2M~ 5.5–6%2%
Washington State (option 1)0%N/A0%5%
Washington State (option 2)3%$3M~ 4–4.5%2%
Washington State (option 3)5%$3M~ 7%1%
Columbia5%Through Series A (highest US cap seen)~ 9–9.5%
Univ. of British Columbia5%2M CAD (≈ €1.25M)~ 5.5%1–2%
Univ. of WaterlooNoneNoneNoneNone
ETH Zurich (express license)2–4%None2–4%0–2%
EPFL≤ 10%None≤ 10%≤ 5%
TUM (Fast Track)≤ 7% (virtual)None≤ 7%Can be included
Denmark Model A (royalties only)0%N/A0%1–6% (scorecard)
Denmark Model B (equity + royalties)≤ 5% (capped by law)~ €2.5M~ ≤7%Half of Model A

From JUNI's research report. "—" indicates information that was not available or not conclusive; it does not imply the presence or absence of a term. Approximate fully-dilutable equity equivalences use the deep-tech test case defined in the report.

02 · The empirical record

Higher uni stakes → lower success rates

−1.8pp
Probability of raising VC drops for every +1pp of university equity.
Hellmann et al.
+38%
UK spinout VC funding 2024, as uni stake fell 22 → 16%. Wider market fell 19% in same period.
Beauhurst / TenU, 2025
15 / 10%
Median UK uni stake in failed vs. surviving or exited spinouts.
RAEng / Beauhurst
+36%
Cambridge (~12.6%) vs. Oxford (~24.3%): half the equity, 36% more exits (between 2011-2021).
Independent UK Review

Across European investor surveys (SPRIND, RAEng, spinout.fyi) and our own interviews, 10% university equity appears consistently as the upper bound for an investable spinout. HTGF estimates two of three German knowledge-based spin-offs are hardly financeable today simply because of their contractual terms. (Orrick, OLNS #10, 2022)

03 · The hard questions — and what the evidence says
Equivalence"Our 15% fully dilutable stake is equivalent to a 5% stake with anti-dilution protection, once you account for dilution."

The math doesn't support it. A 5% stake protected to €10M post-money equals roughly 7% unprotected, not 15%. Matching a 15% fully dilutable stake would require protection past €100M raised, far beyond the highest US protection caps on record. Dilution protection through Series A is also rare, not standard, in the US. Oxford's own faculty have publicly called the equivalence claim wrong.

Run the numbers in the equity calculator →

UK comparison"But the UK does well with high stakes. Why can't we?"

UK universities succeed despite their terms, propped up by £2B+ of dedicated university-affiliated capital (OSE, Cambridge Innovation Capital, Parkwalk, IP Group, UCL Technology Fund) and EIS tax relief; this is infrastructure with no continental analogue. Without that backstop, investable deal terms matter more here, not less. Even inside the UK, terms are correcting: 22 → 16% in a single year, with VC funding +38%.

Fair share"Universities deserve more than single-digit equity for the IP they created."

Single-digit stakes are how the institutions making real money got there. MIT averages $64M/year in tech-transfer revenue (2019–25); Stanford and Duke similar. Those returns come from small stakes across hundreds of companies built over decades, not from extracting a larger share of fewer deals. Cheryl Cathey (Stanford OTL): "plant a lot of seeds." Over-negotiating reliably misses the breakouts and damages the innovation pipeline, as founders avoid working with difficult TTOs.

Doesn't apply here"These are US numbers. They don't apply to Germany/Europe."

They apply directly. European funds — Wilbe, Creator Fund, Extantia Capital, Zero Carbon Capital, Fly Ventures, and many others — publicly state they will not invest above 10% university equity, with several setting their threshold at single digits. Creator Fund's Jamie Macfarlane says quality funds "walk away" from spinouts where the university holds over 10%. Fly Ventures' Matt Wichrowski calls anything above that "a huge red flag." The convergence is investor-driven and holds true internationally.

Rigidity"Standardization is too rigid for our IP-heavy spinouts."

The Berlin Model is built around exactly this concern. Five tracks × three corridors, calibrated by a published IP scorecard, place each deal proportionate to the university's actual contribution. Tracks cover equity, royalties, blends, and fixed-payment options, including "IP for free" where appropriate. The same structure accommodates life sciences, deep tech, and software, and works across institutions with different needs and portfolio management capacities. Standard processes and contracts per track keep execution fast; negotiation reduces to the few decisions that need judgement.