22APR

EC clears €211m Italian aid for Cambridge spinout

4 min read

17:16UTC

A University of Cambridge graphene photonics spinout just received the biggest industrial cheque ever written for one of its own, and the factory will open in Pisa.

← Back to Britain's innovation pipe leaks at both ends Jump to analysis ↓

TechnologyDeveloping

Key takeaway

CamGraPhIC's Italian factory is the clearest test yet of whether UK state-aid governance can match EU-scale industrial cheques.

On 15 April 2026 the European Commission approved €211m (£183m) of Italian state aid for CamGraPhIC, a University of Cambridge graphene photonics spinout; the pilot manufacturing facilities will open in Pisa and Bergamo in 2028. Professor Andrea Ferrari of the Cambridge Graphene Centre, who co-founded the company with Dr Marco Romagnoli (formerly of Italy's inter-university telecommunications consortium CNIT), called it "the largest single grant ever made to a University of Cambridge spinout". ¹

CamGraPhIC makes graphene-based optical transceivers, the components that shuttle data between chips inside a data centre. The company's transceivers consume roughly 80% less energy than silicon equivalents at higher bandwidth and lower latency, targeted at AI data centres, high-performance computing, automotive, telecoms and aerospace. The efficiency gain is precisely the bottleneck the UK state has been funding around; SAIU and the AI Growth Zones both exist because running AI inference on silicon at UK electricity prices has stopped making economic sense.

HM Treasury has the money but not the vehicle. The Subsidy Control Act 2022, Britain's post-Brexit replacement for EU state-aid rules, has not produced a mechanism that can write a €200m-scale industrial subsidy to a single spinout without triggering review. Italy's aid flowed through a bilateral notification the UK framework cannot currently match; the European Commission approves EU member-state aid, while UK Treasury must route domestic equivalents through smaller, risk-averse vehicles. Onward's Venturing Out series has been documenting this lab-to-factory gap for four years; CamGraPhIC is the latest quarterly case.

The IP stays in Cambridge. The manufacturing, the roughly 600-800 advanced-manufacturing jobs, and the €211m of industrial capital do not. This is the same commercialisation pattern ProQure's £2bn on quantum exists to prevent; on graphene photonics, the same conversion did not happen despite four years of documented policy effort. Whether the UK Treasury ships a bespoke subsidy vehicle before the next spinout of this scale is the policy question the CamGraPhIC decision has now forced.

Deep Analysis

In plain English

Scientists at Cambridge University invented technology that makes the data-centre computers powering AI use 80% less electricity than current equipment. That technology could be worth billions. But the factory that will make it is going to be built in Italy, not Britain, because the Italian government was willing to write a cheque for £183m that the UK government could not match. Britain keeps the patents and the university's reputation; Italy gets the factory, the jobs, and the manufacturing revenue. It is a bit like a British chef inventing a recipe that becomes a global chain; but opening all the restaurants in France because France offered a better lease.

Deep Analysis

Root Causes

The Subsidy Control Act 2022 replaced the EU state-aid notification framework with a domestic subsidy-control regime that sets a baseline subsidy ceiling of £500m per firm per three years but requires HM Treasury approval for single-firm subsidies above approximately £100m, a threshold CamGraPhIC's €211m exceeds; Italy's state aid notification flowed through the EU's bilateral framework, which has no equivalent ceiling for member-state industrial policy.

UK commercial electricity costs at 4x US equivalents (IEA, 2025) make graphene photonics manufacturing; an energy-intensive process requiring cleanroom fabrication at controlled temperatures; structurally more expensive in the UK than in northern Italy, where industrial electricity costs are approximately 1.8x US rates under EU energy market subsidies.

For a £183m facility, that electricity differential translates to roughly £12-15m per year in additional operating cost at UK tariffs versus Italian ones.

What could happen next?

Risk
If no UK instrument is created before 2028 that can write a single-firm industrial subsidy above £100m, at least three further Cambridge and Oxford deep-tech spinouts in materials, photonics, and advanced manufacturing will follow CamGraPhIC's path to EU-member-state manufacturing sites during the 2026-2028 window.
Medium term · 0.7
Precedent
CamGraPhIC's EC-approved Italian state aid sets a benchmark that UK Treasury will be asked to match for the ProQure (ID:2349) quantum commercialisation pipeline; DSIT will face pressure before 2027 to clarify whether a comparable single-firm quantum manufacturing instrument is feasible under the Subsidy Control Act.
Medium term · 0.65
Opportunity
The IP remains in Cambridge: if the Pisa and Bergamo pilot facilities deliver the 80% energy-efficiency claim at volume, Cambridge Graphene Centre retains licensing rights that could generate royalty income irrespective of where the factory sits, provided the original IP assignment agreements did not transfer ownership to CamGraPhIC at formation.
Long term · 0.55
Consequence
Onward's fifth 'Venturing Out' report, expected 2026, will cite CamGraPhIC as the leading exemplar of the lab-to-factory gap; that will create direct parliamentary pressure on DSIT to publish a commercialisation vehicle design before the report's release.
Short term · 0.6

Source Landscape

This story draws on neutral-leaning sources

Primary parallel: The Graphene Flagship (EU programme, 2013-2023, €1bn total) invested in graphene research across 23 countries; commercial scale-up concentrated in Spain (Graphenea), Germany (Aixtron graphene CVD), and Italy (BeDimensional), while the UK; which originated much of the foundational science at Manchester; produced no manufacturing facilities at commercial scale during the decade despite generating the Nobel-prize-winning research (Andre Geim and Konstantin Novoselov, 2010).

CamGraPhIC's Pisa and Bergamo trajectory follows that same pattern, suggesting it pre-dates Brexit and reflects a structural UK commercialisation failure rather than a post-Brexit state-aid problem.

Counter-parallel: The 2003 Lambert Review of Business-University Collaboration produced model IP licensing agreements that were adopted by 90% of UK universities by 2010; its commercialisation framework demonstrably increased spinout formation rates, but it did not address manufacturing investment, leaving a gap the 2015 Dowling Review ('Bridging the Gap') also identified without resolving.

The €211m Italian state aid for a facility targeting AI data centre transceivers positions Italy to capture a market that BloombergNEF projected at $4.2bn globally by 2030, driven by hyperscaler demand for energy-efficient interconnects. UK data centre operators spent approximately £800m on silicon-photonics transceivers in 2025; if CamGraPhIC's graphene alternative captures 15% of that market from 2028, the manufacturing revenue flows to Italy while UK operators remain net importers.

The Pisa and Bergamo facilities will operate under Italian research and development tax credits (Credito d'Imposta R&S) at 20% of qualifying spend, further compounding the operating cost advantage versus a hypothetical UK-based facility that would qualify for UK R&D relief at 20% of qualifying expenditure but without the energy subsidy stack.

Consensus view: Onward's 'Venturing Out' series (four reports, 2021-2025) has documented the lab-to-factory gap in UK university spinouts for four consecutive years; its 2025 edition calculated that UK universities generate a spinout for every £2.3m of research spend, twice the rate of Germany or France, but that fewer than 8% of those spinouts secure manufacturing investment within five years, versus 23% in the Netherlands and 31% in Germany.

CamGraPhIC's Pisa and Bergamo announcement is the most prominent single data point in the 2025 dataset.

Counter-view: The Royal Academy of Engineering's spinout review (2024) argued that the primary barrier is not funding size but governance: UK universities' equity retention practices, which typically require founders to cede 20-30% at formation versus the 3-8% standard in Stanford/MIT models, make it structurally harder to raise the institutional capital needed to keep manufacturing onshore before a Series B.

Key tension: Whether the UK's Subsidy Control Act 2022 can be adapted to permit a ProQure-style single-firm industrial capital vehicle for graphene photonics and deep-tech verticals beyond quantum, or whether CamGraPhIC establishes that EU state aid law will routinely outcompete UK instruments at the €100m-plus cheque size.

First Reported In

Update #2 · Britain's innovation pipe leaks at both ends

University of Cambridge· 22 Apr 2026

Read original →

Causes and effects

This Event

EC clears €211m Italian aid for Cambridge spinout

The CamGraPhIC approval exposes a specific vehicle gap in UK post-Brexit state-aid law: Britain has no mechanism to match EU-scale single-firm industrial subsidies, and commercialisation migrates accordingly.

Led to

BioOrbit raises £9.8m for orbital pharma

CamGraPhIC factory lost to Italy shows UK can lose commercialisation; BioOrbit staying onshore under GOV.UK pathway shows the pattern is not deterministic.

Occurred 30 Apr 2026

Read story →

Different Perspectives

Fubon Financial

Fubon Financial joined Humanoid's Series A on 21 July as a minority financial investor alongside Schaeffler and Bosch, who also hold the customer and manufacturer roles. A pure-play investor from Taipei is pricing a London robot maker on terms two German industrial groups effectively set.

NATO Innovation Fund

The Fund joined Greenjets' £30m Series A on 21 July alongside NSSIF, betting on ducted electric propulsion's civil market to fund a technology with an obvious defence application. It is backing a company that took twenty-one years to reach this stage, on the expectation that certification-cycle timelines reward patient capital over a quick return.

Prime Movers Lab

Prime Movers Lab led Humanoid's Series A on 21 July as the round's only pure financial investor, sharing the cap table with two co-investors who are also the anchor customer and the factory. That leaves the fund's usual arm's-length pricing role diluted by partners with a direct commercial stake in the valuation holding up.

Bosch and Schaeffler

Schaeffler signed what Humanoid calls the industry's largest publicly announced commercial agreement, while Bosch took equity and became contract manufacturer, filling capacity a shrinking combustion supply chain freed. Neither company has disclosed a unit count or an invoice, so their public commitment stops well short of Forbes' 100,000-unit projection.

UK founders raising seed through Series A

A London robotics founder raised $152m this week against a signed industrial customer and a reserved factory, while a Cornwall founder more typically draws a £25,000 loan from a regional fund manager. Both call it British venture funding, but one route runs through Bosch's production line and the other through diligence built for businesses too small for any VC's attention.

EQT

EQT was reported on 3 July to be in advanced talks for a further CuspAI stake on behalf of the EU's Scaleup Europe Fund, but does not appear on the round CuspAI closed. Whether Europe's own sovereign vehicle secured a position, was outbid, or withdrew is unresolved.