IP Group — Detailed look at the deeptech portfolio

Following IP Group’s deeptech investor event on 19 September 2023, we have prepared a detailed discussion of the company’s current deeptech portfolio.

However, we believe it is instructive to first define what deeptech is and how it differs from more traditional tech. Deeptech is normally defined as breakthrough technologies based on cutting-edge scientific advances and/or tangible engineering innovation with the ability to disrupt several industries. Another way to describe it, according to Bessemer Venture Partners, is that it is ‘a technology that was science fiction in the past but is reality today’.

Deeptech ventures put strong emphasis on R&D early in their lifecycle and are characterised by a long initial development phase, often with substantial expenditure ahead of revenues and product-market fit. Importantly, each of the subsequent development stages can give rise to new challenges. This influences the deeptech funding cycle, with the authors of The European Deep Tech Report 2023 (Dealroom.co, Lakestar and Walden Catalyst) estimating that European deeptech start-ups raised, on average, their Series A (of over €4m) and third funding rounds around two and four years after their seed rounds, respectively, which is later than the 1.5 and three years on average for other start-ups, respectively. The average company age across the XB100 ranking of the most promising private deeptech companies (introduced earlier this year by XPRIZE Foundation and Bessemer Venture Partners) is 9.2 years.

However, this translates into a stronger competitive advantage in the form of technological edge and intellectual property compared to more traditional tech businesses (an average XB100 company had 122 patent documents). Deeptech companies often have no comparable product in the market. Therefore, they do not have to rely on network effects and market lock-up in the way traditional tech companies do. As a result, while the gestation period of deeptech businesses is longer than for traditional tech, they can still provide strong returns, with the average value-capital multiple (ie the ratio of the latest valuation to the total amount of capital raised) on the XB100 list at 5.5x.

Once a new technology or product matures and scales, it is reclassified as regular tech. Consequently, there is no fixed set of deeptech subsectors, but rather a list that evolves over time. The European Deep Tech Report 2023 highlights the following four emerging (‘new/true frontier’) deeptech segments in Europe:

Other segments considered by the report as deeptech include synthetic biology, advanced materials, robotics, transportation, foodtech and agtech, as well as cybersecurity. The deeptech sector was broken down into nine categories for the purpose of compiling the XB100 ranking: agriculture, AI, aviation, climate, mobility, next-generation biotech, quantum, robotics and space. That said, we note that the above range of ‘deeptech’ sectors may be considered quite broad, with some deeptech-focused investors using a more narrow definition as part of their investment approach.

Europe has an excellent footprint in terms of top-tier universities producing technology spin-offs. Moreover, European governments are committing sizeable capital to boost regional deeptech start-ups. This includes a number of EU-wide programmes, such as:

Furthermore, there are several country-level initiatives that are aimed at fostering disruptive tech start-ups, in particular in Germany, France and the UK. For instance, the UK unveiled a 10-year strategy with £2.5bn in funding to turn the country into a leader in quantum computing, while Germany launched a fund of up to €1bn to invest in deeptech and climate tech.

That said, there are certain deficiencies in terms of private capital dedicated to deeptech investments. This particularly applies to late-stage (growth) capital needed to fuel an extensive scaling phase of deeptech businesses (ie beyond Series A round). Overall, the size of the European deeptech VC market remains inferior to the US market. For instance, VC firms invested €10bn into AI and machine learning start-ups in Europe in 2022 compared to €38bn in North America, according to PitchBook. While this is partly due to the smaller size of the overall VC industry in Europe versus the US, it may be also a result of European generalist VC firms being more risk averse in terms of investing in disruptive technologies at their pre-revenue stage (often preferring software-as-service and e-commerce businesses).

With its evergreen structure (providing funding visibility to founders), a patient approach to grooming university spin-outs (often as an active board member providing support in fund-raising, talent acquisition and strategic decision making), as well as an extensive, global network of lead and co-investors, IP Group helps tackle the persistent funding gap for deeptech businesses in Europe. For instance, both Featurespace and Garrison Technology (some of IP Group’s largest deeptech holdings as at end-June 2023) have US VC investors on board, while Ultraleap’s Series D round in 2021 was led by Tencent. Another example is the introduction of the Australian superannuation fund Hostplus as investor in WaveOptics.

That said, it is worth noting that, more recently, IP Group’s preference has been to invest in deeptech ventures at a slightly later stage. This is to leverage the increasing amount of funding provided to spin-out companies at the earliest stages, partly in the form of government grants through university technology transfer offices (TTOs) and InnovateUK, and partly through the UK tax system as (Seed) Enterprise Investment Schemes (EIS/SEIS) eligible investments, which provide incentives to invest capital at this very early stage. In staying close to universities and other centres of innovation, but encouraging the flow of such capital into high-risk opportunities, IP Group can achieve the best of both worlds – de-risking its early stage investments while still maintaining access to the best deals.

According to its mission statement in the deeptech domain, IP Group is working towards a tech-enriched future by investing in assets that will deliver digital resilience, create prosperity and enable new human capability (therefore contributing the United Nations Sustainability Development Goal (SDG) 9: industry, innovation and infrastructure). To achieve this, the company targets ‘differentiated, defensible innovations with a strong value proposition and the potential to disrupt multibillion-dollar global markets’, which is very much aligned with the definition of deeptech.

IP Group’s key focus areas within its deeptech investments currently cover applied AI (including, among others, in cyber security), next-generation networks, human-machine interfaces and the future of computing (see Exhibit 1). Some of IP Group’s deeptech holdings not classified as applied AI still use AI extensively in their operations (eg Ultraleap, Slamcore, Audioscenic). Moreover, it is worth noting that IP Group also invests in novel energy businesses within its cleantech division (21% of IP Group’s total portfolio value at end-June 2023, not covered in this research note).

IP Group’s technology expertise allows it to take on a greater degree of risk and invest in disruptive technologies before they reach their inflexion point, which in turn provides the prospects for greater returns. The most prominent deeptech exit case within IP Group’s track record is WaveOptics, a human-machine interface business that IP Group invested in back in 2015. WaveOptics is a supplier of AR displays, which allow for the overlaying of virtual objects onto the real world through a transparent surface like glass. The company was acquired by Snap (which used a WaveOptics display for its AR Spectacles glasses launched in 2021) for more than US$0.5bn in May 2021, which IP Group believes to be one of UK’s largest ever venture-backed deeptech exits to that date. IP Group realised a strong MOIC of 7.9x and gross IRR of 54.3% on its WaveOptics investment (IP Group invested in total £7.6m into the business over six years, including an initial investment of £1.5m). The £60.3m consideration was paid in two equal tranches: the first was paid in Snap shares upon completion in May 2021 (IP Group sold the shares shortly after closing the deal) and the second tranche of £31m was paid in cash after 24 months (in H123).

IP Group’s ability to gain exposure to the emerging AR theme is also illustrated by the fact that it was an investor in Ultraleap since the very beginning, providing pre-seed capital in 2014 (the business remains one of its key deeptech holdings, see below). We note that, while the growth of the AR market in recent years has been slower than some expected, the upcoming launch of Apple Vision Pro could spur further demand for similar solutions. IP Group believes that the AR sector is now reaching a point at which hardware capabilities can start supporting the true promise and vision of spatial computing.

Some other notable realisations from IP Group’s portfolio in recent years are listed below. We note that the return IP Group generated on most of these deals was not made public.

IP Group will consider various exit routes for its holdings, including sale to a strategic investor and IPO on the Main LSE market or in the US. While the company floated several of its early-stage holdings on AIM in the past (eg Actual Experience, Applied Graphene Materials, Mirriad Advertising), it is less likely to do so in the future, given the volatile market valuations of listed early-stage businesses that are yet to break even, which make subsequent capital raises more challenging.

IP Group’s deeptech portfolio consisted of 29 holdings valued at £206.4m at end-June 2023, representing 16% of IP Group’s total portfolio value (see Exhibit 2). Three of its largest deeptech holdings (Featurespace, Garrison and Ultraleap) make up 61% of the total deeptech portfolio value (see Exhibit 3).

We note that Featurespace, Teya and Garrison Technology already generate meaningful revenue in the tens of millions sterling and that, overall, around 90% of the deeptech portfolio is already recording revenue. Both Featurespace and Garrison (as well as Teya) target continued double-digit top-line growth. Ultraleap may also turn out to be a strong top-line performer (it is currently in the early revenue stage, ie £1–10m pa), but its revenue growth is dependent on a few large, potential royalty-generating contracts, which are yet to be signed. While the revenue traction across IP Group’s deeptech portfolio suggests a certain degree of maturity (and possibly improving realisation prospects), the private investors backing these companies may decide to hold on to their investments for longer to realise their full potential.

Both Featurespace and Ultraleap were valued by an external thirty-party valuer at end-June 2023 (using a revenue multiple and adjusted funding round method, respectively). Garrison Technology was valued based on its oversubscribed £15.5m funding round in H123, with new investors including Legal and General Ventures and British Patient Capital. IP Group did not recognise any fair value gains or losses on the holding in H123, suggesting that the valuation of the latest funding round was in line with its prior carrying value (which was based on a third-party valuation with reference to future market/commercial events). Below we discuss IP Group’s major holdings grouped by its key focus areas.

IP Group’s largest applied AI holding at present is its 20.4% stake in Featurespace (5.0% of IP Group’s total portfolio value at end-June 2023), a University of Cambridge spin-off founded in 2008, which deploys its proprietary machine learning algorithms (based on adaptive behavioural analytics) to monitor customer data for real-time fraud and financial crime detection. Some Featurespace clients that have implemented the company’s ARIC Risk Hub include HSBC, NatWest, Akbank and Danske Bank. Moreover, Featurespace is providing white-label solutions to payment providers (eg TSYS, Worldpay, Elavon).

In 2021, the company introduced its Automated Deep Behavioural Networks (using deep recurrent neural networks) for the card and payments industry. This allows for the identification of more subtle fraudulent behaviour patterns than what existing models have already been capturing. In a recent podcast, Featurespace’s director of innovation highlighted that he is excited about the applications of generative AI in transaction processing and fighting financial crime, given that these AI models are built to recover and simulate all the information available in a given dataset (as opposed to some other AI models that are focused on a particular outcome, eg fraud or credit default prediction). This full understanding of a dataset can then be used as deep embeddings for downstream models. A generative AI model also allows for the creation of ‘synthetic’ customer data (ie stripped of any privacy-related aspects), and therefore for the resolution of data portability issues, such as the inability to use data for testing, consortia or across different jurisdictions.

The company continues to significantly grow based on its value proposition illustrated, among other things, by the fact that recently NatWest has improved its scam detection rate by 135% using Featurespace’s technology. In this context, we also note that McKinsey highlighted in October 2022 that combining machine learning with other advanced algorithms is where banks can reap the most immediate and significant benefits in their anti-money laundering efforts. Featurespace’s five-year recurring revenue CAGR to 2021 was 74% (with 75% recurring revenue growth in 2021). The company continued growing at a double-digit percentage rate in both 2022 and H123.

IP Group gained exposure to Featurespace via the Touchstone Innovations acquisition in 2017 (with Featurespace’s end-2017 carrying value at £15.6m). Subsequently, IP Group invested £15.1m into Featurespace, with end-June 2023 fair value of its 20.4% stake reaching £64.1m. In November 2022, Featurespace secured funding from the UK and US governments (in the form of the privacy enhancing technologies Challenge Prize).

Other significant applied AI holdings include:

IP Group’s major next-gen network investment is its 23.6% stake is Garrison Technology (2.5% of IP Group’s total portfolio value at end-June 2023), a provider of a remote web browser and related hardware, allowing enterprises and government agencies to prevent local machines from accessing malicious data sources by keeping all activity in the cloud (by means of web isolation). Its solution is primarily aimed at protecting against ransomware and phishing attacks. In recent years, the company reported strong demand from government customers (especially in the UK and US), as well as solid growth across its key commercial accounts (eg Lloyds Banking Group). In its H123 report, IP Group highlighted the large and increasing presence of Garrison in the US.

As a result, Garrison’s revenues in the fiscal year to end-March 2023 reached around £20m after posting a four-year CAGR of 64% (including continued double-digit growth in the last financial year). IP Group expects the business to see further double-digit percentage growth in the current financial year. IP Group gained exposure to Garrison through the acquisition of Touchstone Innovations in 2017 (which backed Garrison’s seed round in 2015). Overall, IP Group has invested £13.8m in Garrison to date, with its end-June 2023 carrying value standing at 2.3x this amount.

Further major holdings in the next-gen network area include:

IP Group’s holding in the human-machine interface domain that is particularly worth discussing is its 17.0% stake in Ultraleap (3.0% of IP Group’s end-June 2023 total portfolio value). The company was founded in 2013 and is one of the leading IP owners in spatial computing and natural user interface technology, with more than 420 patents (of which 75% have strong to very strong blocking components, according to the company). Ultraleap’s two complementary core technologies are hand tracking (where it believes it has built a comprehensive IP moat) and mid-air haptics (where it claims to have the only commercially available solution in the world).

Haptics may be defined as a technology that transmits tactile information via ultrasound waves using sensations such as vibration, touch and force feedback (‘feeling without touching’). The company’s hand tracking solution involves infrared LEDs and two cameras. The LEDs pulse in sync with the camera frame rate, allowing for significantly lower power use and increased intensity. In this way, Ultraleap's hand tracking software takes raw data on the position, orientation and velocity of the user’s hands and turns it into a real-time digital model. According to Ultraleap, this solution is superior in terms of providing the most consistent and reliable data compared to other sources of raw data (eg RGB camera, GPS, laser arrays or LiDAR). IP Group believes that, while there are competing hand tracking solutions, Ultraleap’s technology is considered ‘gold standard’ within the industry.

Ultraleap enables its customers to use these technologies in three core verticals: XR (extended reality, eg in VR headsets used for safety-critical flight attendant training), out of home (touchless interactions with screens/terminals, interactive touchpoints of a brand, museums and theme parks) and automotive (XR, back-seat digital screens and gesture control; see Exhibit 4).

Ultraleap licenses its technologies to key industry players via scalable B2B software-focused solutions. Existing clients that have bought and licensed its solutions include Acer, Autodesk, Qualcomm, Lego, PepsiCo, Skoda, Merlin Entertainments Group, Varjo and Warner Brothers. For instance, Varjo launched two of its latest headsets (Varjo XR-3 and Varjo-VR 3) with Ultraleap’s fifth-generation hand tracking integrated in them. Moreover, Ultraleap’s hand tracking technology is embedded within the Lynx R1 MR headset. Recently, the company launched the next generation of its hand tracking camera product, the Leap Motion Controller 2, which will be compatible with macOS, Android XR2 and Windows.

There are further holdings within IP Group’s portfolio providing exposure to the human-machine interface theme:

IP Group is also invested in quantum computing and other novel computing technologies. The future role of quantum computers is hardly deniable, with a high disruption potential across various sectors, including in particular finance, energy, materials, telecoms and healthcare/pharma. As highlighted in our recent short thematic piece, the quantum computing market size was about US$470m in 2021, with prospective growth to 2026 expected at 30% to more than US$1bn by 2026, and potentially exceeding US$125bn by 2030 (though most of the market revenue growth is projected to come from cloud access services to a quantum computer). IP Group is currently cautious in terms of committing new capital to quantum computing given the increasing hype around the theme.

IP Group holds a stake in Quantum Motion, a company developing fault tolerant quantum computing architectures that are compatible with standard silicon processing. In February 2023, the company raised more than £42m in its latest funding round, the largest ever single investment into a quantum computing start-up in the UK. The oversubscribed round was led by Bosch Ventures alongside Porsche SE and British Patient Capital. All the company’s existing investors (Inkef, IP Group, NSSIF, Octopus Ventures, Oxford Sciences Enterprises and Parkwalk Advisors) also participated. The company will use the funding to accelerate the development of its silicon-based approach to building a cost-effective and scalable quantum computer, especially by developing deeper ties with its manufacturing partners and expanding its central London headquarters.

Intrinsic Semiconductor Technologies developed an innovative approach to non-volatile memory using resistive random-access memory (RRAM), which can read data 10x to 100x faster and write it 1,000x faster than existing solutions. Importantly, it uses standard, frequently used materials (it is based on silicon oxide), as opposed to some other RRAM solutions. The company highlights that it can put efficient non-volatile memory directly on advanced processor chips, revolutionising next-generation microcontrollers and edge processors.

This is a high conviction asset for IP Group’s deeptech team, which believes the company has the potential to become the next unicorn as it seeks to disrupt the US$100bn+ FLASH memory market with its simple yet powerful technology, first created at University College London. IP Group notes that a company at only a slightly later stage of development, Weebit Nano, is listed in Australia with a current market cap of c £300m.

Lumai develops optically trained, photonic neural networks based on free space optics, which it claims has a near-term speed limit that is 100x faster than the human brain and 1,000x faster than traditional electronics. Lumai’s solutions are aimed at bypassing the Moore’s Law limitation and addressing the need for energy efficiency by moving from digital to analogue computation. IP Group’s deeptech team will be looking closely at the near-term growth of Lumai, given the massive interest it sees in the sector on the back of generative AI.

IP Group also holds a stake in Oxford Quantum Circuits, which owns the only quantum computer that is commercially available in the UK. The company’s computer is a complete functional unit, including the control system, the hardware and the software.

In the exhibit below, we present IP Group’s other deeptech holdings (excluding organic and de minimis investments).