1 Introduction

Recent technical advances highlight the potential of quantum technologies in several domains including simulation, sensing and metrology, computation, and communication. This includes achieving quantum supremacy using programmable processors based on superconducting qubits.Footnote 1 The potential applications of quantum technologies range from (1) simulating quantum systems to enhance our fundamental understanding of nature and its applications (e.g. modelling chemical processes in drug development) to (2) achieving unprecedented sensitivity, resolution and accuracy in measurement through quantum sensing and metrology, (3) solving computational problems that are beyond the reach of classical computing by using quantum computers to deploy quantum algorithms, and (4) developing a new generation of secure communication systems.Footnote 2

These advances are of interest to academia, industry and governmentsFootnote 3 due to the implications for (1) basic science, (2) a broad cross-section of industry sectors ranging from computing and IT to pharma, and (3) governments concerned about the ultimate national security implications of quantum technologies. Technology-based innovation is typically driven by the synergy between academia, industry and market forces such as the intensity of competition, new market entrants, availability of substitute solutions, and the bargaining power of customers and suppliers. Governments also play a key role since they set innovation incentives and regulatory frameworks. Additionally, it is often the case that governments are a core market participant since they may also be an end customer for new technological solutions, especially when these have applications in defence and national security.

Governments can provide push and pull incentives. In the context of quantum technologies, examples of push incentives include (1) the grants to fund the basic scientific research that resulted in the first quantum revolution (i.e. the development of the theory of quantum mechanics to understand the principles and rules that govern physical reality at a fundamental particle level), and (2) technology transfer grants designed to take these quantum mechanical principles and use them to develop new quantum technologies as part of the so-called second quantum revolutionFootnote 4 such as the European quantum technologies flagship programme.Footnote 5

The goal of government push incentives is typically to provide funding in the form of grants to accelerate transitions from technology readiness level (TRL) 3 (e.g. experimental proof of concept stage) to TRL 6–7 (e.g. pre-competitive technology prototypes demonstrated in an industrially relevant environment). Pull incentives such as patent protection are designed to incentivise disclosure of novel inventions in exchange for timebound intellectual property rights. Once the pre-competitive technology is sufficiently de-risked (e.g. at the TRL 7 stage), intellectual property rights (IPRs) and market incentives are typically sufficient to enable transitions from TRL 7 to commercially available products (TRL 8–9).

Anticipating the potential implications of quantum technologies in real-world products, commentators from academia, government and think tanks have started to raise concerns, propose governance and regulatory strategies, and suggest policy recommendations focused on quantum technologies. These proposals are often directed to incentives (e.g. push incentives and IPRs), standardisation (e.g. ISO/IEC/IEEE) and regulation (e.g. technology governance).Footnote 6 Such academic contributions and proposals can be better judged by examining them in conjunction with empirical data to facilitate evidence-based policy decisions. For instance, previous research has raised the concern of IP overprotection in the field of quantum technology and proposed reforms to the patent system.Footnote 7 But, is there sufficient real-world evidence that indicates that there is overprotection anno 2022?

What has been the patenting trend over the last 20 years for quantum technologies? Which countries and organisations are leading the quantum patenting activity? What claim formulations are being used to protect these inventions? In what jurisdictions are they being protected? Without more evidence-based IP research to address these types of questions, it is perilous for legal scholars and policymakers to presume they can propose reforms to existing IP and regulatory regimes with reasonable chances of achieving desirable social outcomes. This is especially the case in early-stage deep tech fields such as quantum technologies that require significant high-risk R&D investments and overcoming substantial scientific and engineering challenges.Footnote 8

Perceived under-protection is traditionally understood to cause a lack of economic incentives to innovate in nascent technical fields. In the absence of viable alternatives, it is generally assumed that patent rights, as well as other forms of IPRs, are needed to attract the significant investment required to fund the capital-intensive R&D efforts necessary to encourage innovation and dissemination in high-risk technical fields such as quantum computing. That said, the overprotection of technologies and information may also cause market barriers and hamper healthy competition and sector and industry-specific open innovation.

Patents are an indicator of the pace and dynamics of innovation at the invention stage. Consequently, empirical studies looking at the actual patenting activity can provide valuable evidence to help assess policy proposals related to IPRs, innovation, and regulation in light of patent office data. However, the few patent studies available are either not peer-reviewed,Footnote 9 focus on a narrow patent reviewFootnote 10 or methodological contribution,Footnote 11 or not recent.Footnote 12 In this paper, we report the results of a study designed to map the patent landscape of quantum technologies. In particular, we evaluate the real-world patenting trends over the last 20 years to determine: (1) the growth of quantum technology patents, (2) the technology breakdown and classification of patenting activity, (3) the choice of priority patent office, (4) the types of patent claims and strategies, (5) the subject matter of recently awarded patents, (6) the top patent owners and their characteristics, (7) the dominant patent portfolios in terms of forward citations, and (8) the geographical distribution of this patent activity. These results enable us to critically examine previously published qualitative findings on potential patent overprotection of applied quantum technologies at the current state of the art. Using our landscape results we critically examine if IP overprotection, and in particular patent protection, is posing a problem in the technical field of quantum technologies. Additionally, we analyse the innovation and policy implications of these results in the broader context of quantum innovation initiatives, governance of quantum technologies, and market competition dynamics.

2 What Are Quantum Technologies?

It is tempting to define a quantum technology as one whose operation is governed by the laws of quantum mechanics. However, such a definition is problematic since technically the laws of quantum mechanics govern the behaviour of all physical phenomena at a fundamental particle level. Even narrower definitions such as technologies whose operation exploit the principles of quantum mechanics would include the semiconductor devices used in the latest classical processors, since the channel length of the transistors used are currently in the 5 nm scale which means that quantum mechanical effects must be considered in the design and development of these electronic devices. Thus, quantum technologies are often classified into two generations, and the term is typically reserved for those belonging to the second generation (2G).Footnote 13

First-generation (1G) technologies include transistors, lasers and other devices whose design and operation are based on the principles of quantum mechanics (i.e. devices that directly leverage the scientific discoveries from the first quantum revolution). These 1G devices have resulted in computing and communication technologies that have transformed entire industries and society since the introduction of the first semiconductor integrated circuit (IC). The initial patent on ICs (US Patent No. US3138743A, filed in 1959 and granted in 1964) stated that the “[…] principal object of this invention to provide a novel miniaturised electronic circuit fabricated from a body of semiconductor material […] wherein all components of the electronic circuit are completely integrated into the body of semiconductor material”. Miniaturisation through ICs led to the computer and information technology revolution. 2G quantum technologies are technologies currently being developed that directly harness the unique quantum mechanical phenomena such as superposition, entanglement, tunnelling and quantisation to achieve their underlying principle of operation (i.e. devices whose operation directly exploits quantum phenomena) to achieve advantages over our current state-of-the-art technologies. These include quantum computation, quantum simulation, quantum communications and quantum sensing and metrology.

3 Patent Search Strategy and Landscaping

We developed a search strategy designed to answer the above questions. The strategy follows the recommendations to ensure consistency and transparency of patent landscaping,Footnote 14 as well as the checklist of information for patent landscapes to ensure quality and reproducibility,Footnote 15 but narrowed to answer the specific research questions of this study as opposed to providing a general patent landscape. Similar methodologies have been used to analyse the patent landscape of gene patentsFootnote 16 and drug repurposing.Footnote 17

The search strategy ranges from high sensitivity to high specificity to minimise false positives (Table 1 S4). In particular, the search strategy identifies (1) all patent documents broadly related to “quantum” (S1); (2) core quantum technology patents (S2); (3) patents with claims directed to “quantum”-related inventive concepts (S3); and (4) patents with independent claims related to quantum technologies (S4). While S1 optimises for sensitivity by capturing any patent document containing the keyword “quantum” and related quantum concepts (e.g. qubit, entanglement) to identify any patent broadly related to quantum (i.e. establishing a conservative upper bound of broadly defined quantum-related patents), S2 and S3 optimise for specificity by requiring the keyword to be in the title, abstract, or claims (TAC) for S2, only in the claims for S3, and as part of the independent claims of the patents for S4.

Table 1 Quantum technologies patented in the US (USPTO) and Europe (EPO)

The search achieves a high degree of specificity by further narrowing the results to the Cooperative Patent Classification (CPC) classes established by the United States Patent and Trademark Office (USPTO) and the European Patent Office (EPO) for specific quantum inventions. This leverages the manual classification conducted by USPTO and EPO patent experts to categorise each patent application and granted patent in the relevant CPC class, effectively combining the results of automatic search algorithms with manual expert reviews. The CPC is an extension of the World Intellectual Property Organization’s (WIPO) International Patent Classification (IPC) and is jointly managed by the USPTO and EPO to achieve harmonisation across patent offices and improve patent searching. Our search strategy leverages the CPC classes to classify the patents by broad quantum technology areas, including quantum devices (S5), quantum optics (S6), quantum information processing (S7), quantum computing (S8), quantum cryptography (S9), and quantum communication (S10). The patent documents are further analysed to determine the growth in annual patent activity for quantum technologies, the legal status of these patents, the choice of patent office, the patenting activity by CPC class, the language and inventive concepts claimed, top patent owners in quantum computing, forward citation analysis, and global geographical distribution.

4 Patent Landscape Results

4.1 Patent Search Results

Table 1 shows the results of our search for patent applications and granted patents related to quantum technology published over the last 20 years at the USPTO and EPO. The search strategy ranges from high sensitivity (S1) to high specificity (S2–S10) to minimize false positives. Our results indicate that there have been 236,642 quantum related patent applications and 178,033 grants since 2001 (S1). That said, the majority of these patents are only broadly related to quantum technologies. Thus, S1 establishes a conservative upper bound that captures all the quantum related patents even when quantum technology is defined in overly broad terms.

Searches S2–S10 narrow the search to increase specificity to identify the core quantum technology patents and classify these patents by the specific subfield of quantum technology. We found 20,581 granted patents where “quantum” is a core concept captured in the TAC of the granted patent (S2). Of these, 18,696 have at least one claim containing “quantum” as a limitation (S3) and in 10,318 the more important claims in the patent are directed to a quantum related inventive concept (S4). The results in S5–S10 indicate that most of these patents are related to quantum devices (n = 8965), followed by nanostructures/quantum optics (n = 3282), quantum information processing n = (2057), quantum computing (n = 1603), quantum cryptography (n = 736), and quantum communication (n = 632).

4.2 Annual Quantum Technology Patenting Activity and Legal Status

Figure 1 shows the annual patenting activity for quantum technologies at the USPTO and EPO and the corresponding legal status of patent documents published in a given year. The figure shows (1) the quantum patents granted in the particular year, (2) the rejected/abandoned patent applications, (3) the previously granted patents that expired that given year, (4) the active grants (non-expired patents), and (5) the pending patent applications. Our S2 search results shown in Fig. 1 indicate that there has been nearly a 10-fold increase in the number of quantum technology patents granted per year. There were only 161 patents granted in 2001. For comparison, in 2018 the USPTO and EPO granted 1555 patents corresponding to an overall compound annual growth rate (CAGR) of 15.23%. That said, our results also show a period of relatively low growth (CAGR = 4.05%) from 2003 (n = 608) to 2013 (n = 904), with most of the growth taking place (1) between 2001 and 2003 as a result of R&D efforts and patent filings from the late 1990s, and (2) patent applications filed in the last 6 years (since 2014). Overall, the active patent grants have increased from 111 in 2004 to 2028 in 2021 (CAGR = 18.64%).

Fig. 1
figure 1

Annual patenting activity for quantum technologies at the USPTO and EPO (Search ID: S2) and legal status by publication date of the patent document.

The relative proportion of granted applications to the total number of applications (for years with a small number of pending applications) indicates the patent allowance rate has ranged between 55 and 62%. Overall, 56.89% of the applications filed in the last 20 years have been granted (n = 19, 571 patents) and 43.11% (n = 14,830) were rejected/abandoned. Additionally, 22% (n = 4534) of the granted patents over this period have expired. Thus, approximately 50% (n = 19,364) of all the patent disclosures in the last 20 years are now in the public domain and freely available for society to use.

4.3 Technology Breakdown: Classification of QT Patents

Table 2 shows the top CPC classes, their description and the corresponding number of patents granted by the USPTO and EPO on these classes. Our results show that B82Y* classes related to nanostructures and their applications (e.g. metrology) have the highest number of patents (n > 7000), followed by semiconductor devices (n = 1868) and quantum computing (n = 1603).

Table 2 USPTO and EPO patents classified by CPC class (Search ID: S2)

4.4 USPTO and EPO Quantum Technology Patent Activity

Figure 2 shows the number of published patent documents (i.e. published patent applications and granted patents) on quantum technologies segregated by patent office (USPTO versus EPO). Our results indicate that the USPTO has been the patent office of choice over the last 20 years. In 2001, 63.2% of the quantum technology patents (S2) were published by the USPTO. By 2021 this proportion has increased to 78.8% of the joint EPO/USPTO patents cohort. Overall, the EPO is the priority office for 22.04% of the quantum technology patents (S2) while the USPTO accounts for 77.96%.

Fig. 2
figure 2

Choice of patent office (USPTO vs. EPO) for quantum technologies

4.5 Patents Claiming Quantum Technologies

Figure 3 shows the prevalence of (a) patents broadly related to quantum technologies (S1), (b) core quantum patents (S2), (c) patents with claims directed to a quantum concept (S3), and (d) patents with the independent claims (broadest claims) directed to a quantum concept (S4). Since the S1 search strategy includes any patent including the keyword “quantum” and related concepts (e.g. qubit, entangl* and superposit*) anywhere in the patent document it is highly sensitive but not specific (i.e. in addition to the core quantum technology patents it captures broadly related applications that mention “quantum”). Accordingly, our landscape results are based on more specific searches where quantum technology is core to the patent and accordingly the quantum keywords are used in the TAC (S2), claims (S3), independent claims (S4), and are classified into a quantum technology CPC class.

Fig. 3
figure 3

Relative prevalence of (a) patents broadly related to quantum technologies (S1), (b) core quantum patents (S2), (c) patents with claims directed to a quantum concept (S3), and (d) patents with the independent claims related to a quantum concept (S4)

Figure 3 shows that over 90% of patents that use the “quantum” keyword in the title or abstract (S2) also have claims directed to a “quantum” inventive concept (S3). Additionally, in over 50% of the patents, the quantum inventive concept is claimed in the broadest claims of the patent (i.e. independent claims). Since 2020 there have been over 2000 patents publications per year (S2), with over 1800 patent publications containing claims directed to a quantum concept yearly (S3) and over 1000 yearly patents with the broadest claim including a quantum feature or limitation (S4).

As seen in Table 1, our landscape results show that the majority of the quantum technology granted patents are directed to quantum devices (e.g. semiconductor devices, solid-state devices) and nanostructures/quantum optics, followed by quantum information processing, quantum computing, quantum cryptography and quantum communication. Table 2 shows S2 granted patents categorised by the corresponding CPC class in ranked order. Figure 3 shows relative the prevalence of USPTO and EPO quantum technology patents with claims related to (a) quantum circuits, (b) quantum computing, (c) quantum communication, and (d) quantum algorithms.

Over the last 20 years, patents including claim limitations related to quantum circuits (e.g. quantum and circuit*) were the most prevalent. Quantum communication-related claims were the second most prevalent from 2001 to 2013. However, starting in 2013 patents with claims directed to quantum computing became more prevalent than those on quantum communication. Patents containing claim limitations directed to quantum “algorithms” have been a minority. Patent attorneys may have been intentionally avoiding the use of the limitation “algorithm” to overcome subject matter eligibility restrictions related to the patent ineligibility of “abstract ideas” (Bilsky and Alice). That said, as shown in Fig. 4 there has been a significant increase in patents including the limitations “quantum” and “algorithm” in the claims since 2016.

Fig. 4
figure 4

Relative prevalence of USPTO and EPO quantum technology patents with claims directed to (a) quantum circuits, (b) quantum computing, (c) quantum communication, and (d) quantum algorithms

Figure 5 shows the concept landscape for quantum technology patents (S2). The graph was generated using automatic text analysis of the abstract, title and claims to identify frequently used terms and cluster them by category. Our results indicate that the S2 patents are clustered in six main areas, namely, (1) quantum circuits, (2) quantum dot devices, (3) quantum computing, (4) quantum dot layers, (5) quantum states, and (6) quantum keys. The second layer includes additional details regarding the terms used to describe and claim the respective inventive concepts. Some of these overlap across the primary groups. For instance, “quantum bits” (qubits) are both in patents related to “quantum states” and “quantum computing”. Similarly, the terms “quantum processor” and “quantum control” are part of the “quantum circuits” and “quantum computing” clusters of S2 patents”.

Fig. 5
figure 5

Concept landscape for quantum technology patents (Search ID: S2)

Figure 5 can be thought of as a summary of the technical terms used in the title, abstract and claims of the 20,581 granted patents (S2). That said, this information is highly compressed and it is not a substitute for direct analysis of the actual claim language to examine the claim drafting strategies and scope of protection.

Review of patent grants in G06N10 (i.e. the CPC class for quantum computing) reveals that patent claims are directed to: (1) physical realisations of building blocks for quantum computers (e.g. quantum processors and components for manipulating qubits such qubit control); (2) quantum error correction (e.g. detection and prevention of errors using surface codes, magic state distillation, etc); (3) models quantum circuits and universal quantum computers; (4) applications of quantum algorithms (e.g. quantum optimisation, applied quantum Fourier or Hadamard transforms, etc); and (5) platforms for accessing, simulating, and programming quantum computers (e.g. cloud-based computing, platforms for simulating quantum systems, quantum programming interfaces).

4.6 Recently Granted Quantum Computing Patents

Table 3 lists the titles of the 20 most recent quantum computing patents granted (S8) and the respective assignees. According to 35 CFR 1.72 “The title of the invention may not exceed 500 characters in length […]” and pursuant to MPEP 606.01 if “the title is not descriptive of the invention claimed, the examiner should require the substitution of a new title that is clearly indicative of the invention to which the claims are directed”. Accordingly, the titles can be considered brief (≤500 characters) summaries of the claimed inventions. All these patents were granted in December 2021 and help illustrate the focus and range of subject matter of recent patents in quantum computing.

Table 3 Titles of the 20 most recent quantum computing patents granted (S8)

4.7 Top Quantum Computing Patent Owners

Table 4 lists top patent owners of quantum computing patents (S8) in the last 20 years. In the top 10 with more than 25 quantum computing patents, we find US global companies focused on hardware and software platforms (IBM, Microsoft, Google, Intel), a US defence technology company (Northrop Grumman), a California-based venture-backed ($200M) company founded in 2013 to provide scalable quantum processor technology based on superconducting chips (Rigetti), Honeywell International, and the US Government. Among the non-US companies, we find D-Wave Systems (Canada) focusing on specialized quantum annealing computers and Toshiba (Japan). The top university portfolios in quantum computing are MIT (US), Oxford (UK), Yale (US), Harvard (US), Caltech (US), Stanford (US), U. Maryland & Wisconsin, and Tech. Univ. Delft (Netherlands).

Table 4 Top patent owners of quantum computing patents (S8)

4.8 Citation Analysis and Dominant Patent Portfolios

Figure 6 shows the forward citation analysis of patent portfolios on quantum computing (S8) by assignee (ultimate patent owners). Forward citations are a measure of private value and a proxy for the potential social value of inventions.Footnote 18 Patents without established market values (e.g. in the absence of licensing agreements with negotiated royalty rates) are often valued based on forward citations and related valuation metrics. The patent portfolios with the highest number of forward citations are IBM, D-Wave, Northrop Grumman and Microsoft. The top university is MIT.

Fig. 6
figure 6

Forward citation analysis of patent portfolios on quantum computing (S8) by assignee

4.9 Global Geographical Distribution

Figure 7 shows top patent offices of issuance for quantum technology patents (S2). The US and China are the top countries of issuance, followed by Japan, South Korea, the EPO, Taiwan, Russia, Australia, Canada, and the UK. While the US continues to lead the field of quantum computing, China is arguably becoming the leader in quantum communications. This is remarkable since most of this Chinese growth in secure quantum communications has taken place in the last five years. Thus, it is expected that Chinese quantum networking and communications devices will soon be present in the global markets. As of 2021, China is now second ahead of Japan, Europe and Australia (Fig. 8).

Fig. 7
figure 7

Top patenting countries of issuance for quantum technology patents (Search ID: S2)

Fig. 8
figure 8

Top patenting countries for quantum computing (CPC G08N10). As of 2021, China is now second ahead of Japan, Europe and Australia

5 Innovation and Policy Implications

5.1 Growth of Quantum Technology Patents

The results of our landscape study show that the USPTO and EPO are now jointly granting around 2000 patents related to quantum technologies per year (S2). A total of 20,583 patents were granted from 2001 to 2021 with an overall CAGR of 15.23% over the last 20 years. Approximately 50% of these patent grants have taken place in the last five years (n = 10, 318), following ten years of relative stagnation between 2003 and 2013 (CAGR = 4.05%).

5.2 Patent Disclosures in the Public Domain versus Trade Secrets

From a public policy perspective, it is noteworthy that approximately 50% of the patent disclosures published in the last 20 years are already in the public domain.Footnote 19 This includes 4536 previously granted patents that have now expired, as well as the 14,830 patent applications that did not result in a granted patent. The teachings contained in these public disclosures are already freely available for society to use. Additionally, contrary to trade secrets these disclosures continue to raise the bar of patentability for follow-on applications, preventing applicants from patenting these inventions, and making it harder to obtain broad scope of protection for subsequent patent claims (i.e. effectively narrowing the scope of the claims that are ultimately granted). In fact, to satisfy the novelty and non-obviousness (inventive step) standards of patentability, applicants often have to give up part of the patent term for their newly filed patents by claiming priority to their earliest year of disclosure that supports their newly filed patent claims. Since the patent term is 20 years from the priority date (not the filing date), claiming priority to earlier patents from the applicant’s portfolio has the effect of reducing the effective term of the patent. As an example, of the 2037 quantum technology patents granted in 2021, 1642 have priority dates ranging from 2004 to 2018 (80.6%).

5.3 Continuum from Classical to Quantum Technology

The majority of the quantum technology patents granted in the last 20 years (S2) are related to nanostructures and solid-state devices and their applications (e.g. sensing and metrology). The nature of the claims and claim drafting practice is nearly indistinguishable from those found in the semiconductor patents that protect the types of inventions that gave rise to the computer and information technology revolution of the last 50 years. The number of transistors in dense integrated circuits (IC) has been doubling approximately every two years (following Moore’s law). As an illustration, the Intel 8008 processor introduced in 1974 contained 2500 transistors and used a MOS process of 10,000 nm, the 2005 Intel Pentium 4 contained 169K transistors (90 nm), the 2017 Intel Xeon contains 8 billion transistors (14 nm), and the 2021 Apple M1 Max has 57 billion transistors (5 nm). Thus, the building blocks of our classical processors now use MOS processes in the 5 nm scale. This miniaturisation has enabled the industry to increase the number of transistors in a chip (i.e. increasing the computational power and functionality of ICs) and their speed, while reducing the cost of production. A consequence of this scaling is that the classical devices used in “classical computers” exhibit quantum mechanical effects that need to be taken into account as part of the design and development of processors (CPU) and graphical processing units (GPU) for our current computing devices (e.g. phones, tablets, laptops, computers). In other words, quantum physical phenomena become unavoidable at the nanometre scale. For instance, at deep submicron and nanometre scales, MOSFETs exhibit quantum mechanical tunnelling from source to gate oxide due to the thickness of the oxide layers (approximately 2 nm), quantum mechanical tunnelling from source to drain when the channel lengths are less than 10 nm, and energy quantisation.

5.4 Defining Quantum Technology Patents

The increasing overlap between “classical” (e.g. sub-10 nm transistors used in current CPUs and GPUs) and “quantum” technologies raises the question: What is a quantum technology patent? In turn, the increasing difficulty in answering this foundational question – as the overlap between classical and quantum devices continues to expand – presents a challenge for proposals advocating for a sui generis quantum patent law regime. How do we define the material scope of this technical field? A few years ago a simple answer may have been to classify any invention operating at the scale of 10 nm or less (i.e. whose operation is governed by the laws of quantum mechanics whereby its design needs to take the quantum law principles into account) as a quantum technology patent. But such an answer would have resulted in the current CPUs and GPUs used in classical computers being categorised as quantum technologies. Similarly, answering this foundational question regarding the boundaries of the quantum technical field with more nuanced criteria such as classifying patents as “quantum technology patents” in the case that they contain claims directed to technologies that “exploit quantum principles” would result in the patents directed to many of our current mass-produced technologies, including semiconductor technologies, electron microscopes, lasers, and nuclear magnetic resonance (NMR), being within the scope of this definition. Finally, answering the question narrowly by limiting the definition to include only those patents disclosing inventions that directly harness the unique quantum mechanical phenomena such as superposition and entanglement to achieve the underlying principle of operation (i.e. those inventions whose operation directly exploits quantum phenomena) would likely result in claim draftsmanship. Patent attorneys would likely devise claim drafting strategies optimised to achieve the desired classification outcome to either avoid being classified as a quantum patent (if the sui generis quantum patent law regime is disadvantageous such as reduction in the patent term) or increase the chances of being classified within the quantum regime if the regime is beneficial for the applicant. Current patent law in the US (35 USC) and Europe (EPC) avoids such claim draftsmanship because the law does not discriminate between fields of technology (i.e. it is technology-neutral in theory, although arguably technology-specific in practice due to the examination guidelines for different technical fields). Furthermore, international minimum standards regarding the availability and scope of IP require that “patents shall be available for any inventions, whether products or processes, in all fields of technology, provided that they are new, involve an inventive step and are capable of industrial application” (Art. 27 TRIPS) and a common term of protection “the term of protection available shall not end before the expiration of a period of twenty years” (Art. 33 TRIPS). Thus, any proposals for reform would need to be mindful of applicable domestic and international harmonisation dimensions. Our results indicate that the USPTO has been the patent office of choice in the last 20 years (accounting for 77.96% of the S2 patents). Domestic or regional regimes that do not achieve international harmonisation would likely create distortions regarding the choice of patent offices for priority filings.

5.5 Growth in Quantum Computing Patents

Our search strategy (S2) identified 3042 quantum computing patent applications which resulted in 1603 granted patents. All these patents have been classified by the corresponding priority office under CPC class G06N10/00. This class is devoted to “quantum computers, i.e. computer systems based on quantum-mechanical phenomena” and contains a total of 1892 granted patents in the last 20 years. From 2001 to 2014, the field stayed highly niche and the number of granted patents in CPC G06N10/10 remained under 48 per year. That said, recently this field has been experiencing substantial growth, as the number of granted patents has increased from 37 in 2014 to 435 by 2021 (CAGR = 42.2%). Thus, our results are indicative of substantial interest, R&D efforts and investment in quantum computing.

5.6 Top Quantum Computer Owners

Notably, an examination of the top patent owners shows less concentration than we see in the classical computing and IT market. In addition to well-known and well-capitalised companies (IBM, Microsoft, Alphabet, Toshiba, Intel, HP), our results reveal that universities, public entities, and new ventures are featured among the top assignees. Strong patent protection is more critical for new entrants and quantum specialised SMEs (e.g. Rigetti, 1QB, D-Wave, MagicQ) than for large-cap IT (e.g. IBM, Google, Microsoft, HP), semiconductor electronics and telecom companies (e.g. Intel, Toshiba, Hitachi, NEC, ST) or defence and security firms (e.g. Northrop Grumman, Raytheon). It is noteworthy, for instance, that D-Wave systems ($1.2 billion valuation) holds more quantum computing patents than Google ($1.7 trillion market cap), Microsoft ($2.2 trillion market cap) or Intel ($220 billion market cap). Incumbent well-capitalised companies such as IBM, Google, Microsoft and Intel can fund their quantum computing R&D from their internal resources (e.g. strong balance sheet and extensive R&D budgets), but the new entrants have to fundraise from external investors largely based on the strength of their IP (given the deep tech nature of the field and the limited prospect of short-term revenues or profits). As an illustrative example, a new entrant (Rigetti) founded in 2013 and currently with around 130 employees holds one of the top 10 patent portfolios in quantum computing (ahead of Toshiba, Intel and HP), which enabled them to fundraise a total investment of $200 million and recently announced plans to become publicly traded on the NYSE at a valuation of approximately $1.5 billion.

5.7 Quantum/AI Hybrids

Both overlap between classical computing hardware and true parallel quantum 2.0 computing devices,Footnote 20 as well as synergies between AI and quantum computational methods result in a broad de facto category of quantum/AI hybrids. Just as heavier ex ante regulatory requirements for high-risk AI Systems as proposed by the novel EU AI Act demand a clear definition of High-Risk AI Systems,Footnote 21 including a dynamic list of examples, we can see an emerging need for clear definitions of what “quantum patents” and quantum/AI hybrid systems, products and services entail. Providing legal definitions and a material scope of these phenomena which cannot always be technology-neutral has consequences for innovation policy strategies regarding IP and antitrust laws.

Preferably, such descriptions and classifications should also be aligned with and embedded in quantum and AI-related horizontal and sector-specific regulations outside of IP and competition law, such as trade law, AI-enabled medical devices,Footnote 22 product safety regimes and liability rules. An approach of methodically linking core legal quantum and AI characterisations to other areas of the legal system, even implanting these novel doctrines in existing regulatory structures would foster regulatory coherence, complementarity and interpretability while avoiding fragmentation and forum shopping.Footnote 23 The uncharted cross-disciplinary field of quantum/AI hybrid systems represents a once in a lifetime chance to establish a globally harmonized groundwork of commonly practised and mutually agreed upon legal designations.Footnote 24

5.8 IP and Antitrust

The potential that key quantum computing patents may be owned by a handful of large companies and universities, could raise concerns towards technology transfer, equal access, and fair competition. As the IP toolkit is not designed to prevent or repair winner-takes-all effects or similar innovation distorting effects on its own, additional innovation interventions may be needed.Footnote 25 In the case of quantum computers, this means that antitrust laws and intellectual property laws will likely have to work together in concert to avoid exacerbating existing inequalities. In this light, we recommended earlier that “to encourage fair competition and correct market skewness, IP law needs to be complemented with antitrust law”.Footnote 26

Intellectual property law and antitrust law are intertwined, as encouraging sustainable innovation through the intellectual property system may increase as well as reduce competitiveness.Footnote 27 Therefore, what would be the strengths and limits of antitrust law in promoting technology and know-how transfer while also preventing monopolies and a quantum divide?

Foremost, as antitrust law prohibits monopolisation and coordinated refusals to give access to quantum technology deemed “essential” to compete in the market, patent holders may be required to provide FRAND licenses of essential quantum computational patents.Footnote 28 That said, when put into action to prevent market skewness, antitrust law obliges a proof of dominance, usually at least 50% market share, plus evidence of abuse of that market power.Footnote 29

Moreover, and contrary to popular belief, antitrust is not meant to avoid winner-takes-all effects per se. Antitrust rules do not discourage market dominance, provided it stems from fair and equitable competition.Footnote 30 In the contemporary, experimental state of the quantum computing art that is characterised by an emerging, incomplete marketplace, antitrust in its current form seems mostly ineffective to ensure fair competition. Ideally, IP and antitrust operate in unison to form healthy, quantum technology-specific relationships between patents, trade secrecy and competition law, while offering legal clarity about their interface with privacy and data protection regulations.Footnote 31,Footnote 32 This calls for elucidation and reform of both doctrines.

5.9 Quantum Technologies & Market Competition

Against this backdrop, several initiatives that could help address potential concerns regarding fair competition, knowledge sharing, declining market dynamism, and barriers for market entrants, can be imagined. First, the European Commission recently presented its Digital Services Act (DSA) package, as part of the overall European Digital Strategy.Footnote 33 The DSA aims to warrant that incumbent dominant market players can be challenged by both novel quantum start-ups and established competitors, so that the Single Market stays competitive and open to new ideas and innovations, and consumers will have something to choose from. The DSA will apply to the technical fields of quantum and AI as well. Second, in the US, the essential facilities doctrine could be used to open and revitalise the quantum computing ecosystem, by awarding access rights to facilities for which there is no realistic alternative present in the market.Footnote 34 Third, the introduction de lege ferenda of a pro-quantum-specific antitrust enforcement mechanism that safeguards the emerging quantum marketplace from dominant incumbents using their market power to distort developments in the quantum computing realm.Footnote 35 Fourth, free, government-funded access to quantum computational power in the cloud.

5.10 Government Funded Quantum Innovation Initiatives

That quantum technology is becoming a major competitive factor in the global “power game” becomes evident not only when analysing IPR-related pull mechanisms, but also when looking at recent push initiatives and investments. According to a Quantum Resources and Careers (QURECA) report,Footnote 36 national governments have invested over $25 billion into quantum computing research by mid-2021. Other reports claim that by September 2021 more than $1 billion in venture capital has been invested into the industry – more than the previous three years combined.Footnote 37

In order to keep up with increasing investments in quantum technology regions such as the US, China and India,Footnote 38 one of the most important European initiatives and a major facilitator and orchestrator of push incentives is the European Commission’s Quantum Technologies Flagship initiative.Footnote 39 It received full support from the EU Member States and commenced in 2018 with a total budget of €1 billion.Footnote 40 While the so-called “first quantum revolution” related to the “unearthing of the rules of the quantum realm, which led to the invention of tools such as lasers and transistors”,Footnote 41 the goal of this initiative is to support Europe in the highly competitive race for what has been referred to as the “second quantum revolution”. This includes the development of potentially life-changing technologies that could have an enormous impact on the market.Footnote 42 Consequently, the initiative focuses on the promotion of systems that are rather close to market maturity,Footnote 43 such as quantum-communication networks, ultrasensitive cameras and quantum simulators that could support the creation and design of new or improved materials.Footnote 44 Moreover, the initiative seeks to support less matured technologies with high market potential, such as general-purpose quantum computers and high-precision sensors that could be used in mobile phones.Footnote 45 To reach the optimal outcomes for the QT flagship, the EC has set up a High-Level Steering Committee (HLSC), which involves leaders from industries and academic institutions involved in quantum technology.Footnote 46

Importantly, EU Member State investments are increasing, sometimes exceeding EU investments. For example, in May 2021, Germany announced it will invest €2 billion in quantum computing and related technologies over a period of five years, “under a plan that dwarfs that of almost every other country, with the education and research ministry committing €1.1 billion by 2025 for R&D, while the economy ministry will contribute €878 million to develop applications”.Footnote 47 Even smaller EU Member States such as Denmark and The Netherlands, are investing heavily in quantum technologies, for example by establishing and funding new innovation centres and public private partnerships focusing on quantum technologies.Footnote 48 Meanwhile, in the United Kingdom, the government already invested £270 million in 2013 to set up the UK National Quantum Technologies Programme, with a quantum roadmap following in 2014.Footnote 49 The second phase of the programme commenced in 2019 and runs until 2025. Including industry contributions, the Quantum Technologies Programme has invested more than £1bn in the sector.

5.11 Governance of Quantum Technologies

These substantial, strategically targeted investments will have implications not only for the future of science, business, and society at large, but also for the national security (e.g. military) capabilities of nations. Hence, it is not surprising that international organisations are stressing the need for effective and equitable governance frameworks to best utilise the promises of quantum technology, as well as to address its future risks.Footnote 50 On 19 January 2022, the World Economic Forum (WEF) therefore published a report on Quantum Computing Governance Principles, which attempts to provide a first roadmap for these emerging issues across public and private sectors.Footnote 51 The principles have been co-authored by a “global multistakeholder community composed of quantum experts, emerging technology ethics and law experts, decision-makers and policymakers, social scientists and academics”.Footnote 52 The principles have been informed and guided by the coming hybrid model of classical, multi-cloud computing, to establish a quantum-computing adapted framework with best-practice principles and core values. The governance principles have been categorised into nine themes, i.e. (1) transformative capabilities, (2) access to hardware infrastructure, (3) open innovation, (4) creating awareness, (5) workforce development and capacity building, (6) cybersecurity, (7) privacy, (8) standardisation, and (9) sustainability. The recommendations about these topics pay particular attention to a set of seven core values that resemble many of the basic principles in computer, data and AI science, i.e. (1) common good, (2) accountability. (3) inclusiveness, (4) equitability, (5) nonmaleficence, (6) accessibility, and (7) transparency.Footnote 53 Future research and related initiatives will need to clarify how the design, content and enforcement of the legal and IPR framework directed to quantum technology, and in particular patents, relate to these governance goals, and must decide what role they can and should play in the future. It is expected that the principles will be operationalized by quantum technology impact assessments, as suggested in scholarly literature.Footnote 54

6 Discussion

These types of empirical results based on patent landscape studies should be taken into account when designing, assessing and proposing legal and regulatory reforms related to quantum technologies. For instance, the results our patent landscape study for quantum technologies suggest that the patent system is currently incentivising public disclosure in a technical field where trade secrets may be a commercially preferable IP option due to (1) the early-stage nature of many of these technologies, (2) the market structure, and (3) the potential business models. Similarly to pharma, the R&D of most quantum technologies is capital intensive and high-risk. In the pharma industry, patent rights are a key incentive because R&D for new drugs is costly, but once the composition is known, creating a substantially equivalent generic is orders of magnitude cheaper, especially for small molecule compounds. The asymmetry between the investment required to innovate (i.e. create a new safe and effective drug) versus imitate (i.e. create a substantially equivalent generic) is very large in pharmaceuticals. However, while most quantum technologies share with pharma the need for costly R&D, in general, quantum technology products are more difficult to reverse engineer (i.e. imitation is also costly). A fortiori, even absent quantum IP rights, or in case quantum patents would be waived, pledged or nationalized, recreating complicated machines such as quantum computers would require significant hi-tech facilities, monetary resources and know-how, in particular cleanrooms, hi-quality production lines, supply chains and an experienced, skilled workforce.Footnote 55

Some quantum technologies such as quantum computing hardware are likely to be centralised in physically secured facilities with customers accessing them over the cloud with permissions determined through role-based access control by the quantum computing service providers. Thus, since the product is not “publicly available” as in the case of pharmaceutical drugs, it is much easier to protect it through trade secrets. The expected business model for quantum computing is likely to resemble the current model for cloud-based computing dominated by Amazon/AWS (33%), Microsoft Azure (21%), Google Could (10%), Alibaba Cloud (6%) and IBM Cloud (4%).

In such an environment, incumbents are incentivised to keep their technical breakthroughs as trade secrets since (1) trade secrets are not timebound; (2) they have other sources of competitive advantage derived from their established market position (e.g. the top two providers have over 50% of the total market share in cloud-computing); and (3) they can fund the quantum R&D from internal sources.

However, when new entrants (e.g. D-Wave, Rigetti) in quantum computing start disclosing their inventions and obtaining patent protection, this creates a competitive risk for incumbents since these specialised new entrants may be able to (1) obtain more valuable patents (i.e. with broader scope protection) during the early stage of the field because there is a limited number of inventions that can be used by patent examiners to limit the scope of the proposed claims based on novelty, inventive step (non-obviousness) and sufficiency of disclosure (i.e. written description, enablement and best mode); and (2) leverage the patent portfolio and other IPRs to raise funding from venture and growth capital in order to develop specialised quantum technology commercial solutions that could pose a disruptive innovation threat to the current technology incumbents. Consequently, incumbents that may otherwise prefer to keep their key inventions as trade secretsFootnote 56 as opposed to pursuing patent protection (especially for inventions whose expected ROI from commercial exploitation is 10 to 20 years in the future) are encouraged to also disclose their inventions through the patent system at this early stage. Compared to trade secrets, these dynamics are generally more beneficial to society by promoting innovationFootnote 57 since everyone benefits from the public disclosure and all patents eventually become part of the public domain (i.e. within 20 years of the priority date).

From a public policy perspective, it is noteworthy that nearly all of the 1892 quantum computing patents that have been granted over the last 20 years will be in the public domain by the time the quantum computing market reaches a size comparable to the size of the current classical could-computing market. In fact, already 9.23% of the granted patent applications classified in CPC G06N10* (quantum computing) are now expired. Large-cap incumbents with leading market positions are likely to benefit and favour proposals that weaken patent rights while new quantum technology entrants are more likely to benefit from stronger patent rights, since they have a greater need for patents both (1) as a source of competitive advantage during their 10–20-year incubation period and (2) to help raise patient capital to fund long term R&D efforts. Accordingly, most quantum technology start-ups that are hardware-oriented are likely more dependent on strong patent rights as in the case of pharma and biotech ventures.

Innovative developments in quantum computing also take place in secrecy rather than in the open. Based on leading academic papers and breakthrough qubit experimentsFootnote 58 one could argue that China is now on par with the US in the quantum computing field, but according to our study, this does not yet translate into a higher number of USPTO/EPO patents, or patent quality, when compared to US companies and universities. That said, as of 2021 China is now second, ahead of Japan and Europe in patent filings. Reasons for keeping vital information a trade secret or a state secret might be anthropogenic risk assessment, the competition for technological dominance, or national security concerns. In general, it is easier not to disclose information, or declare R&D a secret in a government funding scenario.Footnote 59

Instead of being technology-neutral as in the case of patent law,Footnote 60 push incentives, governance and regulatory sandboxes can be technology-specific in application, and play different roles across different industries, sectors and technologies, such as quantum technology, biotechnology, nanotechnology and artificial intelligence applied in healthcare, with different incentive, reward and disclosure of information outcomes as we deal with different actors, products and services, even within the quantum domain.Footnote 61 IPRs can play a central role in governing technologies.Footnote 62 Thus, to help regulators obtain a complete overview of the quantum 2.0 field, sector and domain-specific empirical research should clarify the role that IPRs and push incentives play in application areas such as quantum sensing, simulation,Footnote 63 computation and communications. Moreover, we should learn from history and compare the patent concentration levels in the current early-stage market to those in the early semiconductor landscape in the 1950s–1980s, with the main transistor patents owned by companies such as GE, Fairchild, Bells Labs and Texas Instruments, by performing quantitative historical exegeses. In addition, empirical methods should be applied to assess the specific types of IP rights incumbents, and quantum start-ups are using in their value creation strategies, preferably per market sector and quantum subdomain.Footnote 64 These are important and exciting opportunities for further quantitative and multimethod research.

7 Conclusion

The results of our general patent landscape study show that (1) the USPTO and EPO are currently granting around 2000 patents per year related to quantum technologies; (2) the overall CAGR over the last 20 years has been 15.23%; (3) approximately 50% of the patent disclosures published in the last 20 years are already in the public domain (including 4536 granted patents that have now expired); (4) the majority of the granted patents relate to nanostructures, solid state devices and their application; (5) the nature of the claims and claim drafting practice is nearly indistinguishable from those found in the semiconductor patents that gave rise to the information technology revolution; (6) the increasing overlap between classical (e.g. sub-10 nm transistors used in current CPUs and GPU) and quantum technologies presents a challenge for proposals advocating for a sui generis quantum patent law regime; (7) after a period of limited patenting activity (2001–2014), the field of quantum computing has experienced substantial growth recently with the number of granted patents increasing from 37 in 2014 to 435 by 2021 (GACR=42/2%); (8) there is currently less concentration among the top quantum computing patent owners than in classical computing and IT markets; (9) in addition to well-known large cap companies (e.g. IBM, Microsoft, Alphabet/Google, Toshiba, Intel, HP), specialised SMEs (e.g. D-Wave Systems), new ventures (e.g. Rigetti), universities (e.g. MIT, Oxford, Yale, Harvard, Caltech, Stanford, U. Delft, U. California) and government entities (e.g. US Gov, Japan Science & Tech, Gov. Abu Dhabi, Korea Elect Rest. Inst.) are featured among the top assignees; and (10) the USPTO has been the patent office of choice in the last 20 years, accounting for 77.96% for the granted quantum technology patens by the EPO and USPTO.

The results of the patent landscape study suggest that the patent system is currently incentivising public disclosure in “quantum computing”. This is in the public interest since trade secrets may be a commercially preferable IP option due to the early-stage nature of many of these technologies, the market structure, and the potential business models (e.g. quantum cloud-computing as a service). From a public policy perspective, large-cap incumbents with leading market positions are more likely to benefit and favour proposals that weaken patent rights, while new quantum technology entrants are more likely to benefit from stronger patent rights. Push incentives, governance and regulatory sandboxes can be technology-specific in application, and play different roles across different industries, sectors and technologies, such as quantum technology and medical AI, with different incentive, reward and disclosure of information outcomes.