Semiconductor Industry: Commercial & IP Considerations
With the recent sale of UK semiconductor unicorn Graphcore to Softbank Group Corp. we look at the commercial and IP considerations which are so important to implement at the start up stages of any semiconductor technology business.
In this first article of our semiconductor series, we look at how to create and protect IP assets, the commercial considerations and how these are often unique to this sector. Getting this right is crucial to creating and preserving the value of any semiconductor business.
What is a semiconductor?
Semiconductors are the electronic component in most hardware technology – they exist in smart devices, mobile handsets through to accelerators for AI and machine learning technology. Small but mighty – the digital economy would be unable to function without them.
What is the current size of the semiconductor market?
Global revenue from the semiconductor market hit $601.7 billion in 2022 according to Gartner. Whilst 2022 was a record breaking year, 2023 saw a slight decline in revenues. However, Gartner and other analysts predict the sector will grow significantly between now and the end of this decade.
Key sectors influencing the growth in the semiconductor market are AI, data centres and storage, wired and wireless infrastructure, industrial electronics, automotive and consumer electronics.
Underpinning this growth is the requirement to create, protect and monetise intellectual property rights (IP) and associated commercial rights.
Semiconductor IP: Intellectual Property Rights
Semiconductors and the products which embed them, or use them on the edge, may be protected by a range of IP:
- copyright (written designs, instructions, layouts, software and applications etc);
- patents (the high level functional design which could be protected by one or a family of patents);
- trademarks (the brand name and design);
- trade secrets (the way the circuit operates, how any wafers and chips are made and tested, the source code to any proprietary software stacks);
- designs (the chip layout and circuit design);
- topography (chip layout and circuit design arrangements);
- databases (databases which interoperate with products).
IP may also exist in any blocks or cores which are reusable units of logic, cells or integrated circuit layouts.
Semiconductor Patents Review
Unsurprisingly, given the revenue growth in the semiconductor market, related patent applications have increased to 69,190 global patents in 2023, up 59% from 43,380 five years ago.
UK patents in the semiconductor sector accounted for just 0.3% of global semiconductor patents in 2022, falling behind Europe. The global semiconductor patent race is dominated by the US and Asia; the number of patents within these regions has been supercharged by the significant government investment in these regions.
Semiconductor companies which do not have the benefit of this level of government support, need to consider strategic investment in patent portfolios along with technology licensing and cross licensing to stay relevant and competitive.
Incentivising staff early in the design stage, and working with specialist patent attorneys to develop a patent strategy which targets key regions, will ensure a more limited budget has greater impact. So often we work with businesses where the very early technology has not been protected robustly, either due to a lack of time, funds or poor advice.
Semiconductor Research & Development
Research and Development (R&D) activity is often the acorn from which many semiconductor trees grow.
This can include R&D projects with universities, academic and research organisations.
The UK Semiconductor Advisory Panel is focused on an initiative to support commercial R&D and SME growth looking to facilitate an open foundry ecosystem, better access to chip design tools and prototyping facilities. This initiative is also seen in many other countries internationally.
Crucial to any R&D is to understand the objectives and expectations of each party and ensure an appropriate R&D agreement is in place. An R&D agreement should articulate the obligations of all parties, funding commitments and the ownership of foreground and background IP. It should also tackle the treatment of any jointly owned IP; often this can be complex when there are multiple parties, deliverables are not clear at the outset and the IP is not fully known. Taking a proportionate approach is also crucial as legal budgets are often small at this stage. If equity is being provided in return for contribution, this needs to be clearly identified along with tax implications; so as not to impact on future investment opportunities.
Semiconductor Early Access & Piloting
Early access is a common phase for testing and piloting new products. Semiconductor early access enables customers’ to pilot semiconductors in customer devices and environments. It also enables important customer feedback and assesses the appetite for new products. The costs of making available early access technology products can be significant, building and shipping a small amount of prototype units worldwide and managing feedback requires a dedicated team and the right commercial approach.
Whilst it is easy to get excited about the opportunity to get technology into the hands of early access customers; it’s important to think this phase of product development through and consider sensible contractual bumpers.
Issues with Semiconductor Early Access and Piloting
Issues to consider include:
- who will own IP in any feedback?
- what if the customer seeks to benchmark - are there any controls or conditions on benchmarking and onward public disclosure of benchmarking results?
- how will units be delivered, set up and if needed, how will they be returned?
- who will pay for the early access stage?
An Early Access agreement is required to protect IP rights, maintain confidentiality around feedback and control other commercial terms of any access such as benchmarking.
Collaboration within the Semiconductor Market
Collaboration opportunities arise at many stages, during R&D, Early Access and in tandem with supply negotiations; collaboration will vary depending on the technology and customer requirements.
As with the other earlier stages (Early Access and R&D) entering into a suitable agreement is crucial but can be more complex as this relationship will be more future looking. A Collaboration agreement may cover an exclusive flavour of technology, customised for a specific customer; it may create a feature which is switched off for certain fields of use or customers, and it may involve third party IP. Collaboration agreements are thereby complex and unique and often involve cross IP licences and mechanisms for IP pools and allocation around joint IP.
Commercially, for scale ups the timing of these relationships can be a blessing and a curse. On the one hand, collaborating with a customer can be great for investment (investors love to see collaboration - it gives confidence that the technology has legs and hopefully a future supply contract); but it can also be all consuming for the software and development teams which can detract from the overall product roadmap and de-prioritise other customers.
Semiconductor Supply Chain
At the supply stage, the focus shifts to the commercial considerations and ensuring supply chain contracts are in place and in good shape.
Technology Licence Agreements
In some instances discussions with customers lead to a Technology Licence Agreement (TLA) where the technology is suitable for a field of use, but which is not core to the founder companies product roadmap and strategy. This can be helpful for startups when there is a small sales team or the business does not have the capacity to fund product support and development in a specific field of use.
A TLA can licence the IP and some/all of the technology for a specific field of use (such as defence), in a specific region (e.g. US) which can generate licence fees and/or royalty income with less investment and allow the founder company to focus on its core market and customers.
TLA’s are variable in nature and templates should be avoided; key considerations are around the scope and exclusive nature of any licence; term and any conditions; deliverables, foreground and background IP, cross licencing; licence fees and minimum royalties.
Open Source Software
Many semiconductor technologies rely on opensource software within associated software stacks which are licensed alongside the supply of chips, devices or semiconductor IP blocks. Open source software may also be in standalone libraries, code snippets, functions, frameworks or entire application provided to customers.
Licences for open source software can be broken down into 5 categories:
Public Domain
Public domain software is protected by copyright but typically has been released for public use on a free of charge basis with little or no restrictions. Clearly just because there are little or no restrictions does not mean the code base is secure or of robust quality to be included as part of a commercial product.
Permissive
Software released on a permissive basis typically enable users to use and distribute the software with few conditions or restrictions. Examples are the MIT licences.
Copyleft
Software released on a copyleft basis typically allows for modifications and distribution of the code but there are various flavours, so tread carefully here. Some licence terms oblige you to release the code you have modified on the same licence terms, others extend to the entire application to which the modification applies so it is easy to inadvertently “open up” software unexpectedly. Examples include the GNU AGPL v3, GNU LGPLv3 and GNU GPL v3.
Weak or partial copyleft
Software licensed on a weak/partial copyleft aims to allow linking to open source libraries without the obligation of a fully copyleft type of obligation. This is useful when you want to dynamically link to a GNU LGPL licensed library entirely, but distribute under any licence (even a proprietary one) with minimal requirements. Static linking and modifying the library is more complicated, and using LGPL licensed files can trigger copyleft obligations. Other weak copyleft licenses include MPL and Eclipse.
Dual
This covers software which has been made available by the copyright owner on an open source basis but also under different licenses for specific user types; this is common for developer situations and it’s important that code developed in a developer trial is not incorporated into a commercial product without upgrading to the full commercial license.
Open Source Software Best Practice
There is often a misconception that use of “free” open source software carries a low IP risk, but businesses need to pay attention to the licence terms for any open source within their supply chain. Customers will want a detailed breakdown of all open source software (with associated licence terms) used in associated software stacks etc. and this will need to be maintained and regularly updated.
Managing the licence arrangements and ensuring developers follow these terms is integral to protecting proprietary IP within any software roadmap from inadvertent leaks and licence infringement. Good internal governance here is also important to give early access and strategic customers confidence to invest in collaborations.
Longer term, being able to articulate the software used within any product portfolio and having it well organised is also necessary for smooth due diligence at any sale/investment stage.
Finally, considering what support will be available for these open source components within any wider product offering is critical. Assessing the risk of supporting open source code, at scale must be factored into any roadmap. All commercial customers will expect support and any business which deploys Open Source Software, needs to consider the support available and the costs and risks going forward.
Standard Essential Patents (SEP’s)
A standard essential patent is a patent which is essential to the operation of a defined technical standard. Standards can cover ideas, products, services or ways of operating, and are critical to making sure technologies can interact and work together. We come across these commonly in mobile phones, wireless connectivity, navigation systems and smart meters.
In many cases, standards require the use of specific technologies protected by patents known as SEP’s. Licensing of SEP’s for use in products is subject to specific licence arrangements usually via standards organisations. It is imperative that if SEP’s are relevant, businesses comply with the required licence terms and any additional contractual rules applicable to the relevant standards body. The membership arrangements of certain bodies can be complex and involve cross patent licensing and defence arrangements so these should always be reviewed by appropriate legal advisors.
Balancing these commercial and IP considerations from the outset will be crucial to maximising growth and opportunities from those sectors which are going to need semiconductors well into the future.
Rebecca Steer advises semiconductor start and scale ups and advised unicorn Graphcore on its spin out from XMOS. She has acted as interim General Counsel for a number of semiconductor businesses, and supports technology businesses on a range of legal matters.
