Negotiating Intellectual Property Rights

FUBU didn’t distribute billions worth of product because it had the most cutting-edge designs. It grew because of brand identity, and more importantly… licensing, licensing, licensing! We licensed everything: Boots. Bags. Women’s. Entire territories- Europe, Asia. Even retail operations. And when I look back at my most successful companies from Shark Tank, I see the same pattern: Bombas- licensing with brands like Sesame Street, Peanuts, and Wicked. Sun Staches- built a massive business through licensing collaborations. This is what I coach my CEOs every day. Licensing is the ultimate OPM. Here’s how to think about it: If you’re taking in a license- You’re tapping into a globally recognized brand. They’ve already done the marketing. You don’t need to spend millions on awareness. If you’re putting out a license- Someone else is doing the heavy lifting: Manufacturing. Distribution. Marketing. Staffing. Licensing is one of the most misunderstood, and most powerful wealth-building strategies in business.

You might be surprised who really owns your local Shell or BP station… Over the past decade, major oil companies have been quietly reshaping their downstream model, shifting from owning and operating fuel stations to Licensing their brands and outsourcing daily operations. The logic is clear: => Focus capital expensitures on higher-return, lower-carbon investments, while keeping strong brand presence through partners. Here are some notable examples: - Shell → Vivo Energy (Africa): Shell exited operations in over 20 African countries, maintaining its brand via long-term licensing. Vivo Energy now runs 2,000+ Shell-branded stations across the continent. - Shell → Vitol (Australia): Shell sold its entire Australian downstream business to Vitol for A$2.9 billion, including 870 retail sites and the Geelong refinery now operating under the Viva Energy brand, still selling Shell fuels under license. - BP (Austria, Netherlands, Switzerland): Gradually selling or franchising retail networks, retaining branding and supply agreements - Shell → ST1 (Norway & Finland): Sold its Nordic retail business but continues under the Shell brand via license across 400+ sites. - ExxonMobil → EG Group & DCC Energy (Europe): Esso-branded stations are now mostly run by licensees and partners, not ExxonMobil itself. - TotalEnergies (Belgium, Germany, Italy): Partnered with Q8 (Kuwait Petroleum) and MOL Group for retail operations, while keeping brand or supply rights. - Chevron (Asia, Latin America): Chevron Caltex continues through franchise and brand-license models. - Eni (Italy, Greece): Converted many stations to franchise/dealer-operated sites. Why this matters 1. Majors are redeploying capital toward energy transition and high-margin trading. 2. Brand without burden: They keep consumer visibility, supply, and loyalty networks without the cost of operating thousands of stations. 3. Local empowerment: Regional and local players are now driving transformation and growth. They execute locally while operating under a global brand umbrella. Does this resonate with you? Does this apply to your market?

Open-source FEM/CFD libraries are free to use. Until you turn your research into a product. You can indeed install and use all of them freely for research or internal work. However, the situation changes once you build and distribute a product that uses these libraries. What matters is the license. Common licenses in the FEM/CFD ecosystem include: 1. MIT (e.g. PyVista) 2. BSD (e.g. ParaView, PETSc, MFEM, NumPy) 3. LGPL (e.g. FEniCSx, deal.II, Netgen, MOOSE) 4. GPL (e.g. OpenFOAM, CalculiX, Code_Aster, Gmsh) Key points: - You are allowed to integrate MIT, BSD, and LGPL software into your own products, including commercial ones - If you use GPL components and distribute your software, you must open-source your code or obtain a special commercial license. This usually blocks closed-source desktop applications and plugins - GPL (unlike AGPL) does not restrict the use of the library on the server side in SaaS or cloud platforms, since the code is not distributed This is a simplified overview. For details, always consult the text of the corresponding license. Has anyone encountered legal issues after embedding open-source libraries into own products?

🔥AI licenses are similar to open source licenses. The Devil is in the details, so take a hard look. Opinion: ✔️There is a “hurry up and we can fix it later” approach to AI related solutions ✔️But, hearing the “focus on the revenue not the risk” is a lesson that we learned before, and it ended with the wave of indefensible infringement and copyright lawsuits for illegally using Open Source Software(OSS) ➡️These scenarios both involve viral licenses. And we have seen this movie before. It ends badly ✔️Originally, most looked at OSS as free software for use by anyone. OSS came with licenses, but no one paid attention to them ✔️That was a popular and expensive misconception. Almost all OSS has a license governing use, and sometimes that license also set limitations for software integration or allows third party data for “business use” ✔️Similarly, many popular AI/LLM and other products that you build upon have commercial licenses,TOUs etc. that govern how your data and technology is owned and used by third parties ✔️Much like an OSS license, if those terms are violated OR unanticipated, the potential consequences could include losing your data rights and your technology AND facing lawsuits from your clients after their data was compromised ✔️Here is an example from the last rodeo with OSS. In 9/2007, the SFC filed infringement lawsuits against the principal developers of BusyBox-claiming violations of its GNU license(GPLv2) and had $$$ a month later ✔️The Xterasys lawsuit immediately followed. Xterasys had to pay $$, stop shipping its products, and MAKE THEIR COMPLETE SOURCE CODE PUBLICLY AVAILABLE SO ANYONE CAN COPY IT as this was part of the viral GPL license they had integrated into the platform(s) ✔️Now we are at full circle with a class action lawsuit in 2023 against Microsoft, GitHub and OpenAI. The claim is that Copilot was illegally “trained” with code from GitHub repositories, violated licensing terms, and therefore Copilot unlawfully reproduces their code ✔️In comparison to data use and AI tools, OSS licenses are simpler to mitigate because they relate to a discrete/definable part of a platform ➡️With AI, training data is ambiguous, and an infringement/ breach claim against your AI platform could mean that you are breaking unanticipated legal obligations ➡️There are no quick fixes. There are right ways and wrong ways to use data sources, and they are situational 🔥The companies who can keep up the with changing regulations/privacy laws and IP challenges etc.-or better yet anticipate them and pivot accordingly will rise to the top. 🔥 “Perfect is the enemy of good” has its merits, and I agree. But if you believe that this is a minor issue, and that ensuring compliance is a perfectionist pursuit, then perhaps we should heed Albert Einstein’s words when he said that “whoever is careless with the truth in small matters cannot be trusted with important matters” #startups #strategy #management #law #education Bold Women Network

🆙 Bumping an #opensource package version can fix #vulnerabilities. 🧾 However, such change can also introduce a LICENSE RISK. For example, the package `ua-parser-js` has 2 versions with recent releases - version 1.0.39 is licensed as `MIT`, and version 2.0.0 is licensed under `AGPLv3`. 🤖 When a #developer, your #appsec #platform, or an AI code assistant, suggests code changes to bump a package, it is important to flag such license review as part of the development lifecycle. 🌻 Arnica does it natively within the developer workflow, on every code push, on every feature branch, providing the opportunity to address this risk before the pull request. Additional observations: 1. Too restrictive change: the `MIT` licensed version has been downloaded 38X times more than the `AGPLv3` version. 2. Maintainers merged #dependabot pull requests with this bump, likely inadvertently. 895 projects depend on the `MIT` version, while only 36 depend on the `AGPLv3` version directly. 3. `AGPLv3` has a strong copyleft requirement, which means that any software incorporates this package must also be licensed similarly. It can take companies out of business if not handled properly. 🌻 Arnica comes pre-configured with a policy that classifies such license as a high risk.

When it comes to using software systems, it's critical to ensure that you *actually* have all rights necessary to "operate" the system. That's why the patent license in the Apache 2.0 license was such an important addition - because even with the license to the source or binary, you might still run into *other* IP issues. Similarly, truly open source licenses to AI systems *must also consider all IP rights necessary to operate them* - which obviously must include sufficient rights to use or reproduce the training inputs that become verbatim or "derivative" outputs at inference time. After decades of explaining issues like these to the public, it is truly bizarre to watch organizations like the OSI "forget" everything. The result of this all? "Open-washed" models will obtain a liability shield in jurisdictions like the EU, leaving you - the user - responsible for their original sins. As usual, regulatory confusion and regulatory capture only seem to expose the public to more risk, not less...

2010: Marshall licensed its name to a Swedish startup. 2023: The startup owned the Marshall brand. A masterclass in brand value creation. When Zound Industries licensed the Marshall name in 2010, it it wasn’t a big thing. A consumer-audio company paying to use a rock-icon logo on headphones and Bluetooth speakers. But Zound were razor sharp. Marshall wasn’t just an amplifier company - it was a set of codes: black leather, gold knobs, and a handwritten logo that symbolised rebellion, craftsmanship, and sound with attitude. Instead of just coping that history, Zound applied it. Every new product was designed to feel like it belonged on stage, even when it lived in a kitchen or hotel room. That’s what I call Applied Heritage: Using the codes of history to design the feeling of legacy. In new products. The results were extraordinary. Marshall evolved from instrument maker to billion-euro lifestyle brand. In 2023, Zound fully acquired Marshall. And this year, the new Marshall Group sold a majority stake for €1.1 billion. The Marshall family, still owning roughly 24 %, became multi-hundred-millionaires. They didn’t protect their past. They letsomeone else apply it. I’ve seen this pattern before. At Bang & Olufsen, we built B&O Play for the same reason A small, independent company that moved faster, reached younger audiences, and reinterpreted B&O’s design DNA for a new era. That’s the paradox of heritage brands. Relevance doesn’t come from preserving what you were. It comes from re-applying for today. 

There is a virus in your software. Your developers didn't mean to plant it. They just wanted to move fast. They used a piece of Open Source code with a "Copyleft" license. To a buyer’s technical team this is toxic waste. If you use code licensed under GPL or AGPL it can be "infectious." It can legally force you to make your entire proprietary product open source. The moment a buyer sees this they see a zero dollar valuation. They see a total loss of Intellectual Property. The Decontamination Process: ☢️ Run an Open Source Software (OSS) audit immediately. ☢️ Identify every library and component your team has used. ☢️ Flag any license that requires "reciprocity." ☢️ Replace the toxic code with "Permissive" licenses like MIT or Apache. Cleaning up your code is much cheaper than losing a deal. Don't wait for the buyer to run the scan. Be the one who proves the tech is clean. We have an Open Source Component List template in our audit kit. Start your technical cleanup before the inspectors arrive. #SaaS #SoftwareDevelopment #MandA #TechExit #Livmo #IntellectualProperty

AI and Open: My Head Hurts All these discussions about "openness" in AI are spiraling into debates over words with oceans of interpretations. Instead of getting lost in semantics, how about we start the conversation with what we do know? "Open" has no agreed definition, but open source does—an OSI-approved license that also guarantees the four freedoms of the OSD: the right to use, study, modify, and distribute software. If we’re serious about AI openness, can we start by aligning on what is tangible and already has an agreed-upon definition? At a minimum, openness in AI should ensure these same freedoms for model weights and architectures. Training code and data are the most contentious pieces—competitively, legally, ethically, and technically. Rather than letting that complexity stall progress, how about the AI creators, whether they be companies, foundations, or individuals, focus first on making weights and models freely available under clear, enforceable, and OSI-approved licenses? Once we establish those foundations, we can tackle the harder problem: how to responsibly open AI training code and data while addressing privacy, consent, and ownership challenges. Neither Meta’s Llama 3.1 Community License nor NVIDIA’s Open Model License is OSI-approved. It would be an interesting exercise to analyze whether an OSI-approved license could meet their needs—or to understand why unrestricted openness in this domain is commercially non-viable.

Unlocking the Value of Your Intellectual Property through Licensing. In today's fast-paced economic landscape, Intellectual Property (IP) licensing has become a vital strategy for individuals and organisations seeking to capitalise on their innovations. I have seen firsthand the benefits of strategic IP licensing in maximizing the value of one’s intellectual property while ensuring its protection. By licensing their IP, businesses can generate new revenue streams, expand their market reach, and boost brand visibility without sacrificing ownership or incurring the costs associated with full commercialisation. IP licensing offers a versatile approach to growing and protecting valuable assets, whether through patents, copyrights, trademarks, or franchising. By licensing their IP to third parties for specific products and services, businesses can leverage their brand's strength while maintaining control over its use. This generates ongoing income through royalties without the licensor needing to produce or distribute the IP-based product themselves, reducing the risks associated with market entry, product development, or scaling. However, while IP licensing can be a powerful tool to unlock the potential of one’s intellectual property, IP valuation is crucial to a successful licensing arrangement. IP valuation determines the monetary value of intellectual property assets, such as patents, trademarks, copyrights, and trade secrets, and is essential in IP-based transactions and commercial agreements. By carefully selecting a licensee, negotiating strong agreements, and actively managing the licensing arrangement, companies can turn their IP assets into key drivers of growth, long-term success, and innovation. Individuals and businesses are encouraged to seek expert legal advice on licensing strategies and explore the possibilities of licensing to maximise the value of their intellectual property. #intellectualpropertylicensing #innovation #ipvaluation #IP #sbp