Over 250 senior IP professionals – including representatives from major licensors and licensees – attended Sisvel Connect 2026. The conference at the heart of the event provided plenty of food for thought

By Jacob Schindler and Joff Wild

This year’s Sisvel Connect took place in Barcelona from 15–17 September. Over the course of the three days, there was a series of constructive patent owner meetings, highly enjoyable social events and a thought-provoking conference, “The Global FRAND Market Today and Tomorrow”, held under the Chatham House rule.

Although we cannot tell you who said what, we can share some of the themes that emerged from the conference. Below we highlight some of the key talking points.

We would like to thank the 250-plus friends, partners and contacts who joined us at the event for making the three days so special.

Standardisation trends favour incumbent SEP leaders

Attendees were presented with a comprehensive overview of the state of play for various connectivity and media standards. The 6G development process has reached a pivotal stage: at the time of the conference, a 3GPP plenary taking place in Madrid was being asked to ratify key decisions made at the working group level in August. Since the start of 2026, a clearer picture of the technology has emerged. The consensus seems to be that support for AI applications and deeper integration of sensing capabilities will be two of the biggest differentiators. Meanwhile, the underlying connectivity tech will retain significant overlap with 5G as it evolves. That’s generally good news for incumbents in cellular SEPs.

The story in Wi-Fi is broadly similar. Wi-Fi 8, which is nearing deployment, prioritises ultra-high reliability over raw speed; while several big players have sketched visions for its follow-up, Wi-Fi 9, in which AI and sensing will play a major role. This underlines the growing trend of convergence between the two major families of connectivity standards. There has been a surge of interest in Wi-Fi standards-setting but again, those with foundational existing portfolios look set to benefit the most as the technology evolves along a more gradual curve.

On the video side of the equation, the release of AV2 is one of the most significant recent developments. Understanding the direction of travel here demands substantial legwork compared with following the progress of truly open standards bodies. Sisvel Tech engineers have been studying the specification for well over a year, and AV2 is just as indebted as its predecessor to foundational innovations in video coding developed outside the Alliance for Open Media consortium. One of the more interesting titbits shared is that AV1/AV2 developers increasingly acknowledge its place within the broader multi-decade story of codec development rather than selling it as something wholly novel. Sisvel’s efforts to foster an AV1 pool and to educate the market about the specification have no doubt played a role in this.

Licensing transparency needs practical limits

“What is our competitor paying?” This is the first question that a licensee-side negotiator will likely be asked when pitching a deal to management. Many of the same executives, though, resist disclosing even the existence of an agreement to the market.

Patent pools are a major source of transparency in the market, with many programmes publishing patent lists, royalty rates and detailed licensing terms. Most announce all of their licensees too – provided that they can secure permission to do so. In the world of bilateral licensing, things are often considerably murkier.

Delegates were told that practical steps for cutting through these tensions include anonymised figures; controlled access for advisers through mediation, arbitration or court proceedings; and reciprocal sharing of otherwise confidential information.

However, there does come a point at which requests for more transparency stop facilitating a deal and stray into hold-out territory – especially when the material already provided should suffice to establish the credibility of an offer. The objective should be enough transparency to give both sides the confidence to proceed. That requires licensors to provide meaningful FRAND justification; but it also requires licensees to recognise when legitimate questions have been answered.

Strong enforcement matters

There are no cookie-cutter solutions in licensing negotiations; each one is different. However, although most deals are concluded amicably, they are all inked in the shadow of litigation. The only variable is the depth of the darkness cast by this threat.

Frequently, it is the lightest shade – a far-off possibility as the sun shines and talks progress smoothly at a steady pace. Other times, though, the night is fast approaching: disagreements are profound, the sides are far apart and the same issues are hashed out for years with no resolution in sight.

Across the market, there is an understanding that patents are a right to exclude. Without viable means to exercise that right, there is no incentive for those practising on an invention to seek permission to do so. That is a simple business truth.

This makes strong enforcement a matter of considerable interest – and it was a subject that came up in several conference sessions. For licensors, it is important to have a credible threat. In most cases, that will include the ability to secure injunctive relief. Licensees, on the other hand, argue for proportionality and claim that in many cases, monetary recompense for infringement is enough.

Although new litigation forums are emerging – Brazil, India and several countries in Africa were all mentioned – the two that consensus suggests are the most important are Germany and the Unified Patent Court. The combination of speed, cost and remedy that these venues afford create strong dealmaking incentives.

Five questions to ask before buying an AI tool

IP teams are being pitched an ever-growing range of AI products. The message from one panel was not to dismiss these tools – many already perform valuable tasks and their capabilities are improving quickly. But healthy scepticism was strongly advised. Here are five questions worth asking before you buy.

“What is the output based on and can we check it?”

Outputs should be linked to sources that can be inspected by a human expert. The panel heard examples of hallucinations including invented claim elements, references to products that did not exist and even the combination of several products into a single fictional system.

“Does the product deliver rigour or merely plausibility at scale?”

There is still a significant gap between producing large volumes of plausible analysis and delivering something approaching attorney-quality work, attendees heard. Insights that at first glance seem impressive often prove technically or legally subpar on close examination by an expert.

“What happens to the information that we put in and get out?”

For lawyers and IP professionals, confidentiality and privilege are fundamental. Buyers must understand whether information entered into a product remains protected and what contractual or technical safeguards apply. Who owns the output and whether privilege applies are among the key considerations for the output or work product.

“Does the provider understand the work?”

The strongest providers are likely to combine machine learning expertise with a deep understanding of, and preferably hands-on experience in, the IP world. Real-world licensing expertise is a scarcer asset than a powerful AI model or a polished user interface – bear that in mind when evaluating tools.

“How will we monitor it after deployment?”

These are far from being ‘set it and forget it’ solutions. Teams need to closely monitor the AI tools that they deploy, stay in touch with external providers to ensure that errors are addressed and keep tabs on product developments.

AI tools can already process information and perform routine analysis at a speed and scale that no human team could match. But verification and, most importantly, judgement are still human responsibilities. Successful AI adopters will be those who understand both what these tools do brilliantly and where they remain capable of getting things badly wrong.

The foundation of it all

One word consistently cropped up throughout Sisvel Connect: “Trust.” It was repeated many times during conference sessions, in meetings and at social events.

Yes, AI is here, is changing so much and will have an even greater impact in the future; but when you boil all patent dealmaking down to the core, it is about skilled, knowledgeable human beings getting together to reach a point where they can all agree to sign on the dotted line. If they trust each other – have confidence that they are being told the truth, are being given the right data points, won’t hear one thing today and another tomorrow, and so on – negotiations stand a far better chance of success than if not.

Trust can exist even when parties and people disagree to the extent that they end up facing each other in a courtroom. Building it is a skill; maintaining it an imperative; losing it a disaster. There may be quick wins in not playing it straight, but in the longer term you will struggle to develop the kinds of relationships and contacts on which a sustainable licensing business is built. Trust is the oil that makes the patent dealmaking world run smoothly. If you do not understand this, you are not doing yourself or your business any favours.

We had three extraordinary days in Barcelona with a brilliant group of senior professionals drawn from the entire ecosystem – and they spoke to, not across, each other. Our aim was to create a space that gave them the confidence to do that and we broadly succeeded. However, we can do much more. That’s the task for the 2027 event, from 28–30 September, once again in Barcelona.

In the meantime, thanks to everyone who joined us this time around – you were magnificent. Let's do it all again next year. Look out for your invitation …