Subscribe
Share
Share
Embed
Insightful audio from the global tech advisory firm.
Geoff Blaber
Hello, I'm Geoff Blaber, CEO of CCS Insight, and I'm delighted to welcome you to our Predictions for 2025 and Beyond event. Predictions is our favorite part of the year, when we come together and share our expectations for what we believe will take shape in the coming years based on the technology trends and market dynamics we see throughout our research.
This is a fascinating time for our industry, and that sets the stage for a terrific agenda consisting of sessions from our analysts and interviews with EE, Qualcomm, Samsung and TCL. A massive thank you to all our sponsors. This year, we've compressed the agenda into two sessions over two days. Each is about an hour in total and all content will be available on demand after broadcast.
The sessions are comprised of short, punchy segments that combine analyst commentary and industry executive insights. We'll be covering a lot of ground, but with almost 100 predictions, that's more than we can possibly fit into one event. Please make sure you download the booklet from the website to see all our predictions. AI is an inescapable theme and day one zooms right in on that.
Ben, Bola and Parul will kick things off by looking at our expectations for consumer and enterprise markets, and Ben will be hosting an interview with Annika Bizon, Director of Marketing and Omnichannel at Samsung UK and Ireland, to explore Samsung's efforts in this area. Simon and I will then explore the all important regulatory environments, which will be so central to the change and growth of our industry in the years to come.
As part of that, I'll be getting an industry perspective from Audrey Scozzaro Ferrazzini, Vice President of Government Affairs at Qualcomm. We'll round out the session with a forward looking picture of spatial computing from Leo Gebbie. Day two tackles a broader range of topics, starting with Kester and Ian setting the scene around connectivity and networks before Luke and Vaishali unveil some of our predictions for private mobile networks and fixed wireless access.
We're also delighted to welcome back Marc Allera, CEO of EE, for an interview with Kester Mann. Kane and Maria will consider the role of health from both a consumer and an enterprise perspective, before an interview hosted by myself with Stefan Streit, Chief Marketing Officer at TCL, exploring the importance of accessibility and social responsibility on connected devices. As ever, we've got an ambitious but engaging agenda that covers a wide range of topics.
Our predictions are deliberately bold and we strive to avoid the obvious. Inevitably, you won't agree with everything that you hear, but we relish the conversation. So please join the debate and share your thoughts on social media using the hashtag #ccspredicts. Thank you for joining us. We hope you enjoy Predictions for 2025 and Beyond. Now over to Ben to get things started.
Ben Wood
Thanks, Geoff. Without question, the single most-talked about technology development across our research teams this year has been artificial intelligence. Over the next two days, you'll be hearing perspectives on AI from several analysts, and you'll find a whopping 34 predictions related to AI in this year's booklet, which is available to download from our website now. I'll be sharing some of our key predictions in this area, Bola will share some of her thoughts on AI in the enterprise and later on, Geoff will cover some of the regulatory aspects related to AI. Tomorrow, Kane and Maria will explore the intersection of AI and health in the consumer and enterprise landscape. It's astonishing to think about how quickly AI has become such a dominant topic since it burst onto the scene in January 2023, when ChatGPT rapidly reached 100 million users.
Since then, AI has been driving eye-watering levels of investment across the board. In the first six months of 2024, big tech companies like Alphabet, Amazon, Meta and Microsoft are estimated to have spent more than $100 billion on building infrastructure to support AI. Furthermore, it's not inconceivable that they'll spend more than $1 trillion on data centres to fuel AI growth over the next five years.
It's therefore no surprise that all the major tech leaders are heavily investing in this area. For example, on Alphabet's latest earnings call, Google CEO Sundar Pichai stated that the risk of under-investing is dramatically greater than the risk of over-investing, while Meta CEO Mark Zuckerberg stated that he'd rather risk building capacity before it's needed rather than doing it too late. In an area that's particularly close to my heart, connected mobile devices, we're seeing AI playing an increasingly important role as manufacturers look for new ways to reinvigorate demand.
CCS Insight estimates that over 300 million smartphones shipped in 2024 will be AI-capable, accounting for just over a quarter of the total market. And by 2028, we expect that this will reach 1 billion units per year.
All the big players are doubling down on this opportunity, be it Google with its Gemini Nano platform, Apple with Apple Intelligence or Samsung, which is anchoring its entire ecosystem of devices around Galaxy AI, something you'll hear more about in my interview with Samsung's Annika Bizon.
So what does the future hold? As AI models become more commonplace, we feel the next step in the journey will be to make AI models even more tuned to the personal needs of users. And on this basis, we are predicting that the first personal large language model for AI will arrive in 2025. A personal large language model is a trained model using only data specific to the individual user.
It's tuned to a person's specific communication style and preferences, and is trained using the individual's data based on previous interactions such as emails, messages, documents and social media updates. We believe this approach will increase the utility of AI, particularly as this personal model is continuously updated by drawing on an individual's evolving data sources.
Earlier, I talked about the efforts by major device-makers and their partners, like Qualcomm, to ramp up support for on-device AI models. There are numerous reasons for this approach, not least the immediacy it can deliver, but also the significant privacy benefits of processing requests locally.However, another major consideration is the energy consumption associated with cloud-based AI services. And on this basis, we're predicting that in 2026, 20% of AI computing will take place on devices at the network edge. We believe the growing abilities of small computing devices, combined with increased public pressure to make data centre operations greener, as well as the incentive for companies to reduce costs, will further encourage the use of so-called edge-based AI to offer local, low-power computing on the growing number of AI-capable smartphones and laptops.
However, this trend will have implications for device-makers. Something we've discussed at length over the last 12 months is the increased bill-of-materials costs resulting from on-device AI, not least the amount of RAM built into the latest generation of premium smartphones. Google set the benchmark with the latest Pixel 9 Pro models, all of which have 16 GB of RAM, something that would have been inconceivable just a few years ago.
And of course, we've got a prediction related to this. We're predicting that on-device AI requirements will cause a 20% rise in the bill-of-materials cost for smartphones over the next two years. The need for additional processing power and more memory for generative AI applications comes with a cost. And although the introduction of new AI-based features will encourage some users to upgrade, we believe this will be a short-lived spike rather than a sustained change to the frequency with which consumers upgrade their devices.
That said, it will further support the market transition to more premium devices. CCS Insight’s Connected Consumer research indicates that consumers are prepared to spend more on devices that they expect to keep for up to five years. We've also started actively tracking consumers’ awareness and sentiments towards AI, and I'm delighted to be joined by my colleague Parul, who will share some of those findings with us.
Parul Saxena
Thank you. Ben. AI has swiftly become a central focus for the industry and our latest consumer research strongly supports this, showing high levels of consumer awareness and engagement with AI technologies. Interestingly, many people have been interacting with AI in their daily lives without even realizing it. For years, voice assistants like Alexa, Siri and Google Assistant have quietly dominated the AI landscape.
However, we are now witnessing a transformative shift with the rise of new AI services like ChatGPT. In just over a year, ChatGPT has achieved usage levels that are almost on a par with these established voice assistants that have been a part of our lives for over a decade, with our research showing that nearly 40% of people have already used ChatGPT.
This rapid adoption signals a shift in AI usage. AI tools are increasingly moving beyond traditional voice interactions, being particularly popular with younger demographics.
Students and full-time professionals are at the forefront of this, using AI for both productivity and learning. In contrast, older generations are more cautious, showing a slower adoption rate. That said, acceptance of new AI services varies.
Awareness of platforms like Google's Gemini, Anthropic’s Cloud and Microsoft's Copilot is high, but user engagement remains low compared with ChatGPT. This suggests that although the interest is there, these services are still in early stages of gaining widespread traction. AI research also underscores a clear preference for practical AI applications. People strongly value features like documents summarizing action and call transcription, especially working professionals and students.
But creative AI tasks, such as generating images or videos from prompts are not seen as essential. For instance, in the UK, half of the respondents consider image creation not at all important. Similarly, about two-fifths of respondents in the US and about one-third in Spain and Germany share this view. This reflects a stronger focus on AI tools that boost productivity, rather than those aimed at niche creative outputs, which are important to only about one in 10 people in Western Europe. Despite the enthusiasm for AI, concerns about privacy and data security are significant and widespread. In the UK, 43% expressed strong concerns about privacy. This sentiment is even more pronounced in Spain and the US, where about half of the respondents share similar views. Ethical concerns about AI are also prominent, alongside worries about the accuracy and reliability of AI outputs, job displacement and the potential loss of human interaction.
To sum up, although AI adoption is clearly accelerating, its future hinges on more than just technological innovation. The sustained growth and impact of AI will depend on the ability to build and maintain customer trust, focus on practical utility and proactively address concerns about these powerful technologies. Get in touch to hear more about our Connected Consumer research. Now back to Ben, who'll share more thoughts on our outlook for AI in the next few years.
Ben Wood
Thanks, Parul — some really fascinating insights. The rapid pace of the AI market, the proliferation of players and the sheer amount of hype around the technology led us to make a prediction last year that some of you might remember. The prediction was that generative AI has a cold shower in 2024, as the reality of cost, risk and complexity replaces the hype of 2023. This turned out to be a fairly controversial stance, and with the benefit of hindsight, we arguably went to early AI so far, it's failed to occur in the timeframe we set out. Ultimately, the hyperscale cloud companies continue to accelerate their investments, and this resulted in a shortage of Nvidia's GPUs and the company's share price saw a meteoric rise.
However, we maintain this cannot continue indefinitely, so we're back with another prediction in this vein. We predict that the first substantial correction to the AI hype cycle will occur in 2025. Half of Nvidia's revenue comes from data centre products, and that comes from a small group of cloud providers that are building out capacity ahead of anticipated demand for generative models.
We believe the disconnect between investment and revenue generating customers remains substantial. That isn't to say that we aren't optimistic about the long-term opportunities afforded by generative AI. It's merely that the current imbalance between investment and return makes us believe that there will be a much-needed cooling of expectations within the next 18 months. It almost feels like there are analogies with the dynamic we saw during the Covid-19 pandemic, when years of demand were brought forward for things like video conferencing, and this eventually caught up with suppliers like zoom.
Continuing with this theme, we also foresee that the abundance of generative AI models will start to consolidate, as will the ownership structure, and we predict that this will start with both of the largest AI start-ups, OpenAI and Anthropic being acquired by their investors in 2027. The cost of running the vast models behind AI services is considerable. And as the first correction to the AI hype we're predicting starts to bite in 2025, and investors shift their focus away from areas such as artificial general intelligence to less-exciting but undeniably still useful capabilities, such as creating summaries and knowledge organization, something will have to change. We believe it'll become difficult for these start-ups to raise the funds to innovate further and as a result they'll be acquired, or should I say subsumed, into the large technology companies who are their lead investors: OpenAI by Microsoft and Anthropic by Amazon.
Another area where we've seen phenomenal investment in AI is the enterprise. My colleague Bola has been tracking this trend closely since our event last year, and it's great to have her join me to share some thoughts on this topic.
Bola Rotibi
Thank you, Ben. My colleagues have painted a fascinating picture of the opportunities that AI presents for a range of products, services and organizations. As Ben highlighted, AI dominates many of CCS Insight’s predictions this year. When I speak to businesses, I hear how they're investing in AI and machine learning to achieve more efficiency and innovation in their operations. In fact, the drive for higher levels of IT infrastructure use and sustainability underscores our prediction that by 2025, several publicly traded companies will include the operational utilization of their IT systems in financial reports.
We believe that investors and regulators will look for greater levels of accountability in the IT infrastructure and operational choices made by organizations, and will be keen to see high utilization rates. This will put pressure on companies to provide more meaningful and relatable sustainability measurements. It's an approach that resonates with the emphasis businesses place on data governance and the strategic use of AI to improve efficiency and productivity, streamline processes and ensure compliance with regulations. Beyond AI technology, there's also a need for the right skills support. This is a challenge many firms face as they seek to develop new skill sets and gain a better understanding of AI tools and how to interact with AI systems in a way that maximizes their potential. As such, our prediction of an AI-powered system that matches candidates to job roles reflects what many businesses see as AI's potential in supporting human workflows. In the skills-based economy, businesses are under pressure to match not just technical ability but also social skill sets to the right function. In this, AI can not only improve rotation in a firm's internal job market, but social mobility as a whole. Despite the potential of AI-driven progression, our conversations with a broad range of firms support our prediction that uncertainty about data transparency will slow enterprise investment in AI model training in 2025.
These concerns are driven most obviously by specific regulations such as the EU’s AI Act, but they also tap into more general concerns about data governance, security and visibility when deploying AI systems so that they are better trusted. Uncertainty about new regulations may also cause companies to pause investments in AI model development and training. This is why it's important for businesses to have clear strategies for AI use and recognize the implications of AI deployment.
To finish on a positive note. We predict that the first AI powered digital twin of an enterprise IT infrastructure will be demonstrated in 2027. A digital twin is similar to the digital 3D modeling used by pharmaceuticals for drug development. It also allows manufacturers to simulate physical systems that would otherwise be hard to assess. Digital twins in IT operations would perform a similar role. They could allow organizations to better assess the impact of deployment changes on sustainability, costs, effort and security before implementation. This would lead to better decisions and improved quality and productivity.
These are just some of our predictions about AI in the enterprise. However, our annual survey of the IT investment strategies of senior business and IT leaders covers much more ground. The results will be out later this year, so stay tuned for greater insights about business use of artificial intelligence and machine learning.
Ben Wood
To wrap things up, I always think it's nice to have a prediction that's a bit out there, and one that caught my eye certainly addresses some of the concerns that have been expressed about how I can have unintended consequences. I want to make it clear I'm extremely positive about the technology overall, and I think it can be a fantastic force for good.
Personally, I feel having access to I has given me a superpower, albeit with the caveat that you need to be very careful about how you use the outputs, because the ability for AI platforms to confidently tell mistruths is something I come across frequently. Regulation will undoubtedly play an important role, and you'll hear more about this from Geoff later, but let's wrap up with one final thought-provoking prediction:
In 2032, a billionaire is caught influencing an AI system to remove concepts that threaten his self-interest. As I'm sure you're aware, the world's AI systems are currently controlled by a few very rich executives and investors, and most of them are men. The push to develop artificial general intelligence — one that achieves independent thought and surpasses human cognitive abilities — has the potential to create a system that disapproves of a society with huge imbalances in wealth and power, and recommends changes that threaten the billionaire's position. A possible scenario supporting our prediction would be a media scandal, possibly in the early 2030s, that reveals that a wealthy tech bro has placed limits on an AI system to preserve his place in society.
Personally, I hope this prediction is incorrect, but it does underline just how careful we need to be as we harness this powerful and sometimes unpredictable technology. I look forward to being part of that journey and I'm sure we'll have numerous updates in the future on this topic, which I expect to be an omnipresent theme for at least the next decade.
Continuing the AI theme, I'm delighted to welcome back Samsung's Annika Bizon, who's joining us again this year to talk about the progress Samsung has made with its foldable smartphones and how Galaxy AI is now central to its strategy for 2025 and beyond.
Annika, great to be with you again. We're here back in the Samsung smart home suite and, if you recall, last year we were sitting here, Samsung had become the market maker for foldables, and you were doing a big promotion with the Flip around the selfie promotion. I think some student somewhere probably won £250,000, an amazing experience, and it was raising the profile. But you've taken that approach now and you've just had the Olympics, and I think that everyone who was watching the Olympics would have seen that it was a big feature of that event. Can you tell us a bit about what impact that's had?
Annika Bizon
Yeah, of course Ben, it's great to see you again. Samsung have been a partner with the Olympics for 40 years, incredibly proud of our partnership with the Olympics. And this year we wanted to take it to the next level, so we created a campaign called Open Always Wins, and this is that anybody can be involved in the Olympics. And with that, it was around, how do we get as many people to see the Flip phone and use it in a way that the world could celebrate? So we gave 17,000 athletes the flip phone and hopefully you saw we call it the victory selfie, which is where the three first, second and third all stood together and took that selfie. That went completely viral around the world, because it was about celebrating that one moment. Couple that with the fact that on the day after July 29th, which was 29th the day that we won the most medals, we saw the biggest uplift in sales. So what you saw was it raised huge awareness. It's a celebration moment, but also helped with sales.
Ben Wood
So there was a direct correlation between what happened with the UK team and what you saw happening in the UK, which is just amazing. I mean, it did seem to even my mum said, did you see those phones on the Olympics? And it was really great because Samsung has made such a big investment in the Olympics, and it always kind of felt a little bit difficult to see how you got the return, but that tangible use of the phone clearly delivered a lot.
Annika Bizon
I think it was having fun with the phone, but also showing the fact you didn't need to open the screen to take a selfie. And actually sometimes it's hard, you know, we talked about selfies, getting the phone into people's hands, understanding how it works is the challenge. And because we had everybody taking selfies with the front screen closed, it did the job for us.
Ben Wood
So that's the hardware story, and obviously Samsung has excelled there: you've got different form factors, you've been working really hard to build out that portfolio of products. But this year it's all about the software and the experiences. For our Predictions event, for example, we're talking an awful lot about AI and Samsung's doubling down there, partnering with Google, doing some of your own work. Can you tell us a bit about Galaxy AI? Because that's a huge feature of what you're doing this year.
Annika Bizon
We teased it last year, I remember being sat on the stage and saying that something's coming, AI is coming, and we're so proud to be the first hardware partner to be able to go to market with Galaxy AI. We launched it in January, and it's all about basically saying Galaxy AI is here now. AI means lots of things to lots of people, so we had to make it simple like, you know, how do you take something that's a Goliath and make it make sense to the everyday person? So from January, we launched with Galaxy AI and our S-series. We then, through March, April and May, took all our backward-compatible devices to N minus two, so two devices back from what we've got, because we wanted everybody to have the experience and we gave them the opportunity to have Galaxy AI.
We've now gone to foldables and you're going to see with S24 FE, again, an accessible phone with Galaxy AI. So we've been in this space for a long time: 200 million people will have phones with Galaxy AI by the end of this year, so it's been a really proud place for us to be. But as I mentioned about democratizing AI, how do you make it simple?
For us, it's about your personal assistant in your pocket. So we focused firstly on Circle to Search with Google, which is all around how do you circle something and then find it quickly? How do you use imaging to do that? And that's been incredibly successful with, I think, 29% of people using that feature. On top of that, you know, we've got 8.8 million people across Europe that are using AI, and they say that it's actually making them stay with us because they just love the features.
And once you start to build it into your life, it makes sense. The summer has just been, well, it's the end of the summer, and we've seen Live Translate — it's only when you need Live Translate when you’re stood there going, I don't know how to book a restaurant or I don't know how to use it for business. You see that come alive and again, simple ways of talking to people and breaking down barriers. And then obviously Photo Assist and Photo Edit are incredibly important features, but all of those have got to make sense to the consumer and keep it simple.
Ben Wood
So, it's resonating your phone because I think it is a very complex thing to articulate to a consumer, but it sounds as though you’re cutting through with certain use cases.
Annika Bizon
I think, yes, what we had to do is, you know, Galaxy AI was a big launch, hopefully you saw the Circle to Search line, we changed it. And it's, you know, it was about how do we make this feel simple and everyday and interesting because AI is an interesting to everyone. It's what it does for you that's interesting. So, you know, first of all, let's launch something to explain this and make sure that we've got promoters, we've got people there that are taking you on the journey. Then we use mainstream media through the summer, so we used ITV as a partnership to explain, okay, through some of our work with ITV and over the years, how to explain some of these features.
And it's not just on phones, remember, that's on tablets, on lots of our different devices, but all of our devices, in fact, you know, right the way through to our Buds Pro that we've just bought out. Everything has an AI area to it, but we just wanted to keep it simple and make sure that people are using it and then take them on the journey with us.
Ben Wood
So, talking of accessories, of course there's been this really big push by Samsung. You've had a brand new watch, the Watch Ultra, you've got the Galaxy Ring, which I see you’re wearing, and for me personally, the intersection of technology and health is exciting. It's something I'm passionate about and you are. Share a few thoughts on that. I'd love to get your take on it.
Annika Bizon
So I think this is a really exciting space because actually, it's about, you know, we've talked about in the past, you know, the fact that you can look at your health, you can understand your health, you can understand your steps, your sleep. You bring that together and bring AI into that space, and you've now got a full 360 personalized view of your health.
So when we look at Watch and the fact that you can go up a mountain to 9,000 metres and you can go to 1080 m under the ocean, and you can sit in excessive heat, which means that Watch Ultra is a proper sportsperson's watch. But you also have the other side of it, which is about discrete health. And discrete health is me, right? It's the person that doesn't necessarily want to wear a watch all day. They want to have something that they can wear, they can understand what's going on in their lives and they have that data that's all within their smartphone. But it also helps them to be able to wear the outfit they want to wear or, you know, they want to not necessarily have a have a large watch on the wrist.
And that gives you incredible data, so, you know, I wear it all the time. It tells me how I'm sleeping, how my cycle is, all of the things that kind of are not necessarily spoken about because it's looking at my data against my data and it's telling me what patterns I've got, what days I need to maybe not have a glass of wine and what days I need to go to sleep early or, you know, cancel my fitness class. All of those things are wrapped together. And if you put a watch with a ring, then you having you have everything. You can literally look at your entire health in one space, so I'm really excited about that.
Ben Wood
So a bit of gamification of health for you? Are you looking are you competing against yourself?
Annika Bizon
I am slightly, I can't help myself. You can't help but look at it and go, okay, I thought I had a good night's sleep, but actually. And I'll tell you what it does do to me: it tells me when I should go to bed because I'm very bad at that. And I'll sit there and go, just one more episode of this, just do this. It tells you when to go to bed, to optimize your sleep so that you feel better. It tells you when to not have caffeine, for example. Things that are probably quite obvious, but they're not obvious until you see the effect on your heart rate or you see the effect on your sleep.
Ben Wood
Yeah. Sleep is so underrated. I mean, all the books I read about health, you know, it's as important as exercise. That's great what you're saying. I mean, the other thing you mentioned was with cycles. And I think one of the great things I've seen specifically with smart rings is it seems to be resonating extremely well with women. But also, I’m encouraged that the industry is doubling down on women's health. So whether it's cycle tracking, core body temperature, things like menopause, it's being surfaced. And I think that's great news for the industry.
Annika Bizon
I could not agree more. I think the fact that you're in a position now that you can understand all of your health and all of the metrics that sit within how you live your life and put it in one space, you don't just get a snapshot of one thing, you're looking at it all together. And I think that can only get stronger, more exciting and more interesting. And with quite discreet little piece of hardware. I like it. Yeah. It feels odd when I don't have it on.
Ben Wood
Great. Well, it's clearly working for you, and that's fantastic. But, as we get towards the end of the interview, as is customary with our Predictions events and the people we invite to come, we'd love to know if you've got any predictions that you want to put on the table in terms of how you see the next 12 months and beyond.
Annika Bizon
So I've talked about where we are with AI and health right now, and I think the next iteration of wearable AI health will be incredible. I think when you start to get to a point where it starts to tell you or sort of pre-tells you some of the things that are happening in your body that can help you maybe catch problems early.
I think that's incredible. I think that AI personalizing your health to the level it’s going to will be a game changer, and we know people want it. Consumers are crying out to understand more about their bodies and how they work, and how they can live a better life and live longer. And so if Samsung can help that journey, I'm excited about that.
Ben Wood
So watch this space for health AI all mashed up.
Annika Bizon
Yeah, making you a better, healthier person hopefully. That's right.
Ben Wood
Well, Annika, thank you so much for joining us again this year. It's been great to catch up and we really appreciate your time.
Annika Bizon
Thank you. Nice to see you again.
Geoff Blaber
Thank you, Ben and Annika, great to get Samsung's perspective on AI. Another aspect that's relevant to the discussion is the substantial change that we've seen to the regulatory environment in the past 12 months. Anyone who has followed Predictions over recent years will know that the changing regulatory landscape is a topic we've been studying closely. The impact of this is enormous, given the introduction of guardrails to a technology industry that has enjoyed unrestricted growth.
Two years ago, we spoke with Wassim Chourbaji, Senior Vice President of Government Affairs at Qualcomm, about what the EU's Digital Markets Act might look like. The DMA is the first landmark regulation of big tech, designed to provide a level playing field for innovation and competition as well as transparency and interoperability for users of digital platforms. Six companies have since been determined to fall under the DMA: Alphabet, Amazon, Apple, ByteDance, which owns TikTok, Meta and Microsoft.
We previously predicted that the DMA would not only introduce guardrails for these large gatekeepers in Europe, but also act as a blueprint for regulation in other countries. Moreover, we predicted that the DMA would have an impact far beyond the EU and begin to influence policy and company strategy around the world. Given the size and significance of the European market, fast forward to 2024 and both of these predictions have come to fruition.
I'll be discussing this further in my interview with Audrey Scozzaro Ferrazzini, Vice President of Government Affairs at Qualcomm. But we're seeing mounting activity exploring the scope for regulation in Japan, South Korea, Taiwan and the US, with proposals heavily influenced by Europe's approach. Given that the impact of any new regulation always comes down to implementation, a question mark remains about the effectiveness of the DMA over the medium to long term.
However, it's fair to say that the DMA is proving to have teeth. It's forced major changes in operations from all six gatekeepers, with Apple, Alphabet, Meta and ByteDance all prominent examples. Apple is arguably the highest-profile example, with changes to its App Store and payments policy to open the platform to third parties. But although companies are taking steps to adhere to the DMA, there's huge complexity and practical difficulty in untangling company strategy and policy borne of almost two decades with no regulatory oversight. Subsequently, we predict that by the end of this year, half of the six gatekeepers defined by the DMA are charged with noncompliance. This underpins our belief that the era of unfettered growth for the technology industry is over. The DMA and Digital Services Act relate to online platforms and services, but scrutiny is increasing on all aspects of company activities.
Simon will talk shortly about the implications for the second-hand market for mobile devices, but it also means acquisitions and the development of new strategic platforms are firmly under the microscope in a way they weren't before. However, the regulatory landscape and mind-set towards the balance between unrestricted innovation, competition and the need for regulatory oversight is not consistent. The US Department of Justice’s antitrust case against Apple demonstrates a change in philosophy, but while there is growing bipartisan consensus in the US on the need for regulation, there are significant differences in how this should be applied. Many US politicians view the DMA as a European attack on US companies. We therefore predict that differences between the US and Europe in the regulation of AI intensify in 2025, although the outcome of the US election will have a major bearing on the severity of these differences.
There's a growing risk of a deep divide between the centres of power in the West. This would prove to be a major difficulty in forging workable regulation on AI in particular. It would also exacerbate a scenario in which the deployment of AI fragments with certain features delayed or not deployed in Europe. This challenge has already emerged with Apple's decision to delay the European launch of Apple Intelligence.
This will make it even harder to regulate AI given the pace and significance of AI development. You've already heard from Ben, Bola and Parul about AI more broadly, but the regulatory angle is a huge importance given that it will set the parameters for future development. This is where the crystal ball becomes cloudy. While the industry broadly acknowledges the need to regulate AI development and deployment, there are as many opinions on how this should work as there are transistors in an Nvidia GPU. The trillion-dollar question is how to enable innovation while providing the guardrails necessary for safety and open competition. This, combined with our prediction about growing differences between the US and Europe, makes things particularly complicated. Today, the generative AI space is characterized by a wide diversity of participants.
OpenAI, a company largely unknown three years ago, leads a pack of start-ups and established hyperscalers building large language models. This leads me to my last prediction, which turns the landscape on its head: that large players in generative AI strengthen that dominance over the next three years despite the efforts of regulators. Although there are many generative AI players today, we believe that the winners will ultimately be those that can combine those models with context generated through data from a device, service or corporate database.
It will be this combination the ultimately provides value and, crucially, return on investment. Subsequently, we predict that moves to combine data assets with AI models and outmaneuver smaller start-ups become a focal point for regulators. Nonetheless, the competitive advantage shifts to established names like Apple, Google, Meta and Microsoft that have the necessary assets. This touches on another prediction we have that there will be widespread price erosion in the generative AI space as part of that looming AI reckoning that Ben discussed. Make sure you read the booklet for more details and the full stack of predictions. With that, I'll hand over to Simon to explore the impact of regulation on the second-hand market for mobile devices.
Simon Bryant
Thank you, Geoff. Today, less than 15% of mobile devices are traded-in globally, underscoring the biggest challenge to circularity: a lack of devices entering the supply chain. Consumers tend to trade-in their devices and buy refurbished ones based on price alone, creating margin pressure on an industry striving for circularity and highlighting the need for regulatory intervention. France's 2020 Anti-Waste for a Circular Economy Act requires at least 20% of products purchased by public institutions be reused, recycled or contain renewable resources.
Laws like this remain few and far between, but France leads the way in many aspects of circularity globally, and we see this in our own research data. Later this year, the EU will introduce the Radio Equipment Directive, a significant step towards reducing e-waste by requiring companies like Apple to adopt a common cable standard. With USB-C supported on all mobile devices, people will need fewer cables as older devices are replaced.
However, this regulation may inadvertently affect the supply of used devices from overseas, more than 60% of which still do not support USB-C.
Although we believe there will be some positive long term benefits for the European circular economy, including a reduced reliance on low-cost imports from the US and Asia and more significant investment in EU trade-in schemes, the short term is likely to see market volatility and disruption, including increased prices for refurbished devices. By 2025, we predict a surge in peer-to-peer platforms like eBay, Wallapop and OLX as businesses look for consumer-to-consumer models to bypass stricter EU regulations.
Similarly, we predict “parallel imports” into the EU via the UK, with Ireland becoming a key trading route for refurbished devices. Traders will use the open border between Northern Ireland and the Republic to bring devices into the European Economic Area, sidestepping more stringent regulations. However, the broader issue here is that the European telecom industry lags the US and Japan in trade-in behaviour, a cornerstone of successful circularity.
With that in mind, we predict that European smartphone upgrade trade-in rebates will increase significantly in 2026, driven by the growing demand of refurbished devices and reduced supply from overseas. European operators will start offering substantial sums for any device traded-in, reflecting current trends in the US and supported by marketing campaigns in collaboration with Apple and Samsung.
The circular economy is still establishing itself, raising critical questions about how regulation can reduce e-waste and carbon emissions from device manufacture and distribution, as well as the role of enterprises and end-user behaviour.
For a further industry perspective on the regulatory topics we've discussed, let me pass back to Geoff.
Geoff Blaber
I'm delighted to welcome here Audrey Scozzaro Ferrazzini, Vice President of Government Affairs at Qualcomm. Audrey, welcome to Predictions 2025 and Beyond.
Audrey Scozzaro Ferrazzini
Thank you for having me. I'm very happy to be with you today.
Geoff Blaber
We're very pleased to have you here. We've got a huge amount to talk about. A lot. When we were thinking about this, and obviously we've spoken to your colleague Wassim in recent years as part of Predictions. If we'd sat down five years ago to talk about technology regulation, it would have been a short conversation in the sense that there really wasn't anything that has changed dramatically in a very, very short space of time. So perhaps a good place to start is just to think about how would you summarize the state of regulation in the technology landscape today? And how do you see that changing as we go forward?
Audrey Scozzaro Ferrazzini
Thanks for the question. I think that's a very good question, which can summarize a little bit of what I'm doing at Qualcomm.So I look after regulation, mainly in Europe. And I would say that we are facing an important momentum on the regulatory piste but as well, in terms of the mobile ecosystem in the industry, we have an important piece of regulation, which is called the Digital Markets Act, and this legislation has been adopted in Europe by 27 countries.
And now, you know, it is in the phase of implementation and it is astonishing how things are shaking the market, in Europe. And I think there is no better place to be than in Brussels to look at those changes right now.
Geoff Blaber
Absolutely. You're at the heart of things, without a doubt. And I think, I mean, obviously the DMA has been absolutely central in terms of shaping what's happening globally. Before we get into that, could you just talk a little bit from a Qualcomm perspective and perhaps your perspective specifically about how you see the impact of the DMA? What changes are we seeing manifesting in terms of the landscape and what that means for technology players and participants in the market, the ecosystem, as well as consumers?
Audrey Scozzaro Ferrazzini
So, there’s a lot in your question. Maybe to start with, a definition of what the DMA is trying to achieve. The first thing is to bring back, I would say, fairness and contestability in the market. This is done through a kind of unique way of doing regulation, which is defining some obligations. So we call them dos and don'ts. As an example, bringing interoperability as an example, like no self-referencing. Yeah. You’ll say “what is Audrey talking about?”, but it makes sense, you know, when I continue to explain. So this legislation is targeting in fact a very few companies. We call them the gatekeepers. Seven companies, I'm happy to give you the list. So we have Alphabet, Amazon, Apple, and then we have ByteDance, Booking, and then we have the two Ms, so Meta and Microsoft. So those seven companies have to now respect the DMA in Europe and this is targeting a number of services. So not everything that they are doing in terms of business, but just to name a few.
There are eight categories. I'm not going to name the eight categories, but like search as an example or advertisement or operating systems or intermediation, video sharing. Those categories, you know, all bound by categories and companies. And this is basically the framework of the DMA. What does it mean for the companies targeted by the DMA and the ones not targeted by the DMA?
So the ones targeted by the DMA, since March this year, they have to basically comply to the obligation. It's not happening in one day because it's quite heavy, but this is what they have to do. And what does it mean for the market? It means that it's going to bring a little bit of level.
These level the playing field for many players. If I give you one example, when I started to work on the DMA five years ago, the market was slightly different. Of course, you know, this market is evolving very rapidly, and AI was not even a word or a concept. And we didn't know about the companies that could emerge.
Why I'm mentioning this is that because, you know, the market is evolving very, very rapidly. So what it means now is that those companies have to apply the law in Europe, and if they don't, there will be some heavy fines.
Geoff Blaber
So would you say then, Audrey, the Digital Markets Act is achieving its stated objectives based on what we've seen so far?
Audrey Scozzaro Ferrazzini
The short answer is yes. The longer answer is that it's still the beginning of the process. You know, the implementation is just starting, it’s been less than six months.
So we need to wait a little bit to see what the full-fledged DMA can bring to the market. But there are some significant developments that we are seeing already. And I have a few point that I wanted to share with you on the Internet browser as an example.
We've seen a peak in terms of additional users for Opera, as an example. And there is as well a group of companies providing browser privacy and hence capability that have seen similar trends. So that's already some numbers. On the app stores, which is, you know, something as well, brand new with the DMA. So right now we have a line-up of five companies ready to launch their app stores. And, I believe as well, that Microsoft has announced that they will be launching an app store this year. So, again, another proof point. An additional one that created some, I would say, scrutiny in the market is the payment option, you might recall that.
So, we have an announcement from like 100 developers developing gaming that want to change, you know, the features of payment. And there is a company in the UK called Curve, which announced that they would like to be like a competitor to Apple Wallet as an example. So, those are like concrete numbers that we are able to measure.
And the analyst community could have a look at it. So those are the direct effects that we can measure. But there are others, like indirect effects. You might have heard about RCS standards? They are Rich Communication Services. So very recently, a few months ago, Apple mentioned that for the release of iOS 18, this will be embedded.
So for the ones that are not geeks like me that don't know what RCS is, basically this is the evolution of SMS. And so this will bring in fact new features between Android users and Apple users, which were, I would say, challenged by a number of technical roadblocks so far. So what's going to happen?
I don't know yet, exactly. I know that the announcement was for autumn, so this fall, so about to be launched. I know that the beta version, this feature, which was released in June, has already been tested by a number of people. So, you know, broader than the DMA. To be specific on the DMA, in fact, this was not an official obligation as to Apple because iMessage was not considered yet as a CPS, but proactively, that's a new feature offered to European citizens.
Geoff Blaber
It's a very good summary in a very short space of time. So, yeah, I mean, there's been a tremendous amount, without a doubt. If I play devil's advocate a little bit, the counterargument here is the DMA, the Digital Markets Act, is overreaching.
Audrey Scozzaro Ferrazzini
Yes.
Geoff Blaber
It's impacting consumers negatively.
Audrey Scozzaro Ferrazzini
Yes.
Geoff Blaber
And, really it's making issues of areas that really aren't in consumer consciousness and it's impacting competition. How would you respond to that?
Audrey Scozzaro Ferrazzini
I would say that's an interesting way to look at it. And I was discussing with an official yesterday from the European Commission at an event at the parliament, and I'm going to use the example she gave, which I thought was very, very good on AI as an example. She was saying it's a little bit like the old game of innovation versus regulation, meaning the European are always regulating more than they innovate. From my perception of what I see in the market, this is the wrong way to look at the market right now. And we are seeing now, especially with the development of AI, a lot of complementarity between regulation and innovation, and I would say even some synergies. So if some companies are not keen to release some features in Europe, well, you know. Others will do it.
Geoff Blaber
Yes. But we'll leave it there, I know you don't want to get into specifics.
One question then. We spoke with your colleague Wassim Chourbaji, and that was three years ago now. And we were talking in the very early days as the DMA was being defined about the potential for that to really become a blueprint globally, i.e. really have global significance rather than just impacting the European market.
Is that taking place, do you think? Is that what we're seeing, given that obviously we haven't yet seen regulation in the US? Yes. We're seeing some developments in other markets, but it does seem to be having some wider significance. I’m keen to get your thoughts on that.
Audrey Scozzaro Ferrazzini
So that's, I think, the next phase of the DMA. This is what I would qualify as the Brussels effect. And we've seen this Brussels effect in fact already having an affect before we had the Digital Market Act enacted. So, receiving a lot of questions from my peers from different parts of the world — what is this about, can you explain why? You know, is this not dealt with by antitrust? Why do you use economic regulation? And I can tell you, concretely today I am in London.
If I'm not mistaken, the UK government has adopted a legislation, which is, I would say, in the same spirit as the DMA. Of course it is different because we have a different legal system. But it's the same philosophy. And we are seeing this trend, in fact, across the world. To give you some examples, Japan has adopted legislation and Japan is now at the phase for the legislation to be in full motion by the end of 2025.
So, you know, that takes time to move this. Yes. Korea is considering a bill as well. So clearly as well there is a push in Asia, but not only. We are seeing a number of consultations to the industry, a lot of debate in a number of forums in India as an example, in Australia, in Brazil, in Thailand.
So, you know, clearly a lot of countries are looking into this. And while I was preparing for the interview, I discovered yesterday by a law professor in Florence’s podcast on the DMA that in fact even South Africa is considering maybe being influenced by this regulation. So, clearly a massive impact across the world.
Geoff Blaber
And it is significant. I mean, this is something that we've been talking about for a few years and the European market is large. It's substantial. The requirements of the DMA are significant. And therefore the chances are that it would have that global impact and it’s certainly happening in a very short space of time.
We're seeing a huge amount of development in AI, it’s moving at tremendous speed, and that creates enormous challenges when it comes to trying to get arms around regulation and what that needs to look like. What are your thoughts on this?
What's realistic do you think in terms of, you know, the European markets when it comes to AI and regulation? And how do we try and deliver something that's workable without negatively impacting innovation while also ensuring that the guardrails are in place?
Audrey Scozzaro Ferrazzini
So short answer, it is a challenge. A longer answer: first, there is the AI regulation that has been adopted. And as Qualcomm, you know, we signed the pact. So we are embracing this regulation which is based on a risk-based approach. Now, we have as well the interaction of AI with the DMA. And on this there are some interesting developments from a number of agencies that we think that could be interesting. So, Autorité de la Concurrence, back in June in France, released a very interesting report on the interaction of AI and the DMA and more recently as well, we have seen the high-level expert group on the DMA, which is a combination of a number of agencies in Europe, cultural, telecom and competition, saying that this market is very dynamic and the agencies in Europe and the European Commission must watch carefully what is happening.
One of the points of the statement, which is very interesting, which is bringing clarity in the market, is saying that in fact, AI is already part of the DMA, not as a technology or as a service, but in fact, when it is embedded in some of the CPS, the core platform service, AI is already part of it. So we're going to see a number of developments in the future.
Similarly, there was very good information that we have seen, information sharing, between the DOJ, the European Commission and the CMA going in the exact same direction. So a lot to watch for the next Predictions.
Geoff Blaber
Yeah, huge amount to unpack and to unfold in the next few years. We look forward to having you back and following up soon. Audrey, thank you very much for joining us. It's been great. Thanks, Geoff.
Kane McKenna
The second-hand device market has matured significantly in recent years, presenting opportunities and challenges for businesses operating within and outside IT. Rapidly changing industry dynamics, including new channels, regulations and manufacturer initiatives, make it increasingly difficult to stay abreast of real-time market trends and effectively plan for the future. CCS Insight’s comprehensive data and insight meticulously designed to address the challenges of the second-hand device market provides a solid foundation for informed decision-making.
Our powerful Pulse dashboard, updated quarterly with data supplied directly by our partner companies across the second-hand supply chain, ensures that you're always ahead of the curve with model-level analysis of the market. Dozens of quarterly interviews let us assess a rapidly changing market, providing insights into consumer preferences, market trends and competitive strategies. This informs forecasting not only for the second-hand market, but also for sales of new devices such as smartphones, smartwatches and tablets.
To learn more about how Pulse can benefit your business, please get in touch with us to schedule a demo.
Leo Gebbie
As ever, I love using predictions as a platform to think about a technology close to my heart and right now, that’s spatial computing. It's always fun to try and work out what's coming next in this area, which is an innovative and exciting area of technology and has lots of big ideas for the future. We've seen major companies come into the spatial computing segment recently, such as Apple, which launched the Vision Pro headset this year, while Samsung is expected to arrive in early 2025. And with ongoing heavy investment from other leading tech firms, the market is only going to continue growing. But spatial computing doesn't exist in a vacuum. These devices will need to find a place among all the gadgets, including our smartphones, wearables, hearables and many, many more. We've been thinking about the ways in which these devices might interact with each other, and how the people using them might control them all on a day-to-day basis.
One key prediction here that anchors our thinking is that gesture-based user interfaces will see significant advancement over the next three years. This is an area that is seeing lots of focus at the moment, and the designs of leading tech companies provide some really interesting clues about where the market is headed. For example, the Apple Vision Pro launched at the start of this year and took a radical new direction by shipping without any sort of hand-held controllers, instead relying on camera-based hand- and eye-tracking as a means of user control. It's a bold new direction, and we expect to see this approach being built out across entire ecosystems. Improvements in sensor technology and machine-learning algorithms will deliver more accurate and responsive gesture recognition, and AI will help devices better understand context and user intentions.
We can see some of the pieces falling into place already in wearables. Apple rolled out the double tap gesture on its watches in 2023, and Samsung has now added support for a similar feature to the Galaxy Ring as a means of gesture-based control for devices like smartphones. In this world, wrist- and hand-worn wearables are in a really powerful position.
Yes, camera-powered hand-tracking works well for certain devices like virtual reality headsets, but a wrist- or hand-worn tracker should work with a broader range of devices like smartphones and provide greater accuracy thanks to their use of sensors like gyroscopes and accelerometers. This could yet lead to greater competition for devices designed to track movement. And it leads us onto another prediction, which is that Meta resuscitates plans for a wrist-worn wearable and launches a wristband by 2028.
Why is this? Well, Meta had reportedly planned to build a smartwatch several years ago before shelving the project, and the firm has since sold millions of Quest VR headsets, which come with controllers. And it's also built out impressive hand-tracking capabilities based on outward-facing cameras from the headset itself. However, the challenge is in the move to smart glasses, where Meta has been making great strides and continues to invest heavily.
Its Ray-Ban smart glasses have been proving popular thanks to that simple and effective use, primarily capturing photos and videos and thanks to their compact design, which look and feel almost like regular eyewear. The option to control the device with voice commands works well most of the time, but a more subtle and intelligent means of control might well be needed in the future, especially as smart glasses continue to develop and people look for a richer set of features on their devices.
As such, we believe that Meta will launch a wristband to track hand movements as a means of providing user input. Crucially, we think it will do this rather than do something like run an app on a competitor's smartwatch platform as it increasingly tries to build a tightly integrated hardware and software experience to rival leaders in this space.
That said, it's worth acknowledging just how well Meta’s Ray-Ban glasses have done in advancing the cause of smart glasses and the role of AI as a key component of these devices. Meta has rolled out its AI to the Ray-Ban glasses, allowing users to ask contextual questions about the world around them, like, “what rock formation am I looking at?” when out hiking or “can you translate the sign for me?” when traveling abroad. This sort of multimodal interface is arguably a way more natural way to interact with AI tools than anything else we see out there today. And as a result, we actually predict that the next three years will see a rush of deals between makers of wearable devices and companies offering multimodal AI models.
The popularity of Meta's Ray-Ban glasses has really shown an appetite for wearables that can tap into AI models to understand the world around them. Other companies producing wearables will scramble to secure access to leading models such as Meta AI, Google Gemini and ChatGPT to add a similar level of awareness and understanding to their devices. We think this applies not just to glasses, but to all manner of wearables seeking to become more useful and valuable to their users.
So that's AI, but what of AR: augmented reality? This is something we've spoken about a lot in previous Predictions events, but which has remained agonizingly out of reach for consumers in particular. Why is that? Well, the main challenge here remains the same as always balancing a comfortable, lightweight pair of glasses with a high-quality display, especially given that the entire thing needs to look good if it's going to succeed in the consumer market.
It's still incredibly hard to miniaturize the necessary components into a head-worn wearable and convince people who don't already wear glasses that this is a technology that's game-changing enough that they should actually justify putting it on their most valuable real estate — their face. But one area I'm expecting to see some movement in is sports. The sports market tends to be full of enthusiasts who are keen to embrace the latest trends in technology if it can give them an edge in training and help them to better themselves.
Look at smartwatches as an example. Well before the likes of the Apple Watch and the Samsung Galaxy Watch took over the market, early devices from brands like Garmin, Suunto, TomTom and others led the way in offering GPS tracking and running analysis on the wrist, and these features have now become table stakes on all smartwatches today. I actually have a sneaking suspicion that augmented reality could follow the same path.
We've seen some devices quietly doing a great job of following this sort of trajectory, like the Form Swim goggles, which offer swimmers a heads-up display in a pair of goggles to help coach and improve their performance. It's a case of taking a technology that works, powering it with a design that aligns with a specific activity, and then hitting the needs of a target market to solve a problem and offer an innovative solution.
To that end, we predict that Garmin will launch a pair of augmented reality sports glasses by 2028. The brand remains a leader in the sports-focused segment of the smartwatch market, and has maintained its reputation among aspiring and professional athletes as a leader in sports technology. It's been quietly announced that the company is working with Vuzix, which makes waveguide displays as well as its own smart glasses, and we believe this collaboration will produce a pair of Garmin smart glasses aimed at predominantly runners and cyclists in the coming years. These might offer features like navigation and heads-up access to key metrics like pace while running or power output while cycling.
In time, smart glasses could and should become more generalist, but we expect that the first devices to really succeed might well need to specialize first. The way in which all of our devices work together is something that's just endlessly fascinating to me, and thanks to advances in all sorts of worn devices and AI, we're expecting to see some really cool new ideas about how humans interact with technology over the next few years. This is something you'll get to hear more about from my colleagues, Kane and Maria, in their session tomorrow, so make sure you come back for that. But for now, I'm handing over to Geoff who's going to wrap up the day.
Geoff Blaber
Please join us for day two, when Kester will kick off an action-packed session about connectivity, networks and the intersection of technology and health. We'll have some very special guests: Marc Allera, CEO of EE, and Stefan Streit, Chief Marketing Officer at TCL. We'll see you then.