Florals for Spring? AI Smartphones? How Original!!

Turns out, even Miranda Priestly would give in to the next hot thing on Smartphone aisles – no matter how obvious or banal it sounds. Here’s why AI Smartphones are going to be in many tech wardrobes soon.

A fresh-off-the-ramp study on AI Smartphones gives many answers that Priestly’s furrowed eyebrows and stiff upper lip might have asked. The Canalys ‘Now and Next in AI Smartphones’ report shows that by 2024’s end 16 per cent of new smartphones shipped will be Gen AI-capable and this market share of AI-capable smartphones could go up to 54 per cent by 2028 - driven by demand for on-device processing and AI agents.  We understand why AI Smartphones could be in fashion soon and, that too, beyond a fad- in this chat with Sanyam Chaurasia, Senior Analyst, Smartphones, Canalys.

The industry has hardly caught a breath after being abuzz with the idea of AI PCs. Is AI smartphone too soon of an idea? Anything different that they promise?

PCs fit well in the B2B segment and smartphones primarily lean towards the B2C side. AI in PCs is more about productivity. AI Smartphones could focus on entertainment, creativity and communication and also change the productivity aspect. They might replace some tasks on PCs in an efficient way and at a speed and impact that’s way different than current smartphones.

Would this new season of smartphones be about form factors or processor muscle or something else?

The last trend in smartphone form was that of foldables. There was not much marketing drive or FOMO effect seen there. Form factor will not do much in the AI game.

Most marketing talk about AI is around TOPS (Trillions/Tera Operations Per Second) but would vendors also wrestle on memory, power usage etc. when it comes to smartphones?

All SOCs (System on a Chip) on device have a particular AI engine. There are elements like NPU, Quantization technology and AI accelerators worth noting here. Like Qualcomm has Hexagon NPU and MediaTek has APU 790. Both have INT 4 as quantization technology. NPU tends to be highly important as it does most of the jogging in GenAI tasks on a device. NPU and GPU efficiency is also related to power.

Will the AI Smartphone concept also fix Apple Fatigue – as pointed out by many industry watchers in the last few Apple roll-outs?

My personal opinion is that Apple is not struggling that way recently- and every innovation has brought in some incremental value. Every 3 year cycle shows an added utility. Iphone 13 was a big upgrade, for instance. Apple is strong on margins and ecosystem. Even the earlier models get a market fillip when new ones come out. That’s a prominent thing in a market like India.

What surprised you the most in this study?

We are in world of collaboration – OEMs, chipset vendors, device makers. The strategic partnerships in this area would be quite crucial to come up with a product that is both efficient and effective on this concept. Galaxy is trying to do something new here with its own ecosystem. That is different from what we have seen in the Android space that usually competes with the ‘closed Apple ecosystem’ advantage.

Would the AI game in smartphone industry hinge a lot on developers and apps?

Yes, that is a strong inclination. Even the average 6-7 hours spent on any device dovetail around entertainment and creativity and that’s where users would be happy with an experience-upgrade. Some apps would ride well on contextual intelligence and some strongly on privacy that will come in with AI.

How would privacy be promised if AI wants to personalise everything on a smartphone?

When you use such a device you help it to learn itself. On-device intelligence and features will become paramount and the experience they offer will happen in an invisible background. Because AI will do everything seamlessly – as per your unique need and context- and serve it to you right on. Invisible personalisation- that’s where AI magic is. Privacy wins because the apps would be mostly personalised and on-device and not on Cloud.

Do you think this promise will arrest some flow towards dumbphones or less features or no-notifications as have been seen in some users?

Users always associate smartphones with time, which, in turn, is always about productivity. Even with social media apps, it is about some form of productivity of connections. We see a lot of vendor roadmaps and it looks like with AI smartphones your phone will look similar but will not be the same as mine. The difference in usage, context, demographic, AI assistants, multi-modal personalisation etc. – all that would make the smartphone a very personal device in a new way.

Which brands or makers will shine the most?

There are platform players like Google and Apple, of course. Then there are players like Xiaomi that are strong in IoT, smart homes and cars as categories. With Apple, things have changed with their recent AI announcements. Some vendors will offer their in-house LLMs and not collaborate with anyone. Some brands like Samsung could bring AI into the non-premium category-serving users with entry-level models as well. The ‘democratise AI’ campaign will be quite pronounced here. From an innovation angle, Samsung has gone ahead with some steps and has taken the call of creating an ecosystem. In Apple’s iOS, privacy was a corollary to the closed-system while Android etc. had a strong community advantage. With Galaxy AI- appliances, refrigerators, vacuum cleaners etc. come into the ecosystem. Apple, I guess, is silently watching and waiting for others to play their game. Chinese brands have their own agenda and limitations. Global brands will compete with their own models.

What are we missing?

The consumer angle is seldom spoken of in AI conversations. We tried that in our study. We found that India and China have a tech-savvy population so there can be AI inflection here. The US market is about how the market adopts certain features. Latin America is more about entry-level products. Features made in developed markets will find their way in emerging markets. There is also the business model angle which is interesting to note- how Telcos and enterprises will use AI for new services, revenue, ecosystems and partnerships.

-By Pratima H