3 Trends Defining the Next Wave of Mobile Experiences

It’s 2025, and despite what some predicted, app development is very much alive. In fact, it’s thriving.

There was a moment a few years ago when people started questioning whether the world really needed more apps. But the reality is that technology has moved so fast in the last 12 to 18 months that entirely new categories of applications are now possible. Tools that were once experimental are now production-ready. And every new unlock—whether it’s AI, spatial computing, or hardware acceleration—creates new opportunities for developers and businesses to build digital products that weren’t feasible even a year ago.

At Morrow, we’ve had a front-row seat to these changes. We’re seeing a wave of new apps being developed—not because people are reinventing the same things, but because the possibilities have expanded. Here are three trends we think are worth paying attention to.

‍

1. On-Device LLMs: AI That Doesn’t Rely on the Cloud

‍

We’re entering the era of truly local intelligence.

While apps have been integrating cloud-based AI tools like OpenAI and Anthropic for a while, there’s been a noticeable shift: companies are now building or fine-tuning their own models, small language models (SLMs) that are optimised for their specific use cases, and running them directly on the device.

Why this matters:

Offline-first is back: Not all users are online all the time. From fieldwork to frontline roles, apps need to work wherever users are.
Privacy and speed: Keeping data on-device means faster responses and fewer privacy headaches.
Domain expertise: General-purpose models are great, but nothing beats a model trained on your own unique workflows and terminology.

Thanks to tools like Ollama, MLC, and Apple’s Core ML stack, running LLMs locally has never been easier. We expect this to be foundational to a growing number of apps over the next 12–24 months.

‍

2. Computer Vision at the Edge: Cameras That Actually Understand

‍

The leap in on-device AI is also bringing computer vision into a new era.

Previously, running real-time vision tasks meant streaming images to the cloud. But with better chips and smarter models, we’re now seeing teams deploy computer vision directly to mobile devices.

That means:

Detecting objects and environments without relying on a server
Processing video in real time, even offline
Building vision features that are fast, secure, and fully self-contained

And since every smartphone has a camera, the use cases are endless, from identifying equipment in the field, to retail product scanning.

For companies with niche datasets, being able to run their own CV models locally is a massive unlock. The camera isn’t just for taking photos anymore, it’s a real-time interface with the world.

‍

3. Augmented Reality That Finally Works

‍

AR has been a buzzword in mobile for years but for the first time, it’s genuinely delivering on its promise.

We’re seeing a noticeable uptick in clients building new AR-first apps and expanding existing apps to include AR features. And it’s not hard to see why:

On-device AI and computer vision make AR more context aware
Newer smartphones can now handle high-fidelity AR scenes without breaking a sweat
And most importantly, AR is becoming a bridge to the spatial computing future we’re all headed towards

We’re also seeing businesses experiment with mobile AR today in preparation for headsets arriving in the next 18 months. Whether you’re building for learning, logistics, ecommerce or entertainment—AR now makes more sense than ever.

(And yes, we’re more than ready for this shift, our background with ReactVision means we’ve been thinking about cross-platform AR/VR long before it was cool.)

‍

Final Thoughts

‍

The app ecosystem hasn’t dried up, it’s evolving. Rapidly.

We’re building apps today that simply weren’t possible 2–3 years ago. On-device LLMs, edge-based computer vision, and truly performant mobile AR are no longer futuristic, they’re production-ready, and they’re changing what’s possible.