Our top tech trends for 2025: what’s coming next

It’s the start of a brand new year, and here at Osedea we’re fondly looking back but also looking forward.  We’re excited not only about the projects and opportunities this year holds for our team but also about what we foresee as the Great Tech Trends of 2025. 

Augmented reality levels up

In 2015, we saw Microsoft make augmented reality more than a pipedream.  When they released the Hololens 2 in 2019, they essentially provided affordable and production-ready solutions to the factory floors of SMBs.  

We know that manual interventions are not going away.  On factory floors and elsewhere, we will always have need for training, quality control, hands-free commands, remote monitoring, etc.  We believe that in 2025, AR tech can be the force multiplier for workers.

Last year, the Apple Vision Pro and Meta Orion made headways, signaling that true competitors were entering the professional (non-gaming) AR market.  And while AR has been 5 years away from the Early Adopters-stage for nearly a decade now, we’re betting 2025 will see it evolve from that fun gadget you saw at CES to the solution you’re adopting for a double digit increase in worker productivity.

AI tooling as a default

Since OpenAI’s ChatGPT November 2022 release, software developers have been progressively integrating AI in their workflows (in writing both code and emails… let’s be honest)!  According to the Stack Overflow 2024 survey, 62% of developers are now using AI tools in their work - a jump from 44% in 2023.  

We’re not surprised by these stats; in fact we here at Osedea have most definitely contributed to them.  For a while now, our developers have been set up with their own Github Copilot seats.  Just like any tool in our kit, we’ve been intentional as a team in learning how best to leverage AI in our day to day workflows and we’ve found that using AI for the boring stuff is the real productivity boon.  We most definitely foresee AI tooling continuing to revolutionize our work as developers in 2025.

Software Design is difficult to automate - writing code less so.  Given this and our own experience with AI tooling, we foresee the work of developers shifting to a supervisory role in the development process.  Developers will focus on identifying the higher-level design patterns of the software while the code will be more and more generated by AI.  We don’t think this means developers will become less important; rather, we believe it means developers will be able to focus on the multi-disciplinary hard stuff that really matters. 

Agentic AI on full throttle

At Osedea, we’ve been working with both custom AI models and fine-tuned open source models for years and where 2024 got us interested in the Agentic Concept for LLMs, the last few months have made it possible to create low-cost high-return solutions through recently released Agentic AI services.  

Both Microsoft’s Agentic AI and Salesforce’s Agentforce 2.0 late 2024 releases have been greatly facilitating the development of complex LLM-powered solutions that can be tailored to niche business needs.  We think 2025 will only see these advances continue.  As we’ve seen over the years, rolling out complex AI solutions can carry significant costs.  These managed Agentic AI services are now making it possible to construct robust custom solutions that can provide a competitive edge while at the same time reduce development costs.  

On-premise AI computing

AI development differs from traditional software development in the sheer amount of compute resources it requires.  Now we love cloud computing, we really do.  The flexibility is and will remain unmatched.  However, it’s not always easy to justify sending data to the cloud - particularly when working in sensitive sectors.  Data-security compliance is not trivial and can often mean that relatively small AI projects quickly become multi-departmental headaches.  

Enter on-premise AI computing - another 2025 tech trend we see incoming.  

Prebuilt on-premise workstations are extremely expensive for small projects with the cost savings only appearing after a year(+) of continued use. Custom-built workstations are often preferred by small shops but assembling consumer hardware brings its own set of concerns.

NVIDIA’s Project DIGITS was announced on Jan 6th as a viable solution, a prosumer-grade device at a lower price point than custom-built workstations. At only $3,000 USD, we expect this to become the default choice for budding AI teams and PoC deployments.

We’re certainly going to be ordering a few here.

Quantum computing breakthrough

We just had a little quantum computing lunch and learn here at Osedea.  Needless to say, quite a few of us are watching this space excitedly.  

Google’s recently announced Willow is making very concrete steps in the right direction for quantum computing, with their quantum error correction being proven to be more than a theory and moving them forward on their Quantum AI roadmap. 

Now solving computational problems isn’t very applicable in the day-to-day, but for a lot of scientists, this means that at-scale quantum computing is no longer an if but a when.  We believe this will have major implications for our society.  

The Google-originated market hype from the last few months certainly doesn’t mean Quantum Computing is ready for practical applications, but that noise did manage to infuse significant capital in the field which should bring those technologies closer to commercialization in the next few years.  Exciting stuff.  

Robotics

With the continued progress being made in AI, robots are becoming more and more intelligent and autonomous.  Major investment is being made in robotics with key players like ABB, Amazon, and Boston Dynamics as supercharged growth in the industry is forecasted.  Nvidia’s Deepu Talla has talked about “the ChatGPT moment for physical AI and robotics… around the corner.”  Their Jetson Thor which is said will transform the entire robotics industry is set to come out this year.

It’s not a secret that we love our robots, and so we are most definitely excited about these trends.  We feel lucky to have had a front-row seat to both AI advancement and robotics through our robot Spot and the various projects we’ve gotten to work on leveraging both.  From this, we see that training robots is becoming faster and more efficient with roboticists taking from AI methods. Human-robot collaboration (cobots) and use cases for such is increasing which will bring continued growth of robot usage in service sector jobs. This will cause the industry to not only grow more deeply in established domains like manufacturing but also expand to others from construction to education to healthcare.  

In closing

These tech trends excite us (yes, as self-proclaimed geeks, but also… ) as we consider how we can continue playing our part in tech, grow our expertise, and help bring major positive impact to our work, partners, and greater community. Do you have a tech project in mind? Get in touch with us to discuss how we can help bring your ideas to life.