The last couple of years have been among the most transformative I can remember in tech. After nearly two decades writing about, working in and just being a passionate observer of the industry, I’m very confident in saying this is one of the most exciting and fruitful eras I’ve ever seen when it comes to innovation and potential, as well as the nature of products actually being built and deployed.
There’s good reason to believe 2025 will be an even more exciting year – especially in the context of taking the major foundational technological shifts of the past couple of years, and turning them into things people actually use to make their lives better and easier.
Here’s a look at what I think are going to be some of the most exciting developments in 2025 in tech – and for a more firm-wide view from the team here at OMERS Ventures, be sure to look at the predictions from our data and software infrastructure team; our horizontal enterprise software team; and our fintech team, as well as Michael’s recap of 2024 and look ahead for this coming year.
General-purpose AI
This is the year that the conversation around ‘AGI’ and when it’ll happen will finally be put to bed. It’s been increasingly evident over the past couple of years that thinking about AGI as some kind of easily definable end state isn’t really useful or reflective of reality.
What we find instead is that while AI’s trajectory has resembled the development and progress of biological intelligence in some ways, in many others it’s a very different process with a very different progression arc.
There’s been some criticism from skeptics that OpenAI will simply declare that AGI has been achieved, despite the remaining obvious shortcomings of what the product they provide today inherently has. While I don’t deny that they stand to gain from a marketing perspective by doing something like this, I think it’s more fair to say that we will achieve something roughly like what we’ve thought of as ‘AGI’ to date, but that it won’t match human intelligence in all respects: It’ll fall quite short of our capabilities in some areas, and far exceed them in others.
This is the year AI achieves a level of performance and reliability where I think it becomes a default second brain or advisor to just about every computer user, in both work and personal life – thanks in part to the proliferation of OS-level AI assistant integration in Android, iOS and desktop operating systems. It likely won’t feel as significant as the year of ChatGPT’s introduction in terms of paradigm shifts, but I think it’ll actually be much more impactful in terms of peoples’ lived daily experience. Like most things, we’ll realize its impact much more fully in the rearview mirror.
Robotics
I’m anticipating a big year in robotics, in part because we’re likely to see some of the first fruits of the budding marriage between transformer-based AI and real-world physical robots. Big dollars went into this space in 2024, including the late year investment round in Physical Intelligence, and the news that OpenAI is exploring the space – not to mention Tesla’s big bet on its humanoid bot.
Most of the gains here will be in less ‘sexy’ areas like warehouse operations, logistics and shipping, where the operating environments are highly contained, mapped, measured and controlled. Non-deterministic artificial intelligence technologies being adapted to physical world interaction necessarily requires caution and an extremely gradual and staged deployment process, which is something these industries have both appetite and rails for.
One notable exception: I believe this will be a banner year for Waymo and its progress, and a highlight one for autonomous driving in general – despite late 2024 setbacks like GM’s decision to shutter its Cruise ambitions entirely. Waymo is set to expand its program in a number of different markets this year, and there are signs that Amazon’s Zoox will also end up doing a lot more in public in 2025.
Creativity
2024 ended with two major developments in creative technology: The public release of OpenAI’s Sora and the preview of Google’s Veo 2 text-to-video models. Each of these was impressive in its own way, with OpenAI demonstrating big improvements over its original Sora limited preview, launched earlier in the year, and Google showing some frankly incredible abilities in terms of recreating true-to-life people in hyperrealistic full-motion videos (albeit in a more closed beta setting). Meanwhile, X’s latest image generation capabilities are remarkable when considering specifically how well they’re able to produce photorealistic results.
Both video and image models definitely have their own hard limits, much like the transformer-based language models they share some technological scaffolding with. They’re also incredibly resource-intensive.
We’re nowhere near a world where full, production-quality movies can be made by AI, but already these are having a significant impact when it comes to providing a way to translate thought or concept into something tangible. I predict that AI-generated visual content will become a default and requisite substrate for interactions between brand and design teams – especially in a corporate and business context.
For prototyping, storyboarding and concept creation, AI will replace a process that never even really existed, outside of a scant few exceptions where extremely deep-pocketed customers were able to bankroll near-production quality prototype versions of what they eventually wanted made, taking up countless hours of time on the side of both client and provider. This will also become a must-have for any design software tool or software-enabled service provider, likely leading to some roll-up and M&A activity among startups in the field.
Energy
Where 2024 was a somewhat surprising boom year for nuclear power, I believe 2025 will be mostly about bringing that ebullience back down to earth. The realities of nuclear are that it is a highly, highly regulated environment with immense amounts of entrenched incumbent concerns on all sides, and one that faces challenges in terms of popular opinion on top of all of that. Headlines from 2024 about the world’s largest technology companies making investments here aren’t hollow, but only those companies will have the stomach for any kind of serious program, and those will have timelines that mean 2025 will likely not move the needle much relative to where we’re already at.
Instead, I think we’ll see a recovery from the current deflation around climate tech that’s much more focused on other, younger renewable generation technologies, including wind, geothermal, tidal and solar. Projects and companies that focus on new, innovative and more efficient ways to pair these renewable sources with co-located cloud storage and compute facilities, especially those that can be located further from population centers intended for tasks that can tolerate greater communication latency.
A year of ‘new normal’
If I had to point to one guiding theme to describe 2025, I think it’ll overwhelmingly be one in which we settle into a level of comfort with AI, and with having it as a persistent collaborator across most aspects of our lives. As with the advent of every major technology, the past few years have been characterized by extremes: Utopian views on the one hand, heralding an unrivaled golden age for humanity – and hand-wringing terror on the other, imagining us architecting our own obsolescence.
In 2025, we’ll do that thing we always do: Realize neither is right and find out how AI has and will actually change our lives forever, without either perfecting or destroying them in the process.
