5 AI Predictions for 2025 (AI hype dying; real opportunities rising)

Hey there,

It's 2025 and as I look at the greater landscape of AI, I've got 5 predictions on what you can expect this year.

Before we dive in, here's the larger context on why this is relevant whether you're an AI consultant, working in or on a business that's being affected by AI, or an AI enthusiast that likes to stay in the know:

The AI Economy will add $15 trillion to global GDP by 2030.

That's the big number from "Sizing the prize" – PwC's Global AI Study. They're not the only ones who see it either. The 1,300 analytical folks over at the International Data Corporation quote $19.9 trillion by 2030.

Here's the thing, if you dig into either report, you'll see the trendline of growth going up and to the right.

Translation: The AI economy rocket ship isn't slowing down 😎🚀📈

Now, if you heard this tidbit from Google CEO Sundar Pichai, that might seem a bit strange:

💬 "When I look at [2025], the low-hanging fruit is gone…the hill is steeper ... You’re definitely going to need deeper breakthroughs as we get to the next stage."

Heck, it might seem like he's signaling a slow-down…unless you're familiar with the 3 kinds of technology that shape every industrial revolution:

1–Emerging Tech: Emerging technologies introduce new capabilities.
​2–Adjacent Tech: Adjacent technologies enhance the ecosystem, enabling broader adoption and innovation.
​3–Disruptive Tech: Disruptive technologies redefine industries, creating winners and losers.

In the 2nd industrial revolution, the emerging tech was electricity, while adjacent tech was the light bulb, and disruptive tech was the elevator.

When you combined all 3, what the world got were the first skyscrapers – reshaping the urban landscape for over a century to follow.

Now, if you re-read Sundar's quote about the "low-hanging fruit" being gone, then it becomes clearer that he's probably referring to emerging AI tech.

When it comes to the AI Economy though, that +$15 trillion by 2030 is likely to come from the innovation and application of adjacent/disruptive AI tech.

Which brings me to my 5 AI predictions for 2025.

As I go through them – ask yourself, is this an emerging, adjacent or disruptive tech thing?

Whether you're an aspiring AI consultant, AI enthusiast, or owner of a business that's being shaped by AI – thinking this way allows you to see above the "AI forest" instead of a single "AI tree" (read: AI tool) blocking your view.

Let's get into it.

Prediction #1 – The Fall of Coding 📉

By the end of 2025, 80% of coding tasks will be doable by non-coders using AI tools.

Developers who once guarded the gates of tech enablement will be disrupted by tools like Open AI’s o3 which is already scoring in the top percentile of coding assessments.

Basically, coding is shifting further down the value chain. For coders, that means you'll need to add new services to move up, or even stay in the same place.

In 2024, we saw massive layoffs in Silicon Valley. It's also becoming much harder to get jobs in tech.

While this makes the landscape harder for coders, it's a win for 'noncoders'.

We're entering an age where you don't need to take years to learn coding skills. Instead, you can 'skill up' in under 10 hours and do just about everything a coder could've done years ago.

For our Innovating with AI students, this means more opportunities for no-code and low-code building. If you get stuck in some part of automation, you can just use your AI coder to do the job.

Prediction #2 – A Hollywood-quality movie (made with AI) ▶️

We’ll likely see a Hollywood-quality movie created entirely by AI—scripts, visuals, acting, the whole thing.

It probably won't be from a traditional Hollywood studio because AI is causing their employees very high levels of existential dread.

Rather, it might be from the 'creator economy' of YouTubers, who already produce 1-2 hour documentaries. Or, it might be from a slice of the world we'll never see coming like a bunch of kids with a lot of time and curiosity on their hands.

With Sora becoming more accessible, they may use that…or the many eventual competitors it's sure to have. Google Veo 2 will likely be one of them.

While current AI video generation can occasionally produce stuff in the 'uncanny valley' that makes your skin crawl, the tech will only get better with time.

Case in point – this Coca-Cola commercial on YouTube, produced entirely by AI. It's got over 300,000 views and shows we're already on the way there. Huge stuff, since Coca-Cola may spend an average of $1M to create their usual commercials.

At the end of the day, this is already changing assumptions about what it means to be a filmmaker. Great stories will become even more important, while production value (and to some degree acting) may become less important.

Prediction #3 – The Rise of AI agents 🤖

AI tools are software apps with AI that perform tasks (under human direction). They're like your car's cruise control – it does a specific thing but still requires a decent amount of input and guidance.

AI agents, on the other hand, are sophisticated systems that can operate autonomously – i.e. they have a degree of agency. If we continue the car analogy, then it's more like a self-driving car.

I'll have a more detailed explainer coming your way next week but for now, picture this – you could type or speak a prompt to an AI agent, then it'd take control of your mouse and keyboard – opening up applications, typing in spreadsheet, finding stuff online, and more, all to complete whatever 'job' you gave it to do.

The stuff that's in beta right now is literal Star Trek stuff and yet, it's not science fiction anymore – it's here and we're going to see a lot more of it.

Prediction #4 – AI is getting better, faster, cheaper (again!) 📈

AI’s price-to-performance ratio keeps improving. By the end of 2025, we’ll see multiple major players (Google, Meta, OpenAI, and Amazon) rolling out models that are not only more powerful but also less expensive to deploy.

Side note: I also think Amazon remains the sleeper company. At the risk of putting myself out there a little too much, if I had to bet on a surprise winner in the “model wars” it would be Amazon, because they have tons of infrastructure experience from AWS and are well positioned to benefit in unique ways with their shopping and Alexa lines of business.

I mean, just look at their $4 billion investment in Anthropic. That doubles their stake to $8 billion. Also, the deal commits Anthropic to using Amazon's custom-designed chips for training and deploying advanced AI models. While Microsoft captured headlines with their OpenAI partnership, Amazon is building a comprehensive AI stack – from silicon to software.

(Now, if they also create an incredible AI model, sheesh, that would be a pretty amazing coup.)

Speaking of infrastructure though, there's a lot of media fuss about the “cost” of AI computing power. It is a thing – and I also think you'll see datacenters scaling up their solar power (and so using cleaner forms of energy). And since recent human history is a story of inventing more powerful forms of energy to power ambitions, I think you'll see more of this as well.

On the software side of things, you've got companies like Zapier and Make.com, which you could say were "B-rate" tech companies. Now, AI has leveled up their software and you've got a lot of other no-code workflow builders getting into the game.

Prediction #5 – The AI "Gold Rush" continues without government regulation ⚖️

Governments aren’t going to be able to put meaningful guardrails on AI development. Under Trump (with influence from Musk), expect deregulation.

The US tech influence will make it harder for Europe to regulate AI. While some residents are getting cut off from the latest AI tools (like Sora), it seems unsustainable for Europe to intentionally be an AI Luddite.

So, basically there’s no future where AI regulation stops companies from doing whatever they want.

When you combine this with the extremely high economic incentive to create better language models, then you end up with an all-out race to build more useful stuff.

Ultimately I think this ends up with a better outcome for consumers while simultaneously increasing long-tail risks – which I think are mostly about data privacy rather than about robots taking over the world.

My personal stance: I strongly support regulations by the FTC around fair trade. I'm a left-of-center guy and always root for the policy that will be most fair to the most people rather than make the most money for tech/finance moguls.

However, I think also think Europe has repeatedly erred in rushing to codify privacy and AI regulations that make no sense in practice, so I am skeptical of the EU's ability to lead the world on AI law.

Regardless, meaningful AI regulation is not in the cards in the US for AI for at least four years (and probably wouldn’t be with Democrats in charge, though they might pay lip service to the idea). So, government won’t provide meaningful guardrails or slowdown. Which means that as long as AI doesn’t turn into The Cylons from Battlestar Galactica we’ll probably see lots of cool stuff moving very rapidly.

Putting It All Together

The underlying emerging AI tech will continue to accelerate. As a direct result of that, the adjacent tech environment (which many AI enthusiasts and AI consultants operate in) will continue to skyrocket – whether you're on the technical or creative side of things.

As for disruptive tech? You can bet that's on the way too – whether it's AGI (artificial general intelligence) or something else we're not even considering. I mean, heck, who would've predicted the leap from light bulbs to elevators and skyscrapers?

Here at Innovating with AI we're excited about the future, and consider it an honor to be at the forefront of it all with you.

AI is an engine of growth, and the AI economy is real.

And with the AI economy skyrocketing, so will the demand for AI consultants who are ready to lead their clients into the 4th Industrial Revolution (which includes AI).

If that's your cup of tea, then tomorrow's email is for you because I'll be sharing the story of how I invoiced $1.9M from a single client (something I usually share with just students inside our programs).

For now, hit reply – would love to know if any of these predictions were on your AI bingo card for 2025.

To your best year in AI yet,

– Rob
​

PS – if you're curious about starting your own AI consultancy, then click here to join our waitlist. We'll be opening up our virtual doors in a few weeks.