Fact Checked by 
It’s that time again—when tech experts share their predictions for where innovation will take us in the coming year. AI is set to play a starring role once again. Back in December 2023, Bill Gates made a bold claim about AI’s trajectory, saying, “we are 18-24 months away from significant levels of AI use by the general population.” Looking at how ChatGPT has taken off and how Apple, Google, and Microsoft are weaving AI into tools for both consumers and businesses, it’s safe to say he wasn’t far off.
Now, as we look ahead to 2025, another big name in AI has stepped up with some intriguing forecasts. Clem Delangue, the CEO of Hugging Face—a popular AI platform used by millions of developers and major players like Intel and Qualcomm—has shared his predictions for the year. Among his most surprising? He believes 2025 will see “the first major public protest related to AI.”
Delangue’s track record is solid—his 2024 predictions, like rising awareness around the monetary and environmental costs of AI, were spot on. He shared his latest predictions on LinkedIn and X in a straightforward list. However, he didn’t dive deep into how people might react to the rapid evolution of these technologies.
Six predictions for AI in 2025 (and a review of how my 2024 predictions turned out):
– There will be the first major public protest related to AI
– A big company will see its market cap divided by two or more because of AI
– At least 100,000 personal AI robots will be…— clem 🤗 (@ClementDelangue) December 2, 2024
With that in mind, here’s our take on what to expect in the world of AI in 2025.
AI protests and copyright issues
Recent controversies around AI adoption suggest growing tensions that could spark major pushback in 2025. Whether it’s protests over AI taking jobs en masse—imagine something similar to that of the Occupy Wall Street movement—or concerns about how police, government agencies, or healthcare systems might misuse AI, the stage is set for a clash of perspectives.
Here’s my one prediction for AI in 2025 (& a review of how my 2024 prediction turned out).
2025: AI co’s will continue to ignore creators’ objections to unconsented use of their works to train AI.
2024: AI companies will use creative works without the consent of creators. 💯 https://t.co/YOZIFpzmNM
— neil turkewitz (@neilturkewitz) December 2, 2024
Another big spotlight in 2025 will be on the intersection of AI and copyright. In the U.S., a wave of lawsuits over AI and intellectual property is already in motion. George Washington University has compiled an extensive overview of these cases, showcasing just how complex this area has become. Meanwhile, Europe has been slower to see AI copyright disputes land in court. That began to shift in 2024, and key appeals are expected to be decided in 2025.
The groundbreaking EU AI Act, which officially came into force in August 2024, haven’t kicked in yet, so companies have been hustling to prepare.
In February, Chapters I and II of the AI Act took effect, banning AI systems that manipulate consciousness or use deceptive techniques to coerce decisions people wouldn’t normally make.
By May, non-binding codes of practice will be released to help companies better align with the AI Act’s requirements and clarify its interpretation.
Next August, a broader set of rules will come into play, including regulations for general-purpose AI models, oversight of notified bodies, governance standards, confidentiality measures, and penalties for non-compliance.
The year ahead promises to be crucial for shaping how AI is integrated into society while balancing innovation, ethics, and accountability.
AI agents
2025 is shaping up to be a breakout year for AI agents. If you’re not familiar, AI agents are like personal assistants powered by artificial intelligence. They can assess their environment, understand interactions, and tackle specific tasks for you.
These AI agents are set to become even more prominent as the major players in tech ramp up their conversational AI wars. Digital assistants like Alexa, Siri, and Google Assistant are getting serious upgrades. Alexa will integrate Claude, Siri is set for yet another overhaul, and Google Gemini will fully take over Google Assistant, turning it into a conversation-first powerhouse.
Sam Altman says AI agents are going to "really transform things" in 2025 with the ability to use tools and create things of value pic.twitter.com/0Rymh7lWzc
— Tsarathustra (@tsarnick) December 5, 2024
While traditional large language models (LLMs) like ChatGPT work by generating responses from pre-existing training data, AI agents take things further. They can access external tools and systems to gather fresh, real-time information. This doesn’t make them fully autonomous—they still depend on human design, training, and input—but it gives them a huge edge in functionality.
AI agents are great at breaking down a query and figuring out how to deliver the best possible solution by using various tools (including other AI systems). For example, imagine planning a business trip: an AI agent could coordinate your flights, accommodations, and agenda, and even consider factors like traffic and luggage restrictions. It doesn’t just suggest options—it could also go ahead and book everything for you, creating a seamless, efficient experience. OpenAI CEO Sam Altman agrees, stating that AI agents are going to “really transform things” in 2025, and thinks that artificial general intelligence (AGI) is just around the corner, with a breakthrough to come next year.
Putting a face to AI
Bryan Catanzaro, Nvidia’s Vice President of Applied Deep Learning Research, predicts that 2025 will bring AI that’s not just smarter but also more emotionally responsive, creative, and diverse. “Rapid advances in AI are making image and video outputs much more photorealistic, while AI-generated voices are losing that robotic feel,” he explains in a blog.
AI videos will go mainstream by 2025.
Not just because they’re hyper realistic or can mimic actors.
But because AI is changing to game for creators, brands, and storytellers everywhere.
10 wild examples 🧵: pic.twitter.com/8WR9OsYc36
— Min Choi (@minchoi) November 8, 2024
This leap forward will come from better algorithms, richer datasets, and a growing understanding among companies that for AI to resonate with eight billion people, it needs a relatable face and voice. The result? A shift from rigid, turn-based AI interactions to fluid, natural conversations that feel more like talking to a person than exchanging inputs with a machine.
Microsoft’s AI CEO, Mustafa Suleyman, shares a similar vision. In a recent report by ReadWrite, he talked about a future where AI has “infinite memory” and evolves into an extraordinary teacher, doctor, and life companion. “That’s why it’s a copilot,” he says, referring to Microsoft AI’s Copilot project, which is designed to act as a “consigliere”—a wise advisor helping people achieve their full potential.
What sets Copilot Vision apart, he says, is its ability to behave like a genuine friend rather than a simple yes-man. Suleyman highlights how powerful it is when AI occasionally pushes back, offering constructive feedback instead of just going along with everything. “When it occasionally pushes back on you, that’s a profound moment because a true friend would do that,” he says.
AI has a reverse effect on financials
Delangue predicts a major shakeup in 2025 for tech firms, suggesting that a “big company will see its market cap divided by two or more because of AI.” His point? A breakthrough in AI could upend a core technology or business model for a major corporation—similar to how the internet disrupted the newspaper industry by gutting print advertising revenue.
While the AI industry has enjoyed years of relentless optimism, 2024 was a reality check for some big players. A few AI-related companies faced sharp declines in their stock market performance. That said, many bounced back quickly, highlighting the resilience and confidence in the sector. With tech giants like Amazon and Microsoft making massive investments in AI, there’s no doubt that the technology will continue to grow rapidly in 2025.
Gain insights from NVIDIA experts on how industries such as healthcare, telecom, energy, automotive, retail, entertainment, and robotics will be utilizing #generativeAI in 2025. #AgenticAI https://t.co/TuZkmGw27V
— NVIDIA (@nvidia) November 26, 2024
Take Nvidia, for example. The company, which designs GPUs and produces most of the chips powering AI systems, saw its stock drop by 25% in mid-2024. But in the bigger picture, this was just a blip—Nvidia’s stock had skyrocketed by about 200% earlier in the year and later recovered, hitting new highs.
Alphabet, Google’s parent company, experienced something similar. Its stock peaked in July 2024, took a steep dive in October, and has since started to bounce back. Alphabet’s heavy investments in AI, including its autonomous Waymo taxi service and the Gemini AI chatbot, are clear indicators of its commitment to staying at the forefront of this evolving space.
The message for 2025? While there might be disruptions, the AI industry isn’t slowing down—it’s gearing up for even bigger leaps forward.
AI chip wars
Let’s talk about AI chips. The race to dominate the chip market is heating up, and 2025 is going to crank things into high gear. All the major players—Intel, Nvidia, Qualcomm, and Apple—are battling for the top spot in power, efficiency, AI capabilities, and battery life. Intel is entering a tricky phase, and rumors are swirling that Nvidia might finally jump into the desktop CPU market. But let’s be honest—Apple is still the one setting the gold standard.
Apple completely redefined expectations with its M1 chip back in 2020, leaving every other chip maker scrambling to catch up, especially in laptops (and to some extent, desktops). Qualcomm made a solid play in 2024 with its X Elite line, arguably the best response to Apple Silicon so far. However, it’s still unclear if devices running on X Elite are hitting it big just yet. What’s obvious, though, is that Intel, Qualcomm, and Apple are finally delivering what laptop users care about most: power and efficiency.
DRAM memory chip industry sales are seen rising 36.3% year-on-year in 2025, as bit shipments rise 24.7%, media report, citing market researcher TrendForce. Average selling prices (ASPs) will rise 17% in 2025, with AI memory chips called HBM (high bandwidth memory) seeing price…
— Dan Nystedt (@dnystedt) December 4, 2024
Looking ahead to 2025, we can expect some big leaps forward. Battery life for both Windows and Mac laptops is projected to soar to nearly 30 hours—a huge win for anyone who’s constantly on the go. On top of that, the next generation of chips will make running large generative AI models locally a breeze, opening up exciting new possibilities for portable computing.
Even with all this innovation, it’s hard to imagine anyone dethroning Apple as the leader in this space. They’ve consistently raised the bar, and I wouldn’t bet against them leading the way again in 2025.
AI robots (not to be confused with agents)
Delangue’s third prediction takes us into the exciting arena where software meets hardware: “At least 100,000 personal AI robots will be ordered,” he says. This lines up perfectly with developments from companies like Tesla and Figure. It’s also in step with Elon Musk’s ambitious vision for humanoid robots, including plans to deploy them on Tesla production lines.
That said, don’t expect to see intelligent robots instantly flooding homes, restaurants, service areas, or factories. These use cases aren’t quite here yet, but they’re closer than you might think. With aging populations and shrinking labor pools, governments and industries are increasingly exploring physical automation as a solution.
The rollout will be gradual, but in 10 years, personal and industrial AI robots could become as commonplace as the iPhone is today. It’s a fascinating glimpse into the not-so-distant future, where robots could become part of our daily lives.
The end of wearables
Here’s the thing – every AI wearable maker who tried to sell us a standalone AI device in 2024 needs to hear this loud and clear—it was the wrong move. Their struggles proved a simple truth. While artificial intelligence is a game-changing feature when integrated into larger apps and products, it just doesn’t hold up as the sole basis for a dedicated device.
Let’s face it, tools like ChatGPT, Copilot, and Gemini shine when they’re part of the devices we already love—our phones, laptops, and tablets. Nobody’s clamoring to carry or wear yet another gadget just for AI.
The standalone devices like Humane AI, Rabbit R1, Friends, and Plauds? They’re not making it through 2025. The silver lining is the AI powering these gadgets—their “brains,” if you will—will probably find a second life as apps on our trusty iPhones and Androids, where they truly belong. It’s all about keeping AI where it’s most useful – right at our fingertips.
Assisting other industries
Delangue believes there are some interesting breakthroughs ahead in AI for biology and chemistry, especially in areas like drug molecule discovery. He also suggests that China will “start to lead the AI race,” a development likely to catch the attention of the U.S. government and others keeping a close eye on global AI dominance. On a personal note for Delangue’s company, he anticipates its user base of “AI builders” will more than double, growing from 7 million to 15 million in 2025.
We’re also poised to see AI make big strides in construction, engineering, and design, where specialized generative AI models will boost efficiency and spark innovation.
AI may help analyze vast amounts of data from onsite sensors and cameras to streamline project timelines and budgets in construction. By continuously evaluating reality capture data (like lidar and photogrammetry), it will give insights that improve quality, safety, and compliance. The result? Fewer errors and safer worksites.
For engineers, AI will use physics-informed neural networks to speed up solutions for flood prediction, structural engineering, and airflow optimization. This means faster, more precise design iterations tailored to specific spaces, from a single room to an entire building.
In design, tools like retrieval-augmented generation will help architects ensure compliance with local building codes right from the start of the design process. Meanwhile, diffusion AI models will make conceptual design faster and more dynamic.
Featured image: Ideogram
