From Nvidia challengers to fractional CAIOs: Here are our seven biggest AI predictions for 2025

Welcome back to LinkedIn News Tech Stack, which brings you news, insights and trends involving the founders, investors and companies on the cutting edge of technology, by Tech Editor Tanya Dua. You can check out our previous editions here.

2024’s penultimate edition is a special one that focuses on the bold ideas we believe will define AI’s trajectory in the year ahead. Read on for our predictions below.

AI arguably hit its stride in 2024. But what can we expect next?

In 2025, all eyes will be on how the technology’s practical applications could dramatically shape our relationship with work and life.

Every December, LinkedIn News spotlights bold ideas from our editors and experts around the world that they believe are poised to shape the year ahead.

Here are our top predictions for AI in 2025:

In 2024, AI dominated headlines with hype and speculation. In 2025, the conversation will evolve and leaders will focus instead on AI’s tangible impact on the bottom line, says Sridhar Ramaswamy, CEO of Snowflake.

The question is no longer whether to use AI, but how to harness it to deliver meaningful, measurable outcomes.

CEOs and boards must recognize that AI’s most transformative applications may arise in surprising areas. This requires a deliberate, thoughtful approach toward exploring the technology’s potential without succumbing to overinflated expectations.

Businesses must set clear goals around what they are trying to achieve with AI. Whether it’s rapid access to critical insights, faster decision-making, improved productivity or operational efficiency, companies will focus on AI’s proven use cases.

For example, AI-powered tools can now allow a business user to search through data buried within thousands of PDF documents or empower a clinician to quickly identify patterns across patient symptoms with a simple question. We will continue to see companies adopt such use cases, especially as AI agents start to accelerate our ability to automate repetitive, predictable tasks and take action on insights. As these capabilities gain traction, AI will help organizations discover new efficiencies and expand their operations.

We are at a pivotal AI moment — adapt or fall behind. Organizations will need to adopt AI with precision and a clear focus on the outcomes they desire. Those that don’t risk losing their competitive edge. 

As AI executives start demanding greater efficiency and returns on investment from their AI initiatives, NVIDIA 's iron-fisted grip on AI hardware may no longer be a foregone conclusion. 

While the chipmaker has a monopoly on graphics processing units (GPUs) — the tiny server chips that enable AI software to run — and has been one of the biggest beneficiaries of the AI gold rush so far, challengers are circling it on multiple fronts as demand far outstrips supply. 

“The days of using a GPU and brute-force compute to build models and applications will be in the rear-view mirror,” says former Microsoft executive, investor and CEO of the AI community HumanX, Stefan Weitz.

Semiconductor rivals like AMD and Intel are upping the ante with chips and AI accelerators like the MI300X, MI325X and Gaudi 3, even as other big tech players such as Amazon, Google, Meta and LinkedIn parent Microsoft invest heavily in custom chips like Trainium, Tensor Processing Units (TPUs) and MTIA chips to optimize for specific workloads and reduce their reliance on Nvidia.

But it’s not just the big names. A roster of startups like Etched , Graphcore , Cerebras and Tenstorrent are inching toward Nvidia’s turf with new approaches that enable everything from better speed and scalability to energy efficiency and optimizing for specific types of AI models, like transformers in the case of Etched.

“GPUs support many things that aren't generative AI models — whether that's recommendation models or scientific computing — making them a jack of all trades but a master of none,” says Etched co-founder and COO Robert Wachen. “The generative AI market is moving so quickly and AI workloads are maturing to such an extent that new hardware has an advantage.”

To be sure, Nvidia has several locked-in advantages when it comes to its market size, scale and distribution, including its CUDA software platform that’s almost the industry standard for building AI applications. But as AI adoption advances, 2025 may be the year that a growing band of Davids aim their arrows and chip away at Nvidia's Goliath.

The integration of AI into our daily work routines has quickly become less of a novelty and more of a cultural norm. Soon it will become a necessity — something people will begin to expect from their tools at work — and spark a productivity revolution, says Eric Yuan, founder and CEO of Zoom.

The AI we have today will evolve into intelligent digital assistants that will be able to draft email responses, ready for you to send when you sign on to start your workday, help determine which meetings to join and even assist in contract negotiations, reducing your overall workload.

Over time, AI digital assistants will become fully customized, evolving into a “digital twin,” equipped with your work history and institutional knowledge. They will fundamentally change how we work, allowing us to prioritize the creative and strategic tasks that truly require a human touch and thoughtful focus time.

Back in 1926, Henry Ford revolutionized the work week, reducing it from 48 to 40 hours. Now, thanks to AI, a four-day workweek could become the norm — enabling the fifth workday to be used as a mindful retreat from daily tasks, spent with friends, family or hobbies.

In the end, our digital twins will allow us to top Ford — and the implications for productivity, work-life balance and collaboration will be profound. A four-day workweek could truly be within reach. 

Expect to see a growing army of intelligent robots making your life easier — and a lot more automated.

Why? Because recent developments suggest that robots, once confined to sci-fi films, research labs and high-cost industrial applications, are at an inflection point — and arguably on the cusp of becoming mainstream.

After decades of slow progress, breakthroughs in AI models and sensor technology, more affordable hardware and growing availability of large-scale datasets are enabling roboticists and startups to create adaptable and affordable machines that can perform a range of useful tasks across the board.

“The combination of cheaper hardware, better components and advanced AI programming means that building commercially viable robots is now possible without massive capital investment, lowering the barrier for entry and accelerating innovation in the field,” says Ben Hemani , founding partner at Bison Ventures . 

AI-powered robotics are not only already being deployed in sectors like manufacturing, logistics and healthcare, but are also starting to close the gap between industrial prototypes and practical home robots. Companies including Agility Robotics , Tesla, OpenAI-backed Figure and Jeff Bezos-backed Physical Intelligence are a few names leading the charge here.

Using techniques like reinforcement learning and imitation learning, they’ve been able to get robots to adapt their actions and navigate cluttered spaces in real time — and even perform increasingly personalized service-oriented tasks like folding laundry, cooking and cleaning. In fact, cleaning robots are growing faster than all other use cases in service robots, according to Elad Inbar , founder and CEO of RobotLAB. 

It’s no surprise then that experts (not just Elon Musk) are betting that robots will soon hit the mass market, with Hugging Face’s co-founder and CEO predicting “at least 100,000 personal AI robots will be pre-ordered” in 2025 in this LinkedIn post.

AI is set to revolutionize workplace inclusivity by enabling people with disabilities to join the workforce like never before, according to my colleague Neha Jain Kale.  

With innovations like hands-free computer operations and screen reading for the visually impaired, these technologies will reshape how people contribute to their work — and drive outsize business success.

Excluding people with disabilities from the workforce can cost up to 7% of a country’s GDP, according to the WEF 2023 report on AI and disability inclusion. Implementing a disability-inclusive business strategy with assistive AI could lead to a 28% increase in revenue and 30% increase in profit margins for companies.

AI could also help make recruitment more inclusive and offer employees learning opportunities and processes tailored to their specific needs, including for neurodivergent individuals.

“As someone who navigates life in a wheelchair, I’ve seen firsthand how tools like voice recognition, predictive text and adaptive interfaces open doors that were previously closed,”  says diversity and inclusion leader Alister Ong. “These advancements aren’t just about convenience — they’re about giving us the autonomy to perform, contribute and thrive.”

More than 2.5 billion disabled individuals will require assistive products globally by 2030, according to a report by the World Health Organization. “With 75% of employers globally struggling to find the talent they need, embracing assistive AI for persons with disabilities offers a great opportunity for them to tap into this underrepresented and highly capable talent pool,” says Andy Bentote, a regional managing director at PageGroup. 

The AI revolution may be in full swing — but could come to a screeching halt due to a critical skills shortage, especially when it comes to the highest ranks of AI leaders.

There are only 700 qualified chief AI officers in the market, estimates David Mathison, founder and chairman of the CAIO (chief data officer) Club and Summit, who tracks the growth of the title in monthly updates. 

“It's almost impossible to find high-level strategic AI talent, especially for small and medium-sized enterprises,” he says.

One solution that he and others have brought up? Fractional chief AI officers (CAIOs) — part-time executives who are hired on a short-term or project basis to bring in strategic AI expertise, without the commitment of a full-time role. 

Taking the fractional route is not only a cost-effective way for companies to access top-tier AI talent without the rigmarole of traditional hiring processes, but also provides immediate expertise that can help jumpstart their AI initiatives, boost internal capabilities and foster a culture of AI adoption.

“A fractional executive brings the benefit of working with multiple companies,” says John Heasman, the chief information security officer of Proof. “I expect there will be several iterations of the roles and responsibilities of AI executives, and fractional executives will be able to draw on concrete examples of what works and what doesn’t.”

Building future AI leadership is not going to be just about hiring and upskilling; it will be as much about tapping into new and flexible talent solutions that can adapt to rapidly changing needs.

Move over, prompt engineering. As AI agents gain ubiquity, the hottest new role will be that of an AI systems engineer, says Scott Beechuk , partner at Norwest Venture Partners .

AI agents — which autonomously execute on a range of tasks by interacting with and learning from their environment — considerably raise the stakes, adding more complexity to how enterprise AI systems are designed, developed and maintained. 

As a result, this new quality assurance and oversight role will become essential to enterprises as they manage and continuously optimize such AI agents, according to Beechuk. 

Their prominence is only poised to increase in multi-agent environments, where a swath of different AI agents performing a variety of tasks start to interact with each other — like a Resy agent interacting with a Gmail agent to make you a reservation at your favorite restaurant depending on your calendar. 

That’s because even a small error by one agent can wreak havoc, creating a cascading effect that degrades the entire ecosystem. AI systems engineers can keep these environments running smoothly by ensuring that they’re guided by quality data and making quick interventions if things break down.

“People in these roles will have a mix of technical chops, ethical awareness, super communication skills and a non-linear mindset,” says HumanX’s Weitz. “Especially as AI jumps from vertical silos to cross-company systems, it’s key to have people who can see across enterprises and align creativity with machine precision.”

As always, thanks for reading. And click 'Subscribe' to be notified of future editions. Pitch me the interesting investors, founders, ideas and companies powering emerging technologies like AI to reach the inboxes of nearly 1 million subscribers plus thousands more on LinkedIn. Follow me for other tech updates.

What prediction or trend do you expect to have the biggest impact in the year ahead? And why? Share your thoughts in the comments or by posting a video with #BigIdeas2025.