‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ 
 
 

2024: The year that AI grew up

 
 
 
 
 

A note from Think Editorial Team:

 

It's the end of the year—congratulations! You made it. As we prepare to welcome in 2025, let's take a look back at this monumental year in AI—including the open source surge, the race to AGI, agentic advances, governance and much, much more.

What's in store for the year ahead? Our IBM experts share a few thoughts below, but overall, the Think editorial team plans to put the crystal ball down and let the stories speak for themselves. (Though we may ask Granite for a few 2025 cheat codes.) Until then—cheers!

 
 
 
 

The most dynamic year in AI (so far)

"It's such a great time to be alive," says IBM Consulting VP Shobhit Varshney as he reflects on this year's AI breakthroughs. "We're getting to learn how these models learn, and that becomes a good, symbiotic relationship."

 
 
 
See the episode
 
 

This week, our veteran SMEs get together for a special edition of Mixture of Experts, taking a look back at the biggest leaps and bounds we saw with AI models, agents, hardware and products—and forecasting what we can expect going forward.

 
 
 
 

How to win with AI in 2025

 

A new survey from IBM and Oxford Economics shows that 63% of leaders expect AI to boost their bottom line in the next couple of years. But while the excitement around agentic AI is high, many organizations are still figuring out how to make it work for their needs.

"The shift towards agentic AI is redefining leadership," says Robert Zabel, a Research Lead at the IBM Institute for Business Value (IBV). "Leaders are no longer just defining the destination, but also mapping out the path, enabling teams to leverage AI agents in innovative ways."

Agentic AI is just one of the topics that IBM explores in its 5 Trends for 2025 report. Based on surveys of more than 43,000 execs and 4,000 consumers, the report also examines how AI is impacting technical debt, locations of operation, IT budgets, and product and service innovation around the world. It also offers practical, research-backed advice to help leaders leverage AI and come out ahead.

 
 
 
 

What will be the most important AI ethics issue we need to address in 2025?

Literacy—that is, the ability to understand, use and evaluate AI so that it works for everyone, not just developers and data scientists. IBM's Phaedra Boinodiris shares more.

 
 
 
Get the insights
 

Clean energy, extreme weather and automation on tap for the year ahead

 

The sustainability sector is in the midst of a transformation. As AI spurs innovation and industries adopt cleaner practices, change is coming—and fast. Ahead of the new year, we asked industry experts what sustainability trends they will be watching in 2025.

Chris Hopper, CEO and Cofounder of Aurora Solar: "There will be continued growth in the clean-energy sector, driven by key factors like the IRA [Inflation Reduction Act] and the rapid expansion of solar, battery storage and EVs."

David Blanch, Director of Product Management, ESG and Environmental Intelligence at IBM: "Extreme weather and climate-related crises will continue in 2025, leading to costly disruptions."

Kendra DeKeyrel, VP of ESG and Asset Management Products Leader at IBM: "In 2025, organizations with sustainability ambitions and targets should implement AI-powered automation capabilities, including observability, resource management and application lifecycle management."

 
 
 
 

AI, Robot: A year in review

 

"The next wave of AI is physical AI," NVIDIA CEO Jensen Huang said in June. "AI that understands the laws of physics. AI that can work among us."

In 2024, the robotics industry is surging, powered by lower costs along with major gains in computing power, sensors, motors, batteries and, of course, AI, which is giving the sector a major boost. Notably, humanoids are on the move—and they may soon reside not only in warehouses and auto factories but also in spaces where people live and play.

We spoke with Jose Favilla, global leader for Industry 4.0 at IBM, to get his take on the biggest trends in robotics this year, and what to look out for in 2025.

 
 
 

More tech news

 
 
 
 
 

Here's what else you may have missed during the busy holiday week:

 

The US Commerce Department finalized a grant award of up to USD 458 million to semiconductor manufacturer SK hynix to develop and manufacture memory chips critical to the AI supply chain. Funded by the CHIPS and Science Act, the grant will support the construction of a memory packaging plant for AI products and an advanced packaging fabrication and R&D facility in Indiana.

NVIDIA completed its acquisition of AI infrastructure software firm Run:ai. In a press release, Run:ai's founders said they plan to open source the company's software to "[empower] AI teams with the freedom to choose the tools, platforms and frameworks that best suit their needs."

Google introduced Gemini 2.0 Flash Thinking Mode, a new experimental model that unlocks stronger reasoning capabilities and shows its thoughts.

China launched 10 internet satellites into space. They are the first in a series of nearly 13,000 units that will eventually form a broadband mega constellation in low earth orbit (LEO), similar to SpaceX's Starlink.

Chinese firm DeepSeek released DeepSeek-V3, a new open-source model trained on 14.8 trillion "diverse and high-quality" tokens.

 
 
 

Final thoughts

 
 
 
 
 
 

2024 is almost out the door. Let's look back at some of AI's weirder moments this year.

First off, 2024 saw the release of a number of odd, AI-enhanced gadgets. Case in point: This USD 4,799 pair of birding binoculars. And a bidet seat you can talk to. And a GPT-powered talking head that seems like nightmare fuel.

Meanwhile, Casio introduced Moflin, a fluffy AI pet that is less creepy than a Furby but will rate your attentiveness to it on an accompanying app. No pressure or anything.

Can AIs write decent jokes? Actually, they kind of can. They will also laugh at your jokes. This is all a bit more weird-funny than haha-funny, though, so maybe just ask Alexa to make you giggle instead.

In our opinion, though, the best use of AI this year comes from UK mobile provider O2, which developed an AI "Granny" called Daisy designed to waste scammers' time on the phones, sometimes for as long as 40 minutes. It's both weird and helpful! Win-win.

 
 
 
 

Was this email forwarded to you? You can subscribe here