The 3 biggest takeaways from CES 2025 (AI Everywhere)

[Daniel Nenni]

The Consumer Electronics Show (CES) in Las Vegas is coming to a close, but the year’s biggest tech event has brought us a number of intriguing new offerings that are sure to dominate the industry in the months ahead.

Nvidia (NVDA) kicked things off with a keynote by CEO Jensen Huang, while Delta (DAL) took over the famed Las Vegas Sphere to debut new partnerships and technologies for passengers.

Google (GOOG, GOOGL) announced new generative AI features for TV apps, and companies across the spectrum showed off AI-powered smart glasses.

From powerful gaming chips to self-driving cars and everything in between, these are the biggest takeaways from CES 2025.

AI everywhere​

Try as you might, you couldn’t escape artificial intelligence at this year’s CES. And chances are, you won’t be able to for the rest of 2025 either.

Seemingly every company was promoting their own AI capabilities, with Nvidia talking up its latest AI software and robotics firms demonstrating their latest humanoid bots.

Nvidia CEO Jensen Huang delivers a keynote address at the Consumer Electronics Show (CES) 2025. (Artur Widak/Anadolu via Getty Images) · Anadolu via Getty Images

LG said its AI features for smart appliances offered “affectionate intelligence,” while Samsung revealed its “AI for All campaign.” Agentic AI, AI-based around agents that can control multiple foundation models to help cut down on mundane tasks, was also on full display with companies large and small boasting about the technology.

Companies also rolled out a slew of AI-powered smart glasses, intelligent lawnmowers, robots that can grab candy off the shelf, and friendly humanoid bots that’ll offer up a hearty handshake.

It’s going to be some time before robots make their way into our daily lives, but the tech industry is banking on them becoming mainstream.

PC chips get the AI treatment​

AMD (AMD), Intel (INTC), Nvidia, and Qualcomm (QCOM) each announced new consumer and enterprise chips at CES 2025. Nvidia kicked things off with CEO Jensen Huang’s keynote during which he unveiled the company’s latest superchip, the GB10, a stripped-down version of Nvidia’s server-grade GB200.

The company also showed off its latest gaming chips, the RTX 50 series graphics cards for both desktop and laptop PCs.

AMD, meanwhile, debuted its Ryzen AI Max chips for high-powered gaming laptops. Then there were the AMD AI 300 and AI 300 Pro Series chips, which are designed to run AI PC applications for both consumer and enterprise customers.

The chip company also revealed its latest Ryzen gaming CPUs for desktops, laptops, and handheld gaming PCs.

Intel, meanwhile, showcased its Core Ultra 200V line for enterprise computer systems and gaming-focused chips for consumer PCs. Qualcomm, for its part, introduced its Snapdragon X chip for mid-range Windows laptops, setting it up for even greater competition with both Intel and AMD.

CEO Cristiano Amon also teased a graphics announcement in the coming months, which could mean a showdown with Nvidia.

Autos roll out AI capabilities​

You can’t have a CES without a few new cars, and Honda (HMC) did just that at this year’s show, touting its Honda 0 Saloon and Honda 0 SUV vehicles.

Amazon's self-driving company Zoox is testing its autonomous vehicle near its headquarters in Foster City, Calif., on May 24, 2024. (Tayfun Coskun/Anadolu via Getty Images) · Anadolu via Getty Images

Nvidia announced that Toyota (TM) is using its technology to power the automaker’s driver assistance capabilities, and Amazon-backed (AMZN) Zoox gave us a test drive around the Las Vegas strip in one of its self-driving robotaxis.

While CES is wrapping up, there are still plenty of major events ahead in 2025. Amazon, Google, and Microsoft will each host their own cloud and developer conferences; Apple (AAPL) is set to debut its next iPhone; and Nvidia will host its GTC conference in March.

AI is certain to drive a huge chunk of the conversation, but be on the lookout for potential breakthroughs in other areas of tech. After all, there’s always the next ChatGPT around the corner. It’s a long year and it’s just getting started.

https://finance.yahoo.com/news/the-3-biggest-takeaways-from-ces-2025-152538888.html

 

[Daniel Nenni]

It is certainly nice to see semiconductors front and center, well deserved, but TSMC should be doing victory laps with Nvidia, AMD, Amazon, Microsoft, Google, etc... AI in autos has been a major disappointment for me. I expected fully automated cars by now. So much hype back in the day.

 

[OneEng]

Just got back last night. My back and feet are still throbbing .

AI was indeed everywhere (including our own product launch booth ), but it was questionable about how much of it was real vs gas-lighting for attention.

AI is indeed growing like crazy IMO and I personally have plans for it in our back-end. We are marketing these features today .... but it is subtle vs. robots taking over the world (as NVIDIA key note would have us believe .... btw, I think that was a questionable marketing approach).

TV's were still about picture quality and size (and flexibility). Cars are still about style, comfort and price. AI .... IMO, is about making products more useful in their current capacity, not in replacing products.

Yes, autonomous vehicles are a great buzz everywhere .... but first and foremost, they are transportation products for consumers. Customers will primarily purchase a new vehicle based on the styling, comfort, reliability and price. There aren't that many people willing to shell out 8K for Tesla FSD in a car market that consumers are already screaming about high prices in. Hard to make a business case on the 1% market.

Still, AI was everywhere. CES2025 was definitely the year of AI.

 

[Paul2]

This CES was underwhelming, a very low number of products announced, and most were respins of previous designs, rather than new RnD. High inventories, people decided to properly sell out old stock this year. A long lull, akin to Skylake era, when we had years, and years of boring laptops, when previous models looked exactly like one from 5 years ago in everything except for CPU model.

What also needs mention: the number of gadgets also went down, and that signifies consumer sentiment even more than big name purchases. I am bullish on the strategy people used in 2016-2019: quick selling high spec products at great discounts, made from small lots of parts bought at bargain pricing.

Take a look what the sole well doing newcomer to the smartphone arena is doing: HMD.

 

[jiakun0618]

I have been at CES for a few years, and this year it seems many companies are missing, noticeably some leading auto OEMs. Some of the space feels empty. At the same time, it has a record number of Chinese companies attending it. It's interesting to see some Chinese Lidar companies showing presenting next to the US ones. Such a contrast. I went to Lenovo's private showroom, and it seems they are still innovating. I think the combination of camera, multi-modal models, robotics and edge computing can lead us to many more possibilities.

I also talked with some electronics manufacturers. AI glasses and AI toys are really hot. It's easy to design and manufacturing the AI-powered toys and I can image the market size is large.

It's amazing to see that Lidar is being used on autonomous mowing machines. According to the CEO of a Lidar manufacturer, price will drop to lower than $200 per piece this year. When the sensors and chips become cheaper, many things just become more intelligent than ever. That would be a trend for the coming years. Maybe robots are not taking over the world just now, but things are changing day by day.

 

[Daniel Nenni]

This CES was underwhelming, a very low number of products announced, and most were respins of previous designs, rather than new RnD. High inventories, people decided to properly sell out old stock this year. A long lull, akin to Skylake era, when we had years, and years of boring laptops, when previous models looked exactly like one from 5 years ago in everything except for CPU model.

What also needs mention: the number of gadgets also went down, and that signifies consumer sentiment even more than big name purchases. I am bullish on the strategy people used in 2016-2019: quick selling high spec products at great discounts, made from small lots of parts bought at bargain pricing.

Take a look what the sole well doing newcomer to the smartphone arena is doing: HMD.

The last CES I attended was Asia dominated. Is that still the case? Did China companies have a big showing? Samsung is always big at CES. How about Japan?

 

[Xebec]

Another CES take-away is how many times AMD, Intel, and Nvidia said "AI" per minute of keynote.

I think AMD wins because they have "AI" in their product names.

AMD: 2.2
Nvidia: 1.54
Intel: 1.45

 

[Paul2]

The last CES I attended was Asia dominated. Is that still the case? Did China companies have a big showing? Samsung is always big at CES. How about Japan?

Mainland China did not make a big showing, and the atmosphere of stagnation in the mainland tech industry is easily readable in the air, despite the effort to hush it

Mainland is a technical laggard, and it only stopped looking like that from the Western point of view simply because they had some products to show at all at some point, when they had none before. But of course, it's still very big, but nevertheless a laggard keeping importing technology, rather than creating it.

The last place where I seen Windows 98 in my life was a mainland bank office in Shenjhen

 

[blueone]

Yes, autonomous vehicles are a great buzz everywhere .... but first and foremost, they are transportation products for consumers. Customers will primarily purchase a new vehicle based on the styling, comfort, reliability and price. There aren't that many people willing to shell out 8K for Tesla FSD in a car market that consumers are already screaming about high prices in. Hard to make a business case on the 1% market.

Just a personal opinion, but I think autonomous vehicles will be "next year" for more than a few years yet. I've been thinking for years that human drivers and autonomous vehicles are a very risky and unpredictable mix. I wouldn't engage the capability if the feature was free.

 

[Xebec]

Just a personal opinion, but I think autonomous vehicles will be "next year" for more than a few years yet. I've been thinking for years that human drivers and autonomous vehicles are a very risky and unpredictable mix. I wouldn't engage the capability if the feature was free.

Unsupervised driving is definitely a ways away, except for maybe some limited controlled scenarios.

Supervised self driving though is safer than pure human driving. If you had a car with autonomous that you could "supervise" I'd definitely recommend it.

 

[XYang2023]

Just a personal opinion, but I think autonomous vehicles will be "next year" for more than a few years yet. I've been thinking for years that human drivers and autonomous vehicles are a very risky and unpredictable mix. I wouldn't engage the capability if the feature was free.

The same logic applies to robotics. Because Nvidia and Tesla are deliberately bringing humanoid robots into the conversation (likely to add hype and stories to their stock performance), many people are overhyping the concept.

I have discussed this with my supervisor and lab colleagues multiple times. The safety concerns associated with such robots will likely prevent their widespread adoption, limiting them to testing or research purposes.

The types of robots that see widespread adoption are typically vacuum robots or robots that are inherently safe to use.

Lastly, there is the question of whether humanoid robots are even necessary. Many problems can be addressed differently in ways that do not require such robots.

 

[XYang2023]

In China, robots like the ones found here can be purchased at very reasonable prices:

Humanoid robot G1_Humanoid Robot Functions_Humanoid Robot Price | Unitree Robotics

Unitree Robotics Humanoid Robot G1 has the flexibility beyond normal human, can imitate & reinforce the learning drive, combined with force-position hybrid control, sensitive and reliable, can simulate the human hand to achieve the precise operation of the object, welcome to consult with us to...

www.unitree.com

For researchers, these robots are quite useful. However, developing solutions and commercializing them is a different challenge altogether. Personally, I am concerned about the potential market limitations and safety issues. As a result, I prefer to focus on fields other than robotics.

 

[Xebec]

Lastly, there is the question of whether humanoid robots are even necessary. Many problems can be addressed differently in ways that do not require such robots.

I believe the thinking is that humans will treat robots better if they appear human vs something else. The complexity is definitely not needed for all tasks but is necessary if you wanted one robot to do everything a human can do.

The robots at the Cybertaxi event were apparently very articulate but remote controlled by other humans. (Perhaps there is anyway a use case for telepresence for these robots under human control in dangerous situations.)

In Teslas case they tried to fully automate the production of cars and found that humans can do a lot of things better than even dedicated "car building" robots. The Tesla bot may also be a novel approach to address that gap.

 

[XYang2023]

I believe the thinking is that humans will treat robots better if they appear human vs something else. The complexity is definitely not needed for all tasks but is necessary if you wanted one robot to do everything a human can do.

The robots at the Cybertaxi event were apparently very articulate but remote controlled by other humans. (Perhaps there is anyway a use case for telepresence for these robots under human control in dangerous situations.)

In Teslas case they tried to fully automate the production of cars and found that humans can do a lot of things better than even dedicated "car building" robots. The Tesla bot may also be a novel approach to address that gap.

Manufacturing lines are already highly efficient. Additionally, humans excel in dexterity and are remarkably efficient in tasks requiring fine motor skills. Also, we already have industrial robots for a very long time.

When it comes to safety, one major concern is falling. For example, we have two-wheeled robots, each equipped with an arm on a torso. My supervisor warned me that such robots could fall during simple tasks, like opening a fridge door. Initially, I was a bit surprised, especially since they use wheels.

For humanoid robots, integrating them into environments with civilians raises additional safety concerns. For instance, what happens if a child's hand gets stuck near a motor in a joint? These motors can be extremely powerful, and scenarios like this could multiply in real-world use.

I'm confident that someone will eventually make it work, but the question is when. We have yet to see a truly autonomous vehicle outside of controlled, geo-fenced demonstrations.

Personally, when it comes to consumer-targeted applications, I prefer solving problems that robots are designed to address but using alternative, more practical solutions.

 

[XYang2023]

The safety issues and preventive measures could undermine the efficiency of using robots in the first place.

 

[OneEng]

The last CES I attended was Asia dominated. Is that still the case? Did China companies have a big showing? Samsung is always big at CES. How about Japan?

1/2 of the business cards I got were from Chinese companies ..... most of which wanted me to move manufacturing TO China (not a chance of that btw. Too big a risk of future tariffs and already a target of 30% tariffs).

 

[tonyget]

The last CES I attended was Asia dominated. Is that still the case? Did China companies have a big showing? Samsung is always big at CES. How about Japan?

I had a look at Youtube CES related video from various tech channels，the vast majority of products in these videos are from Chinese firms

 

[tonyget]

Mainland China did not make a big showing

But it did

and the atmosphere of stagnation in the mainland tech industry is easily readable in the air, despite the effort to hush it

Quite the opposite，the tech community in China is very optimistic and confident about advancement in technology. Take a look at this video

Mainland is a technical laggard, and it only stopped looking like that from the Western point of view simply because they had some products to show at all at some point, when they had none before. But of course, it's still very big, but nevertheless a laggard keeping importing technology, rather than creating it.

Not according to any recent research paper on Chinese tech power from Western institute/think tank. I read alot related reports on this subject，and pretty much all reports rank China as the second most technical advanced country just after the US，and the speed of Chinese technical progress is unmatched by any other country in the world.

The last place where I seen Windows 98 in my life was a mainland bank office in Shenjhen

Make sense. French airport still uses Windows 3 as far as I know，large organization such as banks or airport is reluctant to change the system once it's installed

 

[Paul2]

But it did

This year it was way less than during previous few CESes, and with trade fairs on the mainland itself, the slowdown is all apparent.

Quite the opposite，the tech community in China is very optimistic and confident about advancement in technology. Take a look at this video

For me, the lack of new stuff on the shelf is the most obvious indicator of how much entrepreneurial activity is turned into widgets. Reading Chinese internets I see some O2O stuff, some gimmicks like Chinese space rocket companies, Chinese EV makers competing making supercars, companies ripping government funds for fundamental physics RnD, supercomputers, but the amount actual new stuff you see in the supermarkets is far less than a decade ago, and it is less novel, and less creative. The mainland wants to appear as it's moving "quantity to quality", but in reality it's going down in both.

 

[Daniel Nenni]

CES is fun to see products that are being tested and introduced but many of then do not make it to market. From what I could see online, this year is still dominated by automotive. I did not see a killer AI product or app yet but it is still early. Hopefully within five years of the ChatGPT introduction, which was at the end of 2022. Otherwise the AI bubble will burst sooner than later, my opinion.