 Hello, and welcome to this Analyst Angle, CES Rapid Edition on theCUBE. I'm Rob Streche, analyst with theCUBE Research. And today I'm gonna provide you some analysis on the top trends that I saw at this year's conference. 135,000 people attended, not quite back to pre-pandemic, but it definitely felt that way. It was really well attended, a lot of enthusiasm, and not just from the people you would expect to be there. There was a broader group of people attending there, and I think that had to do a lot with these top trends that we're seeing there. And those top trends were AI is eating everything. It is in everything, it is all about AI. Everything was AI or labeled AI there. I'm gonna kind of break that down a little bit and kind of give you my view on what was going on a little bit below the covers at some of these companies that we're talking about AI. Sustainability is a must in your marketing, especially if you're going outside the US. It was really clear that people were putting a lot of wood behind the sustainability arrow for lack of a better pun. And you start to look at how they were actually talking about how they were doing circular economy, how they were looking at the power they were using to be clean and renewable, how they were looking at minimizing components. So all of this went into a lot of sustainability messaging that was out there around not only in the B to C market, but the B to B market as well. So again, very key, especially it was seen if you were coming from one of the other countries not named the United States. And then enterprise tech marketing really was being used in heavily and non-traditional B to B campaigns for B to C type products. This will unpack a little bit as well. The number one trend was AI and it was everywhere but how much AI and what type of AI really varied greatly. On the consumer side, I have some very interesting discussions that you had to see what was meant by their AI in their products. Because again, it ranged greatly from actual consumer products containing general models from interaction to the use of chat GPT via API to provide a little summarization and some insights into data that they're already collecting. To me, this was super interesting. One of the very interesting companies that I had to talk to in this space was Blovo. Blovo Animal Health and Blovo Tales is a chat GPT for animal health. It was to help make it easier for pet owners to access veterinary care services and maintain their pet's health and health data. Super interesting and especially if you're a pet owner, you understand the great cost and expense you go through for these pets. So how can you lower that? Because if you're not using pet insurance and there's something major that comes up, it can be a really huge type of incident and cost. So this then lowers the cost of pet caring, gives greater control over the animal's data to those individuals. What they're doing is basically bringing summarization using chat GPT underneath the hood to actually make it accessible so you understand what is meant by different types of tests that are going on. It gives you a summarization. It allows you to then go and get better pet care for that. Similar applications for humans, also using private models, not necessarily chat GPT, were also on display there. Healthcare AI was a major theme for the large vendors and small vendors. And it took a lot of different, I guess you could say variations of what healthcare looked like. One of the ones that I talked to Darucci, who actually won two major awards at CES for their T11 Pro smart mattress, which uses its own general AI model to look for health anomalies based on the data from the sensors in the mattress and what it knows about you. Not only helping you with your snoring by helping you move around on there or giving you a better night's sleep by moving different air sensors and airbags within the mattress, but looking at your overall health indicators and helping you get and reach a deeper sleep and also looking at what it meant to be healthy. So looking at temperature and other types of sensors using these ceramic sensors. Truly AI, truly doing inference at the edge to achieve these results is where they were focused. This way they would use a generalized model, push it down once it's already been trained and then it's fine-tuned and actually the results are fine-tuning based on your body and how you use that mattress. So again, a really good idea about where all of this is going, the data goes back up to their cloud, it's pushed down for some results, that results and tuning that are constantly tuning that model and then the model is deployed at the mattress. Again, very unique in how they actually keep the data, keep it safe, safety was definitely and security was on the minds of all of these vendors. Other AI themes included automotive use cases. I had a discussion with Sandeep Rajani, CEO from Sir Brexium in the Blackberry booth about how AI needs to be more at the car level. And I think this is really interesting. Again, it goes back to these edge use cases for AI when you're doing inference and you're looking at how do I get the data closer or bring the AI to the data? And that training and tuning will be done in the cloud. So again, we started to hear these messages, maybe they're not coming up with their own foundation models but they're using general models and then fine-tuning based on particular data so again, it wasn't all about, hey, I'm going to build the biggest LLM or I'm going to build the biggest foundational model. I'm looking at different models to build that, maybe industry specific models. For this case, many of the use cases such as risk determination in the insurance industry will be able to leverage this inference and data summarization at the edge then act upon it up in the cloud so they can actually predict what is the likelihood of accidents in a particular car driven at a particular rate of speed on a particular highway in a particular location. All of this giving greater information about what is the risk tolerance that maybe an insurance company. These used to be done in all kinds of actuarial tables and still are to a certain extent but this modeling is really unique in the fact that you're getting this edge related data. This highlighted a theme of bringing together models for time series and large language models. This way you could actually understand what was going on over time because time series analysis is something that's been done in ML and AI for quite some time but when you start to bring that together with large language models so you can look at it specific time periods and what were the different parameters going on and ask it in a natural language way. This is really new in how LLMs and SLMs or segmented or small language models are being deployed to help people really get a better understanding. So a lot of what was going on at CES was not just about the AI that was on display in the items such as the mattress or the healthcare app for the pets but it was really what was behind the scenes. That brings us to some of the discussions I had with some of the larger folks there. IBM and AWS and how they were applying AI at the edge for manufacturing in a disconnected and connected manner. Now you may say AWS, IBM, disconnected doesn't factor in with cloud. Well, I think this is the real key is that people understand that if you're going to have safety, security, and be able to address things such as manufacturing, you really need to be there. With IBM we discussed AI in manufacturing, supply chain and what was the sustainability aspects of AI. Talking to Jose Favolivla about how and where models live, the importance of governance and security and the need to bring AI to the data. Again, one of those themes that kept ringing out. Other interesting points that we talked about was the commonality between industries and that foundational models don't need to be different but data will vary highly amongst those. So how you train them, those foundation models and how you then do fine tuning and the fine tuning even goes another step where it may be company specific. That's really the key differentiator in what you're gonna be using these SLMs and LLMs for as well as modeling in general for industrial devices. Another highlight that we believe strongly in is the need for AI to be brought to the data. I can't say that enough rather than the other way around in order to optimize industrial processes and prevent do preventative maintenance types of things. So you wanna understand is the possibility of this machine going to break down? Well, actually you need to know at seconds because when the factory is down, that is costing you money. This is a great place that AI has been actually pursuing for years but now with the advances and the technology that has come out with some of these other models you're able to get at the data and query the data a lot faster by again bringing time series analysis together with LLMs. Also, in a briefing with AWS we talked about how manufacturing is one of the spaces where AWS and the cloud operating model truly meets some partner hardware where users download and install on third party industrial computers or a gateway with Linux or Windows operating systems or on AWS specific hardware such as Outposts or Snow Family. Again, this is where actually AWS code is running at the factory. That was super interesting and when you start to look at hybrid and how AWS is approaching it really this team is on the cutting edge of that. And what we did is we deep dived into AWS IoT site-wise Edge which is the product that you can download and configure and put on those third party hardware devices so that you can get those data. In fact, they have a number of different partners that they have just announced back at re-invent back in early December. Process is streamlined through the AWS IoT site-wise console so again, you can use that and do the deployments and configuration through the console where users can then select data sources like sensors or industrial equipment and apply predefined or custom functions for data sampling and metric computation. Why is that interesting? Well, we also discussed the announcements that were made at re-invent and how AWS is leveraging ML at the edge so with site-wise Edge and how they're also combining that once again with LLMs in Amazon Bedrock for manufacturing. So this is also again helping people get at the data and understand what are the different chances of something happening could be there. I didn't only talk to the big people, I also took a swing through Eureka Park. Eureka Park, if you haven't been to CES is really where a lot of the startups are and it's not just startups from the US, it's startups from around the world. You have different countries actually representing the startups. One of the ones that I visited that was super interesting was Swiss Tech in Eureka Park where all of the startups were located from around the world, including Switzerland. Meaning with Algorized CEO Natalia Laparov for a demonstration of how they're using AI for remote sensing. This was super interesting that the software that they are producing goes beyond traditional perception limits. Things that you can see or hear and understand in space. The offering really complete understanding of how objects and individuals in any setting are really what's going on with them. The demonstration they had was of in vehicle where you could tell the sensors respiration could and the sensors could tell the respiration of what the individual was doing there. Not only could they tell the respiration, how many beats from your heart and how much you were breathing, but there was the difference between children and adults could automatically be seen. So you could think of this as an example where, hey, there's an accident and I wanna understand who's in the car and the telemetry of what's going on. I now know how many people are in the car. What is the rate of breathing? Are there adults and children? It can even tell if there's a pet in the car. This kind of information is a demonstration of bringing AI to the edge. So another demo they showed, you could see this applied to rooms in a building, which is very helpful to super secure environments such as skiffs. They noted that the software runs on ultra wide band chips showing the seamless integration with Corvo's ultra wide band radar chip for enhanced accuracy and reliability in the automotive experience. So again, looking at normal chips that are out there, in fact, we talked about how they are actually in your cell phone today could be using this technology to then go out and understand the environment around you. The applications for this beyond security, beyond safety could be for helping people with various different limitations of sight or hearing or what have you. So it was very interesting to see this and see this in action. Another startup in Swiss tech in the Eureka Park area was StackSync, meeting with Rubin Burden, the CEO and Alexi Farvray, the CTO. I got to see a demonstration of how they were helping bring bi-directional syncing between CRM systems like Salesforce, Microsoft, Dynamics and HubSpot, to databases such as Snowflake, BigQuery and Mongo. Again, bringing the data bi-directionally between the two, but why are they doing this? Well, by doing this two-way sync, StackSync allows DBAs to correct things in the database using SQL and have it show up in CRM systems. One of the things that I thought about was, hey, it would great to have a centralized place to have all the SKUs for a particular set of products that I had at one of the companies that I was building product for. You could go and do that in the database, update it in the database and see it sync up to that, say Salesforce, and that way when people go to actually build out their order forms or go and make offers to customers, it's always consistent. Same time, they can take data back into the database about average deal size or deal discounts. So many other applications for this where you can have that single source of truth be that data warehouse, data lake, and be able to access that through SQL using normal tools, not having to have specialized CRM tools or have multiple versions floating around. Another startup in Eureka Park was over in the European Innovation Council area. That was personally fascinating to me was this company called Dot Lumen who was taking AI for self-driving cars and built a lightweight headset using heptapic responses to guide low vision or blind folks while they were walking through the streets or in this case around a busy CES floor. This to me was one of the best uses of AI for sensing at all from a mobility perspective because it really helped from an accessibility perspective that was only been available for blind or low vision by using guide dogs. And guide dogs do have an expense to them. There's some places they can't go per se because of cost and other things. So being able to bring something that has only really been accessible to people who could afford guide dogs was really interesting. Now we're gonna shift a little bit to sustainability. It was pretty clear that you can not sell in the rest of the world outside the US without having a sustainability message. And in fact, many of the larger vendors were in there talking about sustainability and had different demonstrations of how they were actually contributing back. But not only that, the sustainability on the tech side, there were also a number of interesting advancements on the product side that had come out over the last several years and are really finally starting to be put into action. Mobile off the grid power in solar and wind was big at the show, along with backup power for the home, battery versus generator. There was a whole lot of electric or battery generators being shown off there. As we move towards these new grids, these definitely could be more accessible to people in different areas where they may have only had the option for generator in the past. Also a number of advancements in home appliances from the efficiency standpoint really stood out to me. One was that GE profile combo washer dryer that used breakthrough heat pump technology allowing it to be ventless. The biggest advancement was that you could actually wash and dry a king size comforter in under three hours. Why is this huge? Well, these have been out for quite some time and it used to take you normal time to go and wash the thing like an hour to wash it, a king size comforter, but then it may take you five to six hours, which is not really energy efficient. This working off a 110 volt plug here in the States that would be and being able to accomplish this and the fact that it's ventless. So if you look at cities where maybe construction codes and things of that nature don't allow you to vent out or you really don't want that heat escaping into the atmosphere and other things that are in that using this heat pump technology and keeping it self-contained was truly amazing and watching it actually perform was even more amazing. The last trend that was on display was the tech marketing coming to other industries. As a person who's been in marketing and in product for quite some time, this was one of the most noticeable, I guess you could say trends was that a lot of the messaging that actually tech has used for probably about the last 15 years is now coming to more things. One of the biggest that was at the show and most noticeable was software defined and in particular, the software defined automobile. This was everywhere and it was on full display in Blackberry's booth among others and beyond this and beyond the words of AI being splashed everywhere, software defined and that concept was being brought to every part of the tech ecosystem. But like adopting any set of words, to me, this rang a little hollow in some places and you had to dig deeper to understand what did they mean by software defined, especially where these were hardware companies trying to sell that. In the auto industry, it was pretty clear. You're going to almost fly by wire or drive by wire in a lot of cases. You also were able to have different personalities of these cars and again, like we talked about earlier, all of the different sensors that are gonna happen and all of the different AI that's gonna be enabled to help you in a crash avoidance and other types of situations. It really is software defined and many cars are going this direction. Many of these industrial and consumer tech companies are bound to learn the lessons we've learned over the last few years, bringing marketing from one part of the industry to another is not always the easiest. So that's a wrap and I wanna thank you for watching this analyst angle and wrap up of CES. I'm your analyst and host Rob Streche. You're watching theCUBE, your leader in enterprise technology, news and analysis.