• Nvidia’s CEO Jensen Huang delivered the opening keynote at CES 2025 in Las Vegas
• He expects agentic AI to be a big deal from hereon in and described it as a "multitrillion-dollar opportunity"
• Huang also unveiled new desktop AI supercomputer, autonomous vehicle, and “physical AI” products and collaborations  

Like it or not, ‘agentic AI’ looks set to be the tech buzz phrase for 2025, following the declaration by Jensen Huang, the CEO of AI tech giant Nvidia, during his 90-minutes-long keynote speech that opened the CES 2025 show in Las Vegas that “the age of AI agentics is here”.  

He declared that agentic AI is a “multitrillion-dollar opportunity”, and a development that will change the way in which people work. “AI agents are the new digital workforce… The IT department of every company is going to be the HR department of AI agents in the future,” he added, a statement that will no doubt displease all IT departments. “Today they manage and maintain a bunch of software from the IT industry – in the future they will maintain, nurture, onboard and improve a whole bunch of digital agents and provision them to their companies to use… your IT department is going to become like your AI agent HR,” said Huang.   

The ‘agentic AI’ term is not new – it started to become popular in AI circles during the second half of last year, was highlighted by IBM as the hot new trend in AI in October, and was tipped as a key area of upcoming focus during TelecomTV’s Telcos & AI event in London a month ago (particularly during the Building the telco AI ecosystem session) – but Huang’s keynote reference is sure to make it part of common parlance from hereon in. 

And what is agentic AI? It refers to an AI tool that is capable of “autonomously performing tasks on behalf of a user or another system by designing its workflow and using available tools,” according to IBM. “The system has ‘agency’ to make decisions, take actions, solve complex problems and interact with external environments beyond the data upon which the system’s machine-learning models were trained. AI agents draw not only from databases and networks but can also learn from user behaviour, improving over time,” making them more adaptable and focused than AI tools confined by large language models (LLMs) or small language models (SLMs). “Agents’ adaptability enables them to handle complex, multistep AI applications that traditional AI would not be able to handle, making them a key part of the modern organisation’s process automation strategy,” according to IBM. 

This is an important trend because it’s likely to become a big talking point this year in industry vertical-specific AI developments, including those focused on the telecom sector – it’s hard to imagine that agentic AI won’t be on the (unofficial) agenda at MWC25 in early 2025, for example. 

And it’s certainly on the agenda at CES, thanks to Huang, who has taken his keynote opportunity to pitch the importance of agentic AI (alongside many other things, it must be said), mainly because Nvidia has some new technology to pitch. “To provide a foundation for enterprise agentic AI, Nvidia today announced the Llama Nemotron family of open large language models (LLMs). Built with Llama, the models can help developers create and deploy AI agents across a range of applications – including customer support, fraud detection, and product supply chain and inventory management optimisation,” noted the tech giant in this press release (one of many issued by Nvidia as its CEO addressed the CES crowd). 

In addition, Nvidia unveiled AI blueprints that can be used to build agentic AI applications and a range of partners that have helped to develop these blueprints. “With the blueprints, developers can now build and deploy custom AI agents. These AI agents act like ‘knowledge robots’ that can reason, plan and take action to quickly analyse large quantities of data, summarise and distill real-time insights from video, PDF and other images,” noted Nvidia in this press release.

And to help enterprises make use of AI agents as quickly as possible, Accenture has developed an AI Refinery for Industry, built with Nvidia AI Enterprise tools (including NeMo, NIM microservices and AI Blueprints), to “help enterprises rapidly launch agentic AI across fields like automotive, technology, manufacturing, consumer goods and more.” The first 12 agent solutions from Accenture’s refinery will be available by the end of February and should number 100 by the end of 2025. 

And there’s more… much more.

Agentic AI was just one of the topics on Huang’s keynote agenda, as he also:

• Unveiled the Nvidia Cosmos platform, which “advances physical AI with new models and video data processing pipelines for robots, autonomous vehicles and vision AI” – “The ChatGPT moment for general robotics is just around the corner,” stated the CEO.
• Announced the new Blackwell-based GeForce RTX 50 Series GPUs (graphics processing units) that boast enhanced “visual realism” and “unprecedented performance boosts”.
• Launched Project DIGITS (deep-learning GPU intelligence training system), which “brings the power of Nvidia Grace Blackwell to developer desktops in a compact package that can practically fit in a pocket”. Huang stated: “Every software engineer, every engineer, every creative artist – everybody who uses computers today as a tool – will need an AI supercomputer,” adding, “This is Nvidia’s latest [and smallest] AI supercomputer… It runs the entire Nvidia AI stack.” DIGITS features the new GB10 Grace Blackwell Superchip, which was developed in collaboration with MediaTek. 
• Announced a partnership with Toyota for “safe next-gen vehicle development using the Nvidia Drive AGX in-vehicle computer running Nvidia DriveOS” and noted that Aurora and Continental have also joined the growing list of vehicle companies “developing and building their consumer and commercial vehicle fleets on Nvidia accelerated computing and AI.”
• And launched the company’s Drive Hyperion AV platform, which is “built on the new Nvidia AGX Thor system-on-a-chip (SoC), designed for generative AI models and delivering advanced functional safety and autonomous driving capabilities,” noted Nvidia. “The autonomous vehicle revolution is here,” Huang stated. Hyperion, which Nvidia claims is the “first end-to-end AV [autonomous vehicle] platform,” comprises advanced SoC products, sensors, and safety systems for next-gen vehicles, and is being adopted by the likes of Mercedes-Benz, Jaguar Land Rover (JLR) and Volvo Cars.

AI has been “advancing at an incredible pace,” noted Huang, as has Nvidia’s share price and valuation. The company’s stock was rising ahead of Huang’s keynote speech in anticipation of new AI developments, ending Monday at $149.93, up more than 3.4% for the day and up by 186% compared with a year ago. Nvidia ended trading on Monday with a valuation of $3.66tn (that is not a typo) and early indications are that the stock is set to gain further during Tuesday trading. 

- Ray Le Maistre, Editorial Director, TelecomTV