What’s ahead for tech in 2025, according to an IBM executive

By temporary On Dec 16, 2024 5 0

Since the launch of OpenAI’s ChatGPT two years ago, the tech industry has been consumed by generative artificial intelligence — and the next year will see more businesses adopt the technology, one tech executive predicts.

Enterprises are already looking for technology to scale and implement AI, and the focus in 2025 will be on more efficiency and less risk, Hillery Hunter, chief technology officer and general manager of innovation at IBM (IBM) Infrastructure, told Quartz.

With cost efficiency in mind, enterprises will likely look for smaller, more targeted generative AI models, Hunter said. Businesses could also turn to ensembles — or combining multiple machine learning or deep learning models for more accurate and robust results — that will allow them “to optimize latency, efficiency and accuracy across their AI workloads,” Hunter said.

“For example,” Hunter said, “enterprises would have the option to use smaller, faster and more energy-efficient models on a CPU [core processing unit] as a starting point for data analysis and processing and then use larger models on a GPU [graphics processing unit], as needed, for more complex workloads that require extra performance, all within the same AI Workflow.”

In July, Nvidia (NVDA) chief executive Jensen Huang said the next wave of AI is for businesses, and that everyone at a company “will have an AI assistant.”

“We hope that we can give every single organization the ability to create their own AIs,” Huang said during a fireside chat at SIGGRAPH, a computer graphics conference. “So everybody would be augmented and have a collaborative AI that could empower them, help them do better work.”

As the tech industry focuses on launching new AI products, it’s now AI agents, or software that can complete complex tasks autonomously, that have taken over the chatbot hype. Most AI-powered tools that have launched so far are co-pilots that can help users work more efficiently, but have to be prompted over and over.

In the next year, companies will risk being left behind if they don’t step up their AI efforts, Hunter said. While businesses saw AI become more important this year, they “often struggled with disjointed AI process steps and driving true ROI [return on investment].”

Businesses have to bring AI projects together in 2025, Hunter said, or otherwise create “an increasingly bloated tech environment” with more costs and slower innovation.

“It’s time for enterprises to kill the Frankencloud — a disjointed environment where data is duplicated and sprawled out across various, disconnected locations — and create the holistic view they need to actually leverage and govern the data and models associated with AI, and generate outcomes that support their AI investment,” Hunter said.

Next year will also see enterprises commit to getting out of “tech debt” — debt from failing to modernize IT infrastructure or not having unified IT or cloud-based solutions that are needed to optimize AI.

“AI itself will help modernize IT, but in the year ahead, collaboration across the C-suite is going to be key to getting out of this debt,” Hunter said.