- Experts predicted AI will see a wave of new co-design, right-sizing and business model changes in 2026
- Security will also be a key theme for AI in 2026
- We may even see the rise of analog compute for AI before the year is out
AI developments have come at a blistering pace in 2025. It’s hard to imagine what we might see from AI in the new year, but we asked a range of industry executives to do just that.
The first prediction? Hardware and software roadmaps will get a bit more cozy.
Kira Boyko, Intel Product Line Director for E-Core Xeon Products, told Fierce she expects the need for AI hardware and software co-design will increase in 2026. “The independent evolution of software and hardware is breaking down under the memory-intensive and highly parallel demands of AI workloads,” she explained. “We’ll see an increase in efficiency optimization co-development at every layer of the stack.”
Meanwhile, Arrcus CEO Shekar Ayyar said he expects to see more of a focus around right-sizing AI compute infrastructure and spending. Basically, exactly what happened with the cloud, but now for GPU capacity. “Maximum efficiency” will become the name of the game, he said.
On a related note, WEKA Chief AI Officer Val Bercovici predicted that toward the end of 2026, AI agent pricing could begin to shift from being token-based to outcome-based.
“The cost of inference at scale will create pricing ripple effects throughout next year, catalyzing the rise of new pricing models,” he said.
But what exactly does outcome-based mean? In a nutshell, it means businesses will measure AI value “by results, such as completed software projects or investment returns, rather than units consumed,” he said. “This transition will reshape business models across the AI industry, though it remains risky for providers in the near term.”
On the security front, AI is changing the game in all sorts of new ways. AI agents are creating a broader attack surface for hackers. But that’s not all. Blackberry Secure Communications VP and Chief Security Advisor Christine Gadsby said in the age of deepfakes, identity verification will become a mandatory layer of communication security.
“Proving who you are—cryptographically and continuously—will become as fundamental as encryption itself,” she said.
F5 CTO for Government Solutions Bill Church added that AI governance will be critical to ensuring data sovereignty. “When AI makes or influences a decision that impacts a citizen, there needs to be full traceability - from the model’s provenance to every prompt and output,” he said. “True sovereignty means knowing not just where your data resides, but who holds the keys to it.”
Perhaps the most interesting prediction came from semiconductor company Analog Devices’ VP of Emergent AI Max Versace, who forecast the rise of analog AI compute.
And just what the heck is that? “Analog AI compute uses the physics of the sensing and computing substrate to perform computation, transforming energy directly into AI inference,” he explained.
“Historically sidelined due to scalability and precision limitations, analog compute is reemerging in 2026 as digital architectures face energy, latency, and memory bottlenecks with no solution in sight. This is especially critical in edge environments where real-time responsiveness and power efficiency are a must.”
Versace said he expects to see initial deployments by the end of 2026.