Instead of more cloud, more scale, and more infrastructure, he pointed to something far simpler.
“We will probably see a large transference of where this agentic processing is actually happening, less on the cloud, more on the device,” he said.
That idea sits at the core of what HP is trying to build with HP IQ.
AI today is getting expensive at every level. As models grow larger, they demand more compute, more data center capacity, and more energy. For enterprises, this is quickly becoming a cost problem, not just a technical one. The promise of AI is efficiency, but the reality is that running it at scale is anything but cheap.
Chaudhri believes that the equation needs to shift, and not gradually.
“It’s not really just about offline usage or privacy… the cost is also painful for businesses,” he said, framing on-device AI as not just a feature, but a necessity.HP IQ is the company’s attempt to act on that belief. Positioned as an experience layer across devices, it is designed to bring AI closer to where the work actually happens. Instead of relying entirely on cloud-based models, HP is focusing on running optimized AI locally, within the limits of real-world hardware.
That shift also reflects Chaudhri’s own journey. Before HP, he was one of the key minds behind the AI Pin at Humane, a device that tried to rethink computing altogether. It was an ambitious bet on new form factors, built around the idea that AI could replace the way we interact with screens.
HP IQ is designed to bring AI closer to where the work actually happens.
When I brought that up, and asked whether we’ll see more such radical devices, his response suggested that the vision hasn’t changed, but the path has.
“Today’s an initial foundation,” he said. “We’re building the experience layer first… and then we bring new kinds of hardware into that.”
In other words, instead of starting with a new device and figuring out the ecosystem later, HP is doing the reverse. HP IQ is meant to unify devices first, creating a consistent layer across laptops, workstations, and eventually, even non-HP hardware.
That foundation rests on two key pieces. Devices need to be connected, and they need to share a consistent experience.
“The first step is that we need to have our devices talking to one another… then we need to have a consistent experience layer, which is what you saw through HP IQ,” Chaudhri explained.
Only once that is in place do the more interesting possibilities open up, workflows that move across devices, AI that understands context across screens, and eventually, new form factors that plug into the same system.
This is also why HP is looking beyond its own ecosystem. Through partnerships with Google and Qualcomm, the company is working to extend these experiences into the broader Android world.
“We do think there needs to be an evolution of device-to-device interoperability,” he said, acknowledging that users don’t operate within a single brand’s hardware anymore.
That interoperability becomes even more important in a world where AI itself is distributed. HP’s broader vision hints at a future where compute is not just local to a single device, but shared across nearby devices as well.
“You can imagine that compute being able to be shared within on-prem devices,” Chaudhri noted, pointing toward a model where intelligence is not centralized, but spread across the devices around you.
This doesn’t mean the cloud disappears. Instead, HP is betting on a hybrid approach, where some workloads remain in data centers, but a growing share moves closer to the user.
That philosophy also shapes how HP is thinking about models themselves. Rather than chasing ever-larger parameter counts, the focus is on efficiency.
“We do think that 20 billion parameters is a good ceiling for us at the moment,” he said, adding that advances in compression and optimization could allow smaller models to perform at levels previously associated with much larger ones.
Taken together, this paints a different picture of where AI is headed.
Not one defined purely by scale, but by distribution. Not one where everything flows through massive data centers, but one where intelligence exists across devices, closer to the user.
Coming from someone who helped build one of the most ambitious AI hardware experiments in recent years, that shift in thinking stands out.
Because if the AI Pin was about reimagining the device, HP IQ is about reimagining the system around it.
And if Chaudhri is right, the next phase of AI won’t just be about how powerful it gets.
It will be about where it actually lives.
as a Reliable and Trusted News Source
