We stand at a crossroads defined by a critical problem: the centralisation of artificial intelligence. In today’s dominant model, data and processing are concentrated in remote data centres, creating systems users cannot meaningfully audit or control. This architecture enables convenience at scale, but it also facilitates a subtle transfer of power and privacy. “We have conflated intelligence with centralisation,” observes Lado Okhotnikov, founder of the biotech platform Holiverse. “We built these vast, brilliant minds and asked them to solve our problems. But in doing so, we outsourced our sovereignty.”
The core problems of centralised AI
From Holiverse’s perspective, the centralised AI model presents several structural issues:
- Concentration of power and erosion of privacy
User data becomes a commodity, while control over AI logic and outputs rests with platform owners. - Generic, non-personalised intelligence
AI trained on aggregate data is inherently limited in its ability to understand individual nuance and context. - Structural inequality and fragility
Access to advanced intelligence is tethered to reliable internet connectivity and platform access, excluding many and creating systemic vulnerabilities.
The decentralised alternative
According to Lado Okhotnikov, decentralised AI shifts the paradigm by bringing processing directly to the user’s device. “The future is about intelligences that are personal, private, and profoundly your own,” he states. The advantages of this approach are clear:
- Data sovereignty and privacy
Personal data remains on the device in a closed loop, eliminating exposure to external servers. - Resilient and universal access
Intelligence functions offline, making it accessible regardless of connectivity and resilient to network failures. - Hyper-personalised utility
Models can be finely tuned to individual biological, behavioural, and contextual data. - Aligned incentives
The AI’s loyalty is to the user, not to a platform’s commercial or data-harvesting objectives.
Building private AI: Holiverse’s decentralised approach
This approach is already being implemented. The Holiverse team has begun active development in the field of decentralised AI. “This is undoubtedly a complex and resource-intensive process, and we are engaging some of the leading AI specialists,” says Lado Okhotnikov. “In the near future, we will be ready to present our developments to the world.”
Holiverse’s goal is to create a unique and affordable device with embedded, private AI that operates offline and does not rely on cloud-based storage. “This technology has the potential to significantly reduce the risks associated with AI and make interaction with it truly personal,” Okhotnikov adds.
A leadership choice about the future of intelligence
For Holiverse, decentralisation is both a technical and philosophical imperative. “We are deciding what role intelligence will play in the human story,” Lado Okhotnikov reflects. “Will it be a monitor and a manipulator, or a partner and a protector?” Enabled by advances in on-device processing and federated learning, this approach underpins Holiverse’s work on DNA-based digital twins — where AI analysing a person’s genetic blueprint functions as private counsel on the device, not as a shared service.
“The defining question is simple,” Lado Okhotnikov concludes. “Who does your intelligence serve? If the answer is not ‘you,’ then we have built it wrong. The goal is to ensure that the intelligence which understands your life resides where it should — with you.”
