Research
Diverging Approaches in Physical AI Development
Siftedhigh confidence
Why it matters
- →Physical AI development is diverging into data-driven and architecture-first approaches.
- →Architecture-first models adapt better to real-world conditions, offering commercial advantages.
- →The approach impacts how quickly and effectively AI can be deployed in physical environments.
The development of physical AI is seeing two main strategies: data-driven and architecture-first. The data-driven approach relies on large datasets to train models, similar to methods used in language and vision AI, but faces challenges in adapting to real-world complexities. Meanwhile, the architecture-first approach, rooted in field robotics, designs models to handle real-world unpredictability from the start. This approach, while initially less reliant on data, can lead to more effective real-world deployments and richer operational data, suggesting a potentially more viable commercial path.
Read originalThe story around this
Earlier coverage that leads up to this article, and what followed. Lines connect each piece to the closest one after it, converging here.
More from Sifted
Investment · £540m
Market & Regulationbusiness
Fractile raises £540m Series C
Fractile has secured a massive £540 million Series C, signaling intense capital competition for the UK’s leading AI chip designer. This valuation underscores the strategic pivot toward domestic semiconductor sovereignty in Europe, moving beyond software to hard infrastructure. The funding validates Fractile’s approach to custom silicon for large language models, positioning it as a critical alternative to NVIDIA’s dominance in the European market. Investors are betting heavily on the premise that sovereign AI requires sovereign hardware.
Sifted
Investment · €7.1m
