TacnIQ.ai has raised $1.5 million from In Group Holdings as part of a $3 million pre-seed round, money the company will use to build foundation models for touch — the sense that most of the current AI boom has skipped.

The gap the company is aiming at

Vision and language models have gotten very good. Machines that must make physical contact with the world haven't benefited nearly as much. A robot arm can identify a box on a shelf and still fail to pick it up, because knowing what something looks like says little about how much it weighs, how it deforms under pressure, or when a grip is about to slip.

TacnIQ.ai's argument is that this is a data problem before it is a model problem. Tactile datasets are scarce, fragmented, and usually collected for one narrow application at a time, which means every new use case starts close to zero.

The dataset

The company says it has accumulated more than 5,000 hours of tactile interaction data from experiments and field deployments — what it describes as one of the larger tactile datasets currently available. The collection method is the clever part: sensor-based wearable nodes capture physical interactions as they happen in real work environments, not in a lab rig.

The strategic bet is on breadth. Rather than training a model per application, TacnIQ.ai is training across industries on multimodal data, on the theory that the underlying physics of contact generalizes even when the tasks do not.

Already shipping

Unusually for a pre-seed company in physical AI, the technology is already deployed with paying customers across logistics, construction, e-commerce, hospitality and healthcare. The current commercial hook is not autonomous robots — it is workplace safety and ergonomics, using the same contact data to flag movements and loads that injure workers.

That is a sensible wedge. Safety and workers' compensation costs are line items every operations director already budgets for, making it a far easier sale than a speculative automation roadmap.

Team and partners

Co-founder and CEO Aashish Mehta is running the company across Singapore and California. "Our mission is to build AI models that deliver real-world impact today and transform how engineers design and develop hardware," Mehta said.

The company is also working with Synaptics, the chip designer whose touch and sensing silicon shows up across consumer and industrial devices — a useful relationship for anyone trying to get tactile sensing into hardware at volume rather than as a bolt-on.

What the money buys

Funds will expand the engineering team, advance the tactile foundation models, and scale deployments into additional industrial applications.

The read

Physical AI has become a crowded pitch category, and most of it is still demo-stage. TacnIQ.ai's position is more defensible than most: a proprietary dataset that is genuinely hard to replicate, revenue from a use case that does not require the robotics market to mature first, and a chip partner. At $3 million, it is also cheap enough to find out. The risk is the one every foundation-model startup faces — whether a general tactile model actually outperforms purpose-built sensing in the specific jobs customers are paying for.

Source

TNGlobal — "Singapore's TacnIQ.ai raises $1.5M from In Group Holdings in pre-seed round"

Click here to start customizing