Mecka AI reportedly nears funding at a $500 million valuation
The startup pays people to record everyday movements that robots can learn from. Sequoia is reportedly leading its next funding round.
The briefing
Robot training data draws more investment, Garry Tan challenges limits on learning from rival models, and a medical prototype moves towards commercial use.
Mecka AI reportedly nears funding at a $500 million valuation
Mecka AI is reportedly close to new funding at about a $500 million valuation. It collects human movement data for robot training, a business built on people demonstrating everyday tasks.
Mecka AI is reportedly nearing a funding round led by Sequoia Capital at a valuation of about $500 million. TechCrunch cites two people familiar with the deal. The terms could change; Mecka did not respond to its request for comment, and Sequoia declined to comment.
The company collects and analyses human motion data for robotics training. It pays people to record everyday tasks, such as making coffee or fixing cars, with body sensors and smartphones. Those recordings give robots examples of physical work to learn from.
Before a robot can do the job
This is a different supply problem from collecting text off the internet. A robot needs examples of how people move through a task, including what their hands do and what they see. Mecka gathers that material for humanoid robots and other physical systems. Their education still involves people doing the chores.
The reported financing comes roughly three months after Mecka’s previous raise. The amount of new money being discussed has not been disclosed, nor has the company published its customer list. A valuation is the proposed price of the business, not the size of this cheque.
Money is reaching companies that supply robot training data as well as the companies building robots. Mecka’s reported round puts a price on that interest; its undisclosed customer list leaves demand harder to assess.
Market signal
Garry Tan wants US labs free to learn from rival models
Garry Tan wants US regulators to leave room for smaller American labs to train models using outputs from larger rivals. He argues that this should be allowed through ordinary access, without fraud or stolen credentials.
Garry Tan wants US regulators to leave room for American open-weight labs to use model distillation. The technique trains one model using outputs from another. Tan argues that smaller American labs should be able to use this approach to give the US more open-weight options beyond those developed in China.
The Y Combinator chief’s argument, reported by TechCrunch, comes as Anthropic seeks tougher action against unauthorised distillation. Anthropic has accused Chinese labs of obtaining access through concealed identities, fraud and stolen credentials. Tan distinguishes that conduct from domestic labs using ordinary access routes.
Who sets the limits?
Tan sees it as an overreach for AI labs to dictate what customers may do with the information their models provide. He argues that models trained on broadly available human knowledge should offer more freedom of use, rather than keeping that knowledge behind restrictive terms of service.
He also wants frontier labs to remain viable businesses that can fund further work. His concern is that one proprietary company could gain such an advantage in capital and researchers that it dominates the field. More open-weight alternatives are his proposed counterweight. This is his argument for regulators to consider, not an announced rule change.
Tan wants frontier research to remain fundable while smaller American labs get more freedom to build open-weight models. His proposal puts restrictions on API use at the centre of that argument, alongside the conduct used to gain access.
What to watch
An AI-guided device for medics moves towards commercial use
A device designed to help medics access blood vessels is moving from a research prototype into a company. The transfer has won a federal technology award.
The team behind AI-GUIDE is transferring its vascular-access prototype to AutonomUS Medical Technologies, a startup founded by staff from MIT Lincoln Laboratory and Massachusetts General Hospital. The effort has received the Federal Laboratory Consortium’s 2026 Excellence in Technology Transfer Award.
AI-GUIDE combines custom AI software with handheld ultrasound equipment to help a user place a guidewire and catheter in a blood vessel. Its developers designed it for care outside a hospital, including military medics working with limited specialist support.
A company to carry it forward
MIT’s account traces the project through prototype development and clinical testing at Massachusetts General Hospital. AutonomUS is now responsible for moving the technology towards commercial use. That gives the research a company to handle the next stages, rather than leaving it as a laboratory project.
The startup has secured FDA Breakthrough Device Designation. It has also received an Air Force small-business research grant, alongside private investment and support from other US agencies. The team is developing related technology for peripheral nerve blocks, intended for trauma care and pain management.
For medics and hospitals following AI-GUIDE, the useful news is the transfer into a business that can develop it further. The award recognises that transfer work; it does not announce a commercial launch.
What to watch next
- Money is reaching companies that supply robot training data as well as the companies building robots. Mecka’s reported round puts a price on that interest; its undisclosed customer list leaves demand harder to assess.
- Tan wants frontier research to remain fundable while smaller American labs get more freedom to build open-weight models. His proposal puts restrictions on API use at the centre of that argument, alongside the conduct used to gain access.
- For medics and hospitals following AI-GUIDE, the useful news is the transfer into a business that can develop it further. The award recognises that transfer work; it does not announce a commercial launch.
The takeaway
Money is reaching companies that supply robot training data as well as the companies building robots. Mecka’s reported round puts a price on that interest; its undisclosed customer list leaves demand harder to assess.
The editor’s view
Tan wants frontier research to remain fundable while smaller American labs get more freedom to build open-weight models. His proposal puts restrictions on API use at the centre of that argument, alongside the conduct used to gain access.