Pronto may no longer be just another fast-growing instant home-services platform.
Behind the startup’s rapid expansion and growing investor interest lies a much larger and potentially controversial ambition: building a real-world behavioural data layer for Physical AI and robotics systems.
The strongest indication of this shift comes not from public marketing material, but from internal investor documents reviewed by Entrackr.
According to a memo from Glade Brook Capital, Pronto is “seeking to formalize India’s vast informal labor markets and in the process generate data to help train physical AI and robotics.”
The memo further states that the company is already “piloting real world training data with leading physical AI labs.”
That single line changes how Pronto may need to be viewed.
Until now, the company was largely seen as another instant services startup helping households access workers for tasks such as cleaning, utensil washing, laundry, cooking assistance, gardening and car washing.
Founded in 2025 by Anjali Sardana, the company has rapidly expanded across major Indian cities and attracted strong investor attention.
But the investor memo suggests Pronto’s ambitions may extend far beyond convenience services.
From Home Services To AI Infrastructure?
According to the investor note, Pronto is “developing a data business leveraging its workforce to capture real-world household data for robotics labs.”
The document also claims that early partnership discussions have been “encouraging” and that the company is “moving quickly to commercialize the strategy.”
That points toward a very different kind of business model.
Instead of merely operating as a home-services marketplace, Pronto could potentially evolve into an infrastructure layer for Physical AI companies seeking real-world behavioural training data.
And unlike traditional internet companies collecting clicks, browsing patterns or social interactions, Pronto’s environment is far more intimate: people’s homes.
Why Household Data Matters For Physical AI
The rise of generative AI has already transformed how software interacts with language and digital content.
But Physical AI operates differently.
Robots and embodied AI systems need real-world behavioural training to function effectively in physical environments.
That means learning how humans:
- Wash dishes
- Fold clothes
- Organise objects
- Navigate cluttered spaces
- Handle appliances
- Move around homes
- Interact with everyday environments
Synthetic simulations can help train robots to some extent. However, industry experts say real-world behavioural datasets remain significantly more valuable.
Pronto itself acknowledged this when responding to queries.
The company stated that Physical AI systems require “first-person video of people doing real tasks” in real environments.
According to Pronto, work performed by its professionals “can form a foundational data layer for physical AI.”
Cameras Inside Homes Raise Bigger Questions
Pronto confirmed that it has been running a limited AI-related pilot involving recordings of home-service jobs.
The company said customers voluntarily opt in to recordings and that professionals carry “a small camera that faces outward at the work.”
Customers reportedly receive the footage afterward.
Pronto also said:
- Participation is fully opt-in
- By default, no one is included
- Faces and identifying details are blurred automatically
- No personally identifiable information is uploaded or shared
- Footage is deleted within 48 hours
However, the company’s statements also raise important questions.
If footage is deleted within 48 hours and inaccessible beyond customers, how exactly does it become usable training data for AI labs?
Training datasets for robotics systems generally require:
- Curation
- Annotation
- Processing
- Categorisation
- Structured storage
That creates an apparent contradiction between Pronto’s deletion claims and investor descriptions of commercialisable AI training datasets.
Privacy Concerns Could Intensify
The implications become especially sensitive because homes reveal much more than conventional internet behaviour.
Domestic spaces can expose:
- Lifestyle patterns
- Financial conditions
- Cultural habits
- Family routines
- Consumption behaviour
- Organisational preferences
Even anonymised datasets may contain highly sensitive contextual information.
Under India’s Digital Personal Data Protection Act, 2023, consent must be purpose-specific and clearly defined.
Legal experts note that agreeing to recordings for customer service monitoring is fundamentally different from consenting to data being used for AI model training.
That distinction could become important if household behavioural data is eventually commercialised or shared with third-party robotics firms.
The larger ethical questions are even more significant:
- Who owns household behavioural data?
- Can such data become a commercial AI asset?
- Do users fully understand how Physical AI systems are trained?
- Can anonymised domestic workflows still reveal sensitive information?
As AI increasingly moves into the physical world, these questions may become central to future regulation.
Why Investors Are Interested
Globally, venture capital is rapidly shifting toward robotics, humanoids, warehouse automation and embodied AI following the generative AI boom.
Many investors now believe the next wave of AI dominance will depend not just on large language models or compute infrastructure, but on proprietary real-world datasets.
That makes Pronto strategically valuable.
According to reports, the startup has raised nearly $60 million so far, including backing from Glade Brook Capital and a recent $20 million investment from investor Lachy Groom.
The funding reportedly doubled Pronto’s valuation to around $200 million.
There is no evidence suggesting Pronto is collecting household data specifically for Groom’s robotics interests or affiliated firms such as Physical Intelligence.
However, the overlap between:
- Household workflow recordings
- Investor references to robotics datasets
- Growing investor enthusiasm around Physical AI
is likely to attract increasing scrutiny.
India Could Become A Global AI Data Supplier
Pronto insists its long-term objective is not worker replacement.
The company says informal workers participating on the platform could directly benefit from the AI economy through the data their work generates.
But the broader implications extend far beyond one startup.
If platforms operating inside homes evolve into AI data infrastructure companies, India could potentially become one of the world’s largest suppliers of real-world household behavioural data for global robotics systems.
And in that future, the distinction between:
- Convenience platforms
- Labour marketplaces
- AI infrastructure companies
may become increasingly difficult to separate.
The Pronto story therefore represents something much larger than a startup experimenting with cameras or AI pilots.
It may offer an early glimpse into how the next phase of artificial intelligence will be built — not just through code and algorithms, but through the digitisation of everyday human behaviour inside the physical world itself.
