One investor. One team. One conviction.

The first chapter answered whether private frontier-class AI could run without a GPU estate. The answer was yes. The next chapter is scale.

Six convictions ringed around one centre: sovereign, nothing leaves, yours to own, in production, CPU-first, built in India.SovereignNothing leavesYours to ownIn productionCPU-firstBuilt in India
Most of our best weeks look like the flat part. That is where the work happens.
  • Backed by Thangam Capital, USA
  • An advisory board across health, industry and defence
  • Built in India, for anyone who cannot ship their data out

The investor

Thangam Capital, United States

An ambitious US venture firm backing deep technology, with six companies in the current portfolio and more than $2 million committed to each.

What arrives alongside the capital matters as much as the capital: strategic reach into customers, access to US enterprise buyers, and operating help from people who have taken hard technology to market before.

6Deep-tech companies backed
$2M+Committed to each
USAWhere the firm is based

Portfolio figures as provided by the firm.

6 Deep-tech companies backed, $2M+ Committed to each, and what arrives alongside the capital: strategic reach into customers, access to US enterprise buyers, and operating help.Thangam Capital · United StatesPortfolio figures as provided by the firm$2M+$2M+$2M+$2M+$2M+$2M+6 Deep-tech companies backed · $2M+ Committed to eachStrategic reachinto customersAccess to USenterprise buyersOperating help frompeople who have done it
What this shows. 6 deep-tech companies backed, $2M+ committed to each — and the three things that arrive alongside the capital.Why it matters. What arrives alongside the capital matters as much as the capital: strategic reach into customers, access to US enterprise buyers, and operating help from people who have taken hard technology to market before.

Executive team

People who build and ship.

Chief Operating Officer

Ms. Margaret

Owns operations across Genie and Swayam, and holds engineering, customer success and PMO to a single shipping calendar. Turns technical vision into product that lands on customer infrastructure.

Chief Financial Officer

Mrs. Mary

Owns finance, capital strategy and governance across the group — structuring the runway so engineering never has to trade away sovereignty, security posture or scope.

Chief Customer Success Officer

Dr. Subhash

Owns customer outcomes from first pilot to renewal, and runs the playbook that turns a four-week proof-of-concept into a multi-year sovereign-AI programme.

CEO · Superchat Robotics (Swayam)

Mrs. Ramya

Runs the robotics wing as a full P&L — the humanoid bipedal programme and the AURA dual-drone system, from control theory to production hardware.

Advisory board

Experience that helps us scale.

Capital & Strategy

Mrs. Sree

Brings a family lineage of scaling technology services businesses across the US and India, and the network to unlock institutional capital and US enterprise access.

Enterprise AI

Mr. Shikhar

Head of AI at a large global pharmaceutical company. Runs enterprise AI at pharmaceutical scale — inside the exact regulated-data environment Genie is built for.

Retail & Operations Tech

Mr. Sai

Former CTO of a national retail chain. Ran technology across that retail operation, and advises on the cloud, edge and on-prem trade-offs that decide whether an AI programme lands.

Global Services Delivery

Mr. Dilip

CEO of a global software services firm. Has scaled global services delivery to hundreds of enterprise clients.

Editorial illustration of a clinician and nurse reviewing information at a hospital workstation during the night.
We build for rooms like this one — where the wrong answer about data has a cost somebody else pays.

How we work

Technology built. Advantage compounding.

The core architecture is in production and already serving eight live deployments. The job now is to turn that technical lead into three repeatable commercial engines: Enterprise Intelligence, Embedded Intelligence and Physical Intelligence.

Research keeps widening the lead: better cache behaviour, more efficient expert routing, faster decoding and control loops that close in milliseconds. None of it stays theoretical. It compounds a platform customers can run now.

Tokens per second, week by week

Most of our best weeks look like the flat part. That is where the work happens; the step comes later, once, and usually on an unremarkable Tuesday.

The four research lines

The future of AI isn't somewhere in the cloud. It's wherever it's needed.

You will not need to trust us on any of this. Ask for the box. Pull the cable.

Try Genie