AI Infrastructure
Foundation models, training, retrieval, evaluation, orchestration, and the runtime layers below them.
Early capital for technology-native founders with the conviction to build enduring companies.
We back AI-native and frontier technology founders building intelligent systems from India to the world.
Antarix is inspired by Antariksh — the expanse beyond the visible horizon.
The IX signals intelligence systems: models, agents, infrastructure, computation, and the hidden mathematical layers behind modern AI.
We invest at that frontier.
Most companies are built to compete. A few are built because the future has made them necessary.
We back the founders who see that future early — the ones who recruit through belief, survive through ambiguity, and build with a time horizon longer than the market's patience.
“Build with a time horizon longer than the market's patience.”
ANTARIX · WORKING NOTE 001
Six convictions that orient the fund. We invest at their intersection, not their edges.
Models, retrieval, memory, and agents are settling into the base layer of every category — the way networking and databases once did.
Not only a market and not only talent: a builder-exporter ecosystem with the systems depth to ship product to the world.
Compute, data, automation, and agents will rewrite workflows in software, industry, finance, energy, and the physical world.
The next enduring companies will be built by founders who understand models, hardware, distribution, and human behavior in one motion.
The categories that matter most look small, esoteric, or unfashionable for years before the rest of the market arrives.
The best companies often look non-obvious before they look inevitable. Our job is to recognise inevitability early — and underwrite it with conviction.
Distinct sectors, one underlying motion: intelligence becoming infrastructure across software, industry, and the physical world.
Foundation models, training, retrieval, evaluation, orchestration, and the runtime layers below them.
Vertical AI for industries where domain context, data, and workflow are the moat.
Long-running agents, multi-step reasoning, and the systems that make them reliable in production.
The next generation of editors, runtimes, and observability built around AI-native development.
Pipelines, lakes, vector stores, and feature systems that make intelligence operational at scale.
Identity, policy, model security, and the trust layer for an agentic, machine-speed internet.
Software and intelligence for manufacturing, supply chains, energy systems, and physical operations.
Earth observation, satellite intelligence, communications, and orbital infrastructure.
Embodied intelligence — mobile robotics, autonomous systems, and human-machine collaboration.
Silicon, accelerators, and the frontier infrastructure that compounds AI's underlying economics.
Rails, risk, and intelligent systems for the next decade of global financial movement.
Advanced energy, materials, batteries, and software for a decarbonising industrial base.
India-native categories, global Indian founders, and companies built at population scale.
ANT–IX · MAP / 13 COORDINATES
Not a checklist. A diagnostic. The signals we triangulate when conviction is still forming.
Years inside the problem before they tried to solve it.
Founders who can rebuild the system from first principles, not just describe it.
A reason to spend a decade here that pre-dates the round.
A view of the world that does not require permission from the consensus.
The stamina to absorb compounding ambiguity without losing direction.
Early teammates who joined when only the founder could see the company.
Sharper decisions when the problem space expands, not when it contracts.
Proximity to a non-obvious problem, dataset, customer, or system the market cannot easily reach.
Every month inside the problem makes them harder to replace.
DIAGNOSTIC · 09 SIGNALS · TRIANGULATED, NOT CHECKED.
What we offer beyond the cheque — sequenced to the stage of the company, not the volume of the promise.
Operating partnership on the early decisions that compound the longest.
Sharpening the story that earns trust from engineers, customers, and capital.
Access to senior operators, technical leaders, and the early-team archetypes a category needs.
Warm introductions inside the design partners and enterprises that matter most.
Round construction, investor selection, and timing across Seed through Series B.
A discreet network of global LPs, operators, and strategics — used carefully.
Direct, founder-first availability — across the moments that don't fit a quarterly cycle.
Preparing India-built companies to be legible to global capital from day one.
Founder-investors who have spent the last decade co-building companies — and are now backing the founders bringing intelligence into India's real industries.

Operator and angel investor building consumer and technology brands from zero to one across real estate, fintech, hospitality, and sustainability. Co-founder and CMO of Equity Address and The Venya; currently leading strategy at Go Rewise, the circular supply-chain venture from listed recycler Ganesha Ecosphere. India advisory board member of the APAC CMO Council; mentor at the Abu Dhabi SME Hub. Angel investor across real estate, fintech, deeptech, and crypto — backs founders translating technical depth into brands the market actually remembers.

Serial founder and angel investor with a decade-plus building ventures across proptech, fintech, hospitality, virtual reality, and fashion. Founder of Equity Address — fractional ownership of luxury vacation homes — and The Venya, the global marketplace for boutique escapes; previously built Teleport, the VR-powered hospitality platform, and StitchMe in bespoke clothing. Angel investing across consumer brands and deeptech. Backs founders with operating depth — the stamina to build through ambiguity and the discipline to actually ship.
ANT–IX · PARTNERS / 02 PRINCIPALS
A small, deliberate network of operators, technologists, and capital partners who lend their time to Antarix companies — selected for depth, not breadth.
Intake open. Names announced as commitments are finalised.
We are deliberate about who we add — venture partners, operating partners, and domain advisors are selected for depth, not breadth. Names appear here as commitments are finalised.
ANT–IX · CONSTELLATION / 00 CONFIRMED
Categories before companies. Replace each card with portfolio detail as the fund is announced.
AI Infrastructure
Foundational training, inference, and orchestration.
Agents & Automation
Long-running agents inside knowledge work.
Industrial AI
AI-native operating systems for physical industries.
Spacetech
Orbital infrastructure and Earth observation.
Fintech Infra
Risk, identity, and movement at machine speed.
Climate & Energy
Storage, grid, and advanced manufacturing.
Robotics
Embodied AI and autonomous platforms.
Cybersecurity
Security and policy for an agentic internet.
CATEGORIES SHOWN · COMPANIES REVEALED IN TIME.
India is the rare geography where software, services, data, language, payments, identity, and infrastructure collide at population scale — and where AI-native companies are being shipped to the world from day one.
Figures rounded. Edit in lib/content.ts → india.pillars.
Native demand across enterprise, SMB, and consumer at a scale no other geography offers.
A deep, AI-fluent engineering base now training and shipping models, not just integrating them.
Companies built in India for the world — distribution from day one, not in act three.
Technical founders who turn one unit of capital into outsized engineering velocity.
An adoption surface large enough to forge product instincts unavailable anywhere else.
Software, language, identity, payments, and data converging in one regulatory and technical context.
If you are building an intelligent system that must exist, we want to hear from you.
Submissions are read by the partners. We respond to every founder we believe we can help.
Antarix Capital is designed around early identification of technology-native founders, concentrated conviction, disciplined ownership, and long-horizon compounding.
We believe the next generation of enduring technology companies will be built where AI-native systems meet India-scale markets and global ambition.
Conviction formed before consensus pricing.
A small number of companies, each capable of returning the fund.
Sizing held to thesis, not to round dynamics.
Built for a decade-plus relationship with each company.
The decade ahead will be defined by the moment models, memory, and agents stop being applications and start being substrate.
Not a copy of the US arc, not an extension of the services era. Something rarer: a builder-exporter ecosystem forming in real time.
How we triangulate signal when the company, the category, and the founder are all still becoming themselves.
FIELD NOTES · UPDATED QUARTERLY