Aryan Agarwal, building Gravity, from Bangalore. The default assumption when a 3x founder builds an AI agent platform is San Francisco. The default assumption is wrong for this company; it is also probably wrong for more companies than founders realise. This post is the geography-and-cost reasoning behind the choice, written so future founders can borrow the framework rather than the conclusion.

The argument is not "Bangalore is better than SF". It is "geography is a tool and the right tool depends on the funding model and the failure mode you are designing against". For bet four, after three shutdowns, against a unit-economics failure mode, the right tool is Bangalore.

The geography question

The geography question every founder eventually asks is: where do I build from. The standard answer for AI in 2026 is San Francisco, on the assumption that proximity to investors, conferences, and senior research talent compounds. The standard answer is correct for some companies. It is the wrong default for a bootstrapped solo founder running on a small monthly burn.

The reasoning is mechanical. Bootstrap means runway is the gating factor. Runway is monthly burn divided by reserves. The numerator (burn) is dominated by founder cost-of-living and any team salaries. Both are dramatically cheaper in Bangalore than in San Francisco. So bootstrap runway in Bangalore is multiples longer than in San Francisco for the same reserves. That is not a small effect; it is the difference between three years of runway and seven months.

Runway is the gating factor

Bootstrap is unit-economics enforcement (the argument is in why bootstrapping, not VC). The discipline only works if the founder has enough runway to wait for unit economics to balance. Geography is the lever that buys runway without raising capital.

Bootstrap runway for $50k reserves (illustrative, founder + 1 hire) Bangalore San Francisco Solo founder ~36 mo ~7 mo Founder + 1 engineer ~22 mo ~4 mo Founder + 2 engineers ~14 mo ~3 mo Source: Aryan Agarwal, illustrative bootstrap budget, drawing on Bangalore vs SF cost-of-living and salary medians.
Same reserves; multiples longer runway. Bootstrap discipline only works if you have enough runway to wait for the equation to balance.

This is the entire argument in one chart. Every other consideration, network, conferences, hiring depth, exists inside the constraint of "do you have enough months to find product-market fit before reserves run out".

Talent density and time zones

The standard objection is that San Francisco has unmatched senior AI research talent. That is true. The objection assumes you need senior AI research talent now. Most bootstrapped AI products do not; they need product engineering depth, founder-led design, and operational discipline. Bangalore has those in dense supply.

India produced approximately 1.5 million STEM graduates in 2024 (NASSCOM annual reports), with Bangalore as the largest concentration of working software engineers in the country. The talent pool for the kind of work Gravity needs in year one is wider in Bangalore than the founder needs. The talent pool for senior AI research is wider in San Francisco; that hire happens later, if it happens.

Time zones are the underrated advantage. From Bangalore, the working day overlaps with all of Asia in the morning and with the US West Coast from late evening. That covers 70% of the global SaaS market in synchronous hours without forcing the founder to live nocturnally. San Francisco has the opposite asymmetry; Asia is unreachable in working hours.

What Bangalore actually costs you

Be honest about the trade. Three real costs.

Network density of investors is lower. If you are bootstrapping, this is a non-issue; if you are raising, it is a meaningful penalty. The argument for staying in Bangalore is conditional on the funding model; for VC-funded companies the math changes.

Ambient signal from the AI conference circuit is lower. The hallway conversations at NeurIPS, the parties after the OpenAI dev days, the chance encounters with researchers; those happen less in Bangalore. The replacement is structured: RSS, X, podcasts, deliberate reading. It works for most product-engineering companies; it is genuinely a loss for research-heavy ones.

Hiring senior research talent is harder. Real constraint, not an immediate one. Bootstrap means hiring is deferred until unit economics support it; by the time the constraint matters, the founder has options including remote senior hires from anywhere.

The stack does not care

The infrastructure that runs Gravity does not know where the founder is. Cloudflare Workers run in 300+ cities; OpenAI APIs respond in milliseconds from Bangalore as readily as from San Francisco; Stripe, Auth0, every SaaS the company depends on serves global. The product is built on globally available primitives. That is the structural reason geography is no longer a precondition for serious software companies.

Distribution is also global. The waitlist comes from X, LinkedIn, Reddit, Hacker News; the audiences live in every time zone; the content is read in English by people in San Francisco, Berlin, Singapore, and Lagos. The founder being in Bangalore changes nothing about distribution. The product is global; the audience is global; the costs are local. That is the asymmetry that makes bootstrap from Bangalore rational.

This is consistent with the broader playbook in bootstrapping an AI agent platform in 2026 and the cost reasoning in economics of bootstrapped AI agents.

When SF is the right call

For founders weighing the same question, the rubric I would offer: SF is the right call when (1) you are raising VC and the network density compounds the round, (2) you need senior research talent in the first 12 months, or (3) the category requires Bay Area access to specific decision-makers (enterprise sales into Bay Area unicorns is a real example).

If none of those three apply, geography is a cost lever, and the cost lever points to wherever the founder's existing roots and runway are best aligned. For me that is Bangalore. For other founders it is Lisbon, Bengaluru, Seoul, or Buenos Aires. The argument generalises beyond Bangalore-vs-SF; the right framing is "match geography to failure mode and funding model".

Frequently asked questions

Why are you building Gravity from Bangalore?

Bangalore is bootstrap-tolerant in a way San Francisco is not. The cost-of-living math means a single founder with a small monthly burn can run for years on bootstrap funding. Talent density is high, the time zone covers Asia in working hours and the West Coast in late evening, and the engineering depth is real. The geography matches the funding model.

Do you lose anything by not being in San Francisco?

Three things, real but bounded. Network density of investors is lower; that matters less for a bootstrapped company. Ambient signal from the AI conference circuit is lower; that is replaced by RSS, X, and structured reading. Hiring senior research talent is harder; for a single-founder bootstrap, that is not a near-term constraint.

Is Bangalore the right place for an AI agent platform?

Yes for bootstrap; the answer would be different for a venture-scaled company. The infrastructure stack is global, so the technology runs anywhere. The product market is global, so distribution runs anywhere. The cost base is local, and Bangalore's cost base is bootstrap-friendly. The infrastructure is global; the team is local; the costs are tolerable.

How much cheaper is Bangalore than SF for a solo founder?

Living costs are roughly five to seven times lower; engineering salaries are roughly three to four times lower if hires happen later. The full ratio means the same bootstrap budget runs four to five times longer in Bangalore than in San Francisco. Runway is the gating factor for bootstrap; geography directly extends it.

Would you ever move to SF?

Possibly later, for a specific reason: a category that requires Bay Area network density to win, a hiring round that needs Bay Area senior talent, or a fundraising posture that benefits from proximity. None of those reasons apply to bet four. Geography is a tool. Bangalore is the right tool right now.

Three takeaways before you close this tab

Sources