Runable raised a $21M Series A at a $65M post-money valuation to extend its AI agent from building small-business websites into running the ad campaigns, SEO, and social media that bring in customers. Susquehanna Venture Capital and Nexus Venture Partners co-led the all-equity primary round, with existing backers Together Fund and Array VC also participating. The Bengaluru-based startup, founded in 2025 with a 15-person team, hit a $2M annualized revenue run rate within three weeks of launching payments in March.
The pitch is that building the product is no longer the hard part. Claude Code, Codex, Cursor, Lovable, and Replit have driven the cost of shipping a functional website or app close to zero for anyone willing to prompt. Runable's bet is that small business owners still can't get customers on their own, and won't pay an agency $10,000 for a Google Ads campaign if an agent can run one for a fraction of the cost.
Co-founder and CEO Umesh Kumar framed the shift bluntly in an interview with the outlet that broke the story.
Key facts
- 01Runable closed a $21M Series A at a $65M post-money valuation, co-led by Susquehanna Venture Capital and Nexus Venture Partners.
- 02The startup hit a $2M annualized revenue run rate within three weeks of launching payments in March 2026.
- 03Runable has 1.7M registered users, with the U.S., UK, and Japan as its largest markets.
- 04Users consumed more than 1 trillion tokens over the last 90 days, with 60–70% coming from paying customers.
- 05Founded in 2025, Runable operates with a 15-person team in Bengaluru.
Runable now has roughly 1.7 million registered users, with the U.S., UK, and Japan as its largest markets and users in Brazil as well. Kumar expects Japan to move alongside the U.S. as a top market as soon as next month. Users consumed more than 1 trillion tokens over the last 90 days, with 60% to 70% of that usage coming from paying customers.
The company started as an AI infrastructure play — Kumar and co-founder Saksham Sarda built browser technology for scraping data at scale — before pivoting when users kept asking its browser agent to build slide decks and websites instead. That pivot has now stretched into ad campaigns and marketing, which Kumar calls the "grow" side of the business.
In a test run by the outlet, Runable's agent was asked to build a coffee-subscription website, set up analytics, and attract 100 visitors on a $25 ad budget. It built the site and prepared the ad campaign but stopped short of spending — advertising accounts still need to be connected externally. A similar test on Cursor hit the same wall, requiring a Meta Ads account and a third-party host to deploy.
Runable said it can run ads without customers connecting their own accounts specifically for placements on ChatGPT, via undisclosed partnerships it described as a "soft wedge." That is a narrow but interesting foothold: as AI chatbots become discovery surfaces, the ad account structure that dominates Meta Ads and Google Ads does not yet exist for ChatGPT, and whoever wires up the plumbing first has a real advantage.
Economics are the open question. Kumar acknowledged Runable operates at negative gross margins today, partly because it subsidizes AI usage for its customers. The company is using a mix of models, including some of its own, and is banking on inference costs to keep falling.
“We are seeing this path where you can provide the same quality of inference at almost 10x less cost.”— Umesh Kumar, Runable co-founder and CEO
The other risk is that the model providers Runable depends on are building competing agents themselves. Anthropic and OpenAI have both pushed hard into agentic products in 2026, and general-purpose agents Manus and Genspark target overlapping small-business users. Kumar's counterargument is that Runable takes on the operational glue — deployment, analytics, ad infrastructure — that most nontechnical founders can't stitch together on their own, and that Codex or Claude Code remain better fits for developers writing code locally.
The negative gross margin is the honest tell here. Runable is buying growth on subsidized inference, and the $21M runway is a bet that model costs collapse faster than customer acquisition does. If Kumar is right about 10x cheaper inference arriving on schedule, Runable ends up with a distribution moat that the frontier labs, focused on model quality and enterprise deals, are unlikely to build for a nontechnical shopkeeper in Japan. If he's wrong, this is a subsidy business waiting for a check-in. Either way, the shift from "AI that builds" to "AI that sells" is the more interesting agent story of the year, and Runable is now funded to test it.
Working on something we should cover, or seeing a story we missed? Send leads, documents, or feedback to hello@aichatdaily.com. For sensitive tips, see our secure tips page for Signal and PGP options.
Spotted an error? Email hello@aichatdaily.com with the URL and the issue, or read our full corrections policy.
