Tab Emerges From Stealth With a $300 Million Bet on Texting
Tab has emerged from stealth at a $300 million valuation, TechCrunch reported on Oct. 7, 2026. The company was started earlier in 2026 by Brennan Erbz, Stafford Schlitt and Ammar Amdani.
Tab works by text message. Users message an assistant on iMessage or WhatsApp for help with tasks such as ordering groceries or finding birthday presents. The assistant has its own phone number, computer and wallet. Investors in the round include SV Angel, Valar Ventures and American Spirit. Tab declined to share exact funding details for the round.
The broader context here is consolidation around messaging. There is no app to install and no new client to maintain. For people who already handle work and personal life in chat threads, asking for errands in the same thread is a small change in habit. That lowers friction for users. It pushes complexity to the back end, where tool calls (the assistant's ability to use outside services), session state (its memory of where it is in a multi-step task) and third-party integrations have to work without a custom screen to hide behind.
Looking at what this means for builders, the trio of resources is the detail to watch. A phone number means a lasting identity that can place calls, pass verification checks and receive callbacks. A computer means a workspace that stays active beyond one message and keeps track across steps. A wallet means permission to spend on a user's behalf. Each idea is simple to state and hard to run safely at scale, because permissions must be limited, actions must leave a clear record, and users need a fast way to cancel access when an order goes wrong or login details leak.
In my view, the $300 million figure should be read carefully. Tab did not disclose the amount raised, so the number reflects how investors are pricing risk, not how much cash was invested or how much customer traction exists. Early prices for agent startups often come down to a bet on the team and on control of places where money changes hands, like checkout and purchasing. Groceries and gifts sound simple. They involve unclear requests, substitutions, budgets and clean recovery from mistakes across shops that were not built for machine shoppers.
Looking at what this could enable if the execution holds, the payoff is practical rather than dramatic. Dependable help with low-stakes but annoying tasks would save attention without forcing users to learn prompts, plugins or workflows. I saw this with my own kids, who kept tools that removed a chore and dropped tools that added one. For developers, the same building blocks of limited access, clear confirmation rules and histories users can inspect would carry over to company help desks and internal tools. The value adds up in busywork avoided, not in novelty.
