Pocket Raises $11 Million as AI Note-Taking Hardware Finds a Real Market

Pocket Raises $11 Million as AI Note-Taking Hardware Finds a Real Market

The graveyard of AI hardware is littered with ambitious gadgets that promised to replace the smartphone and delivered mostly disappointment. The Humane AI Pin was pulled from the market and its assets sold off. The Rabbit R1 became a punchline. So when a startup called Pocket announced on June 29 that it had raised $11 million to make yet another AI device, the reflexive skepticism was understandable. The difference this time is the traction: Pocket says it has already sold more than 130,000 of its $129 pucks and is running at a $27 million annualized revenue rate. It is not trying to reinvent the phone. It is trying to take better notes.

The round was led by Accel with participation from Y Combinator, which previously backed the company, and a roster of operator-investors that includes ElevenLabs co-founder and CEO Mati Staniszewski, Vercel founder and CEO Guillermo Rauch, and Opendoor CEO Kaz Nejatian, the former COO of Shopify. That mix of a top-tier venture firm and a bench of hands-on founders signals something the AI hardware category has lacked for two years: investor conviction rooted in shipping product and real revenue rather than a splashy keynote.

A puck, not a pin

Pocket's device is deliberately unglamorous. It is a credit card-shaped puck that magnetically attaches to the back of a phone. You turn on recording before a meeting, and the device captures and transcribes the conversation. An accompanying phone app then generates summaries, answers questions about what was said, builds mind maps, and reformats the raw transcript into templates. The hardware carries two microphones with a roughly 15-meter range, 64GB of on-device storage, up to four days of battery life, and support for more than 120 languages.

Crucially, the basic value proposition is subscription-free. Buy the $129 puck and you get unlimited recordings, transcriptions, and to-do items with no recurring fee, a pointed contrast to rivals that gate core functionality behind a monthly plan. Pocket does sell a $200-per-year tier that unlocks unlimited AI summaries, unlimited queries to its assistant, daily highlights, and file attachments, but the entry point is a one-time hardware purchase that does something useful out of the box.

That framing is the whole thesis. Co-founder Akshay Narisetti, who was a founding member of rival note-taking startup Omi, argues the existing crop of meeting tools solved the wrong problem. "We thought every meeting notetaker was built for online conversations, but nothing was geared towards real-life talk," Narisetti told TechCrunch. "AI really needs a lot of context to work better for us, and a lot of that context exists offline." His co-founder, Gabriel Dymowski, previously founded a blockchain-based document management startup.

Why this niche is working when the moonshots failed

The clearest way to understand Pocket is by contrast. Humane and Rabbit asked consumers to carry a second, general-purpose AI companion and, in effect, to believe it could do a little of everything. Both stumbled because they were solving no single problem well enough to justify their existence alongside a phone that already did those things better. Pocket inverts the bet. It picks one painful, high-frequency workflow, capturing and organizing spoken conversations, and builds hardware tuned specifically for it.

Accel partner Cecilia Wang, who led the deal, framed the appeal around the offline, in-person moments that screen-based transcription tools miss. "You can record on the go, offline, and in the field, which is exactly how lawyers, salespeople, doctors, real estate agents, construction workers, and students use Pocket today," Wang said. "Not only are people present rather than shifting focus to take notes, but more information and insights also get captured than ever before, which otherwise would have been lost. Over time, that accumulation of insights is really valuable: one central place where your ideas, conversations, and thoughts live, rather than scattered and lost."

The market data backs the thesis. Pocket is far from alone. Competitors including Plaud, Mobvoi, Anker, Viaim, and Vibe have all piled into dedicated note-taking hardware, and the category leader, Plaud, recently said its software business crossed $100 million in annual recurring revenue after shipping more than two million devices. When a category has a nine-figure incumbent, it is no longer a novelty; it is a market. The single-use-case AI gadget, it turns out, may be the format that finally works.

The ambient recording problem

Success brings a thornier question into view: what does it mean to have hundreds of thousands of always-ready recorders circulating through meetings, doctor's offices, and job sites? Pocket, like every device in this category, depends on capturing other people's voices, and consent laws vary sharply by jurisdiction. Many U.S. states require all parties to a conversation to agree before it is recorded; the professions Wang cites, lawyers, doctors, real estate agents, sit squarely in territory governed by confidentiality and privilege rules. A device that quietly captures "more information than ever before" is precisely the device that regulators and privacy advocates will scrutinize.

Pocket's on-device storage and offline capability cut two ways here. They are a genuine privacy feature, keeping sensitive audio off the cloud until a user chooses to process it, but they also make ambient capture frictionless. The company has not made consent tooling a centerpiece of its pitch, and that omission is worth watching as the installed base grows. The winners in this category will likely be the ones that build consent, disclosure, and data-retention controls into the product rather than treating them as afterthoughts.

Software is the real moat

Pocket knows the moat is not the hardware. Pucks are commoditizing, and smartphones remain a perfectly capable substitute for anyone willing to run a note-taking app. So the company is racing on software and integrations. It already connects to Google Calendar, Google Drive, OneDrive, Obsidian, Claude, and Cursor, offers webhook support and custom workflows for enterprise buyers, and ships a Model Context Protocol server to wire its AI assistant into external databases. The roadmap points toward automating the after-meeting grind: drafting follow-up emails, updating CRMs, and generating action items. The company is banking on shipping software fast, which is both the strategy and the vulnerability, since Granola, Zoom, Fireflies, Otter, and Read AI are chasing the same outcomes without asking anyone to buy a device.

What to watch

The signal from Pocket's raise is bigger than one company. After two years of AI hardware being synonymous with hype and returns, capital is flowing back into gadgets, but with a sharply narrowed thesis: solve one painful workflow, price it honestly, and let the software compound. Watch three things next. First, whether Pocket can convert one-time hardware buyers into recurring $200-a-year software subscribers, the number that will determine if the business is durable or just a well-timed hardware moment. Second, whether the privacy and consent questions around ambient recording harden into regulation as the installed base scales into the millions. And third, whether the phone makers themselves, Apple and Google, decide that on-device meeting transcription is a feature they should own, at which point the entire category of pucks and pins has to prove it can outrun the operating system.