The Unsolved Problem Nobody Will Shut Up About
Here's a fact that should embarrass the entire software industry: after four decades of building business tools, nobody has shipped a to-do list that actually thinks for you. Notifications? Solved. Kanban boards? Solved to the point of exhaustion. Reminders that follow you into three different apps? Every last one, solved. But an app that takes a half-baked meeting outcome and turns it into a filed issue, a drafted follow-up email, and a CRM update without you touching a keyboard — that's been the white whale since the first shared calendar was ever deployed.
Tana thinks it finally caught the whale. The Norwegian startup raised $25 million in a round led by EQT Ventures at a $110 million post-money valuation, and the waitlist told the story: over 160,000 people queued up before this thing even emerged from stealth. That's not the behavior of a product that needs convincing. That's a product that's been quietly passing around Slack channels like contraband.
What Tana Actually Does (And Doesn't)
Tana is not another note-taking app with a prettier interface. Not exactly. The platform operates as an agentic meeting system where AI agents perform real work during native video calls. The product page makes the distinction blunt: "Tana doesn't take notes. It captures your calls, files the issue, drafts the email and updates the CRM before you hang up, and remembers what everyone promised."
Read that again. It captures. It files. It drafts. It updates. The emphasis sits on verbs that currently consume twenty minutes of your afternoon.
There are actually two products here. Tana Outliner (at outliner.tana.inc) is the note-taking and knowledge-graph builder — the Supertags, nodes, and all that structured thinking stuff. Then there's the Tana meeting platform, the agentic layer that sits on top. The company treats these as separate offerings but the combined vision is singular: one memory that every call contributes to, one place to ask "what did we decide about SSO?" and get an answer with attribution, ownership, and the thread of calls it came from. It's the same grounding problem we unpacked in Grounding Agentic Data Platforms: How SAP's Knowledge Graph Strategy Changes the Game — agents are only as useful as the structured memory they can query.
What makes the meeting capture technically interesting is that it works across Zoom, Microsoft Teams, and Google Meet without inserting an active bot participant. Instead, it captures from your local machine. No meet.tana.ai join link floating in front of clients. No "recording bot has joined the call" awkwardness. That's a meaningful architectural choice for anyone who's ever had to explain to a nervous CFO why a third-party robot is sitting in on the budget review.
The Backers and Their Patience
$25M led by EQT Ventures at $110M post-money. Round also included Lightspeed Venture Partners, Northzone, Alliance VC, and firstminute capital. On the angel side: Lars Rasmussen (co-creator of Google Maps) and Arash Ferdowsi (Dropbox co-founder). The list reads like it was assembled by someone who knows exactly what "enterprise software patience" costs.
Sheila Gulati, founder and managing director of Tola Capital, put the pitch plainly in her interview with TechCrunch: "But this is a miraculous experience. I use it to run our VC firm." She also flagged the competitive landscape honestly — "this is a market that will have real competition and players who want to win" — before circling back to why she backed this particular team anyway: "their vision of productivity is completely different." It's a competition warning worth taking seriously; serial entrepreneur Bhavin Turakhia, for one, is already pouring $30M of his own money into Neo, an AI-native workspace built on the premise that work software needs rebuilding, not patching.
That phrase, "completely different", matters more than it sounds. Most productivity startups describe their vision as "streamlined" or "intuitive" or some adjective that could sell you a SaaS onboarding tutorial. Gulati's insistence that Tana's vision is different in kind rather than degree signals something investors rarely get to say about tools in this space.
Built on the Bones of Google Wave
The lineage is not accidental. The founders carry what the TechCrunch coverage calls "battle scars" from building Google Wave, that famous Google product that launched too early, was too ambitious, got killed before the world had Chrome extensions fast enough to appreciate it, and then leaked its best ideas into everything from Slack threads to Notion blocks over the following decade.
Building Google Wave taught a particular kind of patience. You don't try again unless the lesson was genuinely structural rather than just a timing mistake. Tana's architecture, knowledge graphs that accumulate meaning over time, agents that execute rather than merely suggest, feels like an answer to "what would Wave have been if it shipped when MCP, LLM function-calling, and local processing existed?"
The CEO's stated ambition in the TechCrunch interview is characteristically unbounded: "We envision Tana used by all knowledge workers." No niche. No wedge. All knowledge workers. That's either hubris or the only answer that makes sense for a knowledge graph play, and it's the machine-readable-content bet behind pieces like How GraphRAG and Agentic Data Platforms Are Rewriting AI Search.
Why AI Cloud Infrastructure Teams Need This
Every company building AI cloud infrastructure, whether they're in Bengaluru, Berlin, or Bellevue, faces the same operational tax: coordination overhead that scales superlinearly with headcount. More engineers means more meetings. More meetings means more decisions that evaporate because nobody filed the ticket. More integrations means more context-switching. The tools that supposedly fix this (Jira, Linear, HubSpot, Slack, Gmail) have actually become the problem by fragmenting context across more surfaces than a human can hold simultaneously.
Tana's integration surface tells you exactly who it's targeting. GitHub, Slack, HubSpot, Linear, Jira, Google Calendar, Outlook, and, notably, Claude Code, Codex, Gemini, and Cursor. Those last four aren't project management tools. They're the development environments where AI cloud infrastructure engineers actually spend their hours. Tana is saying: I'll meet you where the work happens, not where the PM wants to check status.
The platform also speaks MCP, Model Context Protocol, which means it plugs into the same agent-to-tool plumbing that most modern AI infrastructure stacks are standardizing on. For a team already running RAG pipelines, agent orchestration, and multi-model deployments, Tana slots into the same architectural pattern rather than demanding yet another SaaS with its own bespoke integration API. It arrives at a moment when agents are reshaping how software gets bought and secured, a shift we mapped in How AI Agents Are Reshaping SaaS Purchasing — and Why AI Cloud Infrastructure Companies in India Matter.
The 160K Question
Here's what the waitlist number actually says. A 160,000-person queue for a product that requires you to let AI agents file your Jira tickets and draft your CRM updates means something has shifted in trust. Five years ago, giving a machine permission to send email on your behalf would've sounded like a security incident waiting to happen. Today, people are lining up for it.
The pricing model is a 30-day free trial with no credit card. That's not B2B enterprise SaaS language, that's consumer app language applied to a team tool. It signals the founders are betting on bottom-up adoption through individual knowledge workers who then drag their teams in, which is the Dropbox playbook (not coincidental, given Ferdowsi's involvement).
Tana isn't going to replace your project management tool this quarter. It doesn't need to. The bet is that the project management tool becomes a destination that agents write to, rather than a workspace that humans have to maintain manually. If that shift lands, the entire category of "tools for teams to coordinate work" restructures around who can execute reliably rather than who can visualize progress in a Gantt chart.
The to-do list was never the unsolved problem. The unsolved problem was work that organizes itself. Tana's 160,000-person waitlist suggests we're closer than anyone expected. Whether the execution holds at scale, whether "captures your call and files the issue" survives contact with a 200-person org that has compliance requirements and a very literal interpretation of who's allowed to write to HubSpot, that's the next chapter. But the waitlist doesn't lie.
