Key Intelligence Insight
Granola entered one of the most saturated categories in enterprise software and won by refusing to play the category's game. While competitors raced to add bots, recordings, and repository features, Granola built a personal tool -- invisible, frictionless, augmentative. That philosophy earned it a cult following among Silicon Valley founders and VCs, then compounded into enterprise deals and a $1.5 billion valuation in March 2026. The real bet is not that Granola builds better meeting notes. The real bet is that meeting context becomes the foundation for a contextually-aware AI workspace -- and that whoever owns that context layer owns a meaningful share of how knowledge workers think.
Founding Story
Chris Pedregal encountered large language models in 2022, seven months before ChatGPT launched. His read was immediate: every tool used for knowledge work would be rebuilt on top of AI. Pedregal had already sold his previous startup, Socratic, an AI-powered tutoring app, to Google. He joined Google post-acquisition, left after four years, and began searching for a co-founder with the right complementary instincts.
He found Sam, a product designer with a background in tools for thought and knowledge management. The two did not set out to build a meeting recorder. They set out to build a contextually-aware AI workspace -- a tool that would know enough about you to help you work faster and think better. The mechanism for getting there: start with the highest-leverage entry point for context collection, where the data was rich and the incumbent tool was weak. Email was too entrenched. Meeting notes was not. Apple Notes was the real competition.
Granola incorporated in 2023, based in London. Pedregal's wife is English; the team relocated there deliberately while building explicitly for the American market -- American English spelling, US-based investors, targeting Silicon Valley founders and VCs as the initial user base. The founding thesis: build a product so good for a specific, influential user type that distribution compounds through word of mouth.
Product
Granola runs as a native desktop application on macOS -- with Windows in development and an iOS app launched in January 2026. It does not join meetings as a bot. It captures system audio locally on the user's device, transcribes in real time using third-party ASR providers (Deepgram and AssemblyAI), and then uses large language models to take whatever notes the user typed during the meeting and make them complete. The user stays in the loop. The AI fills in what was said; the human provides judgment about what mattered.
That design decision -- no bot, no stored audio -- was deliberate and initially controversial. Investors thought Granola was leaving distribution on the table. Every bot-based competitor exposed its product to every meeting participant. Granola exposed itself to none. Pedregal's counter: a great tool is consistent. It works on Zoom, on Slack huddles, on in-person conversations. A bot cannot. Audio storage is invasive. What users actually want is good notes, not recordings.
The product also generates personalized output. Two users in the same meeting -- a founder and a VC -- receive different notes because Granola generates context-appropriate summaries. This is not a minor feature. It is the structural signal that Granola is building a personal tool, not a meeting repository.
Current product lines:
Granola for macOS -- flagship app; local audio capture, real-time AI note enhancement, customizable templates
Granola for iOS -- mobile access and search launched January 2026
Granola for Windows -- in development
Granola Chat -- AI interface powered by Claude, GPT, and Gemini models for querying across single or multiple meetings
Granola Recipes -- saved automated workflows combining meeting context with expert AI prompts for post-meeting action items
Granola Spaces & Folders -- enterprise-grade shared workspaces with granular access controls, launched March 2026
Granola APIs -- personal API (Business and Enterprise tiers) and enterprise API (Enterprise tier only), launched March 2026; MCP server integration for embedding meeting context into external AI tools
The model routing layer is non-trivial. Granola tests all major foundation models and routes dynamically based on task -- note generation uses different defaults than chat. Every time a new model is released, the team re-tunes prompts to maintain what Pedregal calls "the Granola voice." The company maintains eval tooling (Braintrust) to enable model swaps overnight without breaking output consistency.
Transcription -- not LLM inference -- is currently the largest cost driver in the business. Real-time diarization (speaker identification) remains unsolved at quality thresholds Granola will ship. Echo cancellation runs on-device, built internally on open-source frameworks.
The roadmap points toward a deep research mode: querying across thousands of meetings in seconds to surface cross-meeting patterns, identify relationships, and generate dynamically updating documents. The near-term analog is already being tested internally.
Market, Competition & Business Performance
Market
The addressable market is every knowledge worker who attends meetings -- which is to say, most of the enterprise software market. The meeting notes category itself is worth several billion dollars at scale, but Granola's long-term surface area extends well beyond it. The thesis Pedregal articulates is that AI is only as helpful as the context it has. Meeting transcripts are one of the richest, most consistent context sources available for knowledge workers. Whoever builds the best meeting context layer builds the foundation for a broader AI workspace.
That market framing matters because it changes the competitive surface. Granola is not primarily competing with Otter.ai for transcription market share. It is competing for the right to be the AI layer that knows the most about how a knowledge worker thinks, decides, and communicates. That population -- and the enterprise spend attached to it -- is growing, not shrinking.
Competition
The meeting notes category is crowded and commoditizing. Otter.ai and Fireflies.ai have been in market for nearly a decade. Fathom, Jamie, and Convo operate in adjacent positions. Google Meet, Zoom, and Microsoft Teams all ship native AI note-taking features. Pedregal's own read: notes are becoming a commodity. He does not compete with them on features. He competes on depth of context and quality of experience.
The more consequential competitive threat comes from AI-native platforms. OpenAI is building memory layers and agent frameworks that could, in principle, absorb the meeting context use case. Anthropic's products move in a similar direction. Pedregal names these as the companies he actually thinks about -- not Otter, not Zoom -- because they are racing toward the same long-term destination: a contextually-aware assistant that knows everything about how you work.
The mechanism for Granola's current moat: switching cost compounds over time. A user with two years of meeting history in Granola has a data asset that is not portable. The longer they stay, the deeper the context, and the more valuable the AI's output becomes. That flywheel does not exist for a generic AI assistant starting from zero context.
Whether that moat holds against a well-resourced platform player who decides to make meeting context a priority is the defining strategic question for the next 24 months.
Business Model
Granola operates a freemium SaaS model with three tiers:
Plan | Price | Key inclusions |
|---|---|---|
Basic | Free | AI meeting notes, limited history, AI chat, shared folders, templates, multi-language |
Business | $14/user/month | Unlimited history, advanced AI models, integrations (Attio, Notion, Slack, HubSpot, Affinity, Zapier), MCP integration, personal API |
Enterprise | $35/user/month | SSO, enterprise API, admin controls, org-wide training opt-out, priority support, usage analytics |
The Business plan at $14/user/month is priced aggressively against the category. Fathom's paid tier, for comparison, starts at $19/month. The pricing reflects a land-and-expand motion: convert individuals on the free tier, then push teams toward Business, then close enterprise deals on the $35 plan with compliance and governance requirements.
The company contributes 1.5% of subscription revenue to Stripe Climate for carbon removal.
Traction
Granola launched publicly in May 2024. Within months it had become the default meeting tool for a significant portion of Silicon Valley's founder and VC community -- described in reporting as the product "everyone in Silicon Valley uses." The Information wrote about its near-universal adoption in tech circles.
Employee count reached 94 across 169% growth, with 11-50 employees on LinkedIn indicating a lean team relative to the capital raised.
Funding history:
Series B, May 2025: $43 million led by NFDG Ventures; valuation $250 million
Series C, March 2026: $125 million led by Index Ventures (Danny Rimer) and Kleiner Perkins (Mamoon Hamid); valuation $1.5 billion
Total raised: $192 million. The Series C came less than a year after the Series B -- a 6x valuation step-up in under 12 months.
Enterprise customers: Vanta, Gusto, Thumbtack, Asana, Cursor, Lovable, Decagon, Mistral AI
Notable endorsement: Pedro Franceschi, founder and CEO of Brex, cites Granola as core to Brex's AI-native rebuild: "Granola earned our trust by delivering precise, reliable summaries, and helped strengthen our written culture."
Product milestones:
May 2025: Granola 2.0 -- team collaboration and shared folders
Jul 2025: Zapier integration
Dec 2025: Recipes feature for automated post-meeting workflows
Nov 2025: San Francisco billboard campaign on "bot-free" differentiation
Jan 2026: iOS app; Outlook Calendar integration
Feb 2026: MCP server launch
Mar 2026: Spaces, enterprise APIs, Series C close
The valuation trajectory -- $250M to $1.5B in under a year -- reflects not just traction but investor conviction that the meeting context layer is a strategic asset, not a productivity feature. Granola is not the category leader in meeting notes. It may be the early leader in something considerably larger.