Jordan Tigani spent years building Google BigQuery into a petabyte-chomping beast. Now he's betting that most companies don't actually need one.
That conviction—equal parts heresy and shrewd market reading—just attracted $52.5 million in fresh capital for his startup MotherDuck. Felicis led the September 2023 Series B, with participation from Andreessen Horowitz, Madrona, Amplify Partners, Altimeter, Redpoint, and Zero Prime. The round valued the Seattle company at $400 million post-money, more than doubling its November 2022 Series A valuation of $175 million and bringing total raised to roughly $100 million.
Perhaps more revealing than the money: MotherDuck simultaneously announced it was ditching its waitlist. After months of careful cultivation, the company was ready to let anyone kick the tires on its cloud analytics platform built atop DuckDB, the open-source database that's become something of a cult favorite among data engineers who've grown weary of complexity.
An Architecture Born From Skepticism
Tigani's pitch sounds almost quaint in an era when every data warehouse vendor races to tout distributed computing prowess. MotherDuck runs queries across both your laptop and the cloud—what the company calls "hybrid execution." The idea? Most analytics workloads don't actually touch petabytes of data. They touch megabytes, maybe gigabytes.
Why spin up costly cloud infrastructure when much of the computation could happen locally?
It's the kind of contrarian thesis that gets dismissed at conferences until suddenly it doesn't. Tigani and co-founders Ryan Boyd and Valentino Tereshko detailed the technical approach in a CIDR 2024 paper, positioning their creation as an alternative to the Snowflakes and Databricks of the world. Those platforms excel at scale. MotherDuck excels at not needing that scale.
The company formalized its relationship with DuckDB Labs through a partnership that saw the open-source project's stewards acquire shares in MotherDuck. Crucially, DuckDB itself remains MIT-licensed under the DuckDB Foundation—a structure reminiscent of how MongoDB and Databricks threaded the needle between commercial ambition and open-source ethos.
From Zero to 125 (Maybe Too Fast?)
At the Series B announcement, MotherDuck employed 32 people and served nearly 2,000 users. The plan called for growing to 45 employees by year-end 2023, focused on engineering and go-to-market.
LinkedIn now shows 125 employees spread across Seattle, New York, Amsterdam, and San Francisco. That's aggressive expansion—perhaps more aggressive than some venture-backed startups manage comfortably. Whether MotherDuck can maintain its technical velocity while scaling culture remains an open question.
The company reached general availability in June 2024, shipping features like read-scaling and SOC 2 Type II compliance along the way. Customer wins tell a compelling story: ATM.com slashed analytics costs roughly 65% versus SingleStore. FinQore reduced pipeline times from eight hours to eight minutes.
Those aren't marginal improvements. They're the kind of numbers that make CFOs pay attention.
The AI Angle (Because Of Course There Is One)
MotherDuck's architecture includes something called "hypertenancy"—individually provisioned cloud compute instances dubbed "Ducklings" that scale independently per user. It sounds almost whimsical until you realize what it enables: direct integration with AI assistants like Claude, ChatGPT, and Gemini through MotherDuck's MCP Server.
Each agent gets sandboxed compute and traceable SQL. The company prices AI functions at $1 per "AI Unit" for text-to-SQL and embeddings, a calculated bet on enterprises deploying more autonomous AI systems. Whether agentic analytics becomes the next big thing or another overhyped trend—well, that's the gamble.
There's also a WebAssembly SDK for in-browser analytics, letting developers push compute to client devices. The strategy feels scattershot or prescient depending on your tolerance for optionality.
Riding (Or Creating?) a Wave
DuckDB has accumulated 25,000 GitHub stars by December 2024 and over 22 million monthly PyPI downloads as of late 2025. The project's adoption provides crucial validation for MotherDuck's commercial thesis—the same dynamic that propelled MongoDB and Databricks from open-source curiosities to venture darlings.
Described frequently as "SQLite for analytics," DuckDB appeals to developers exhausted by the operational overhead of distributed systems. Sometimes simplicity wins.
The addressable market remains enormous, even if sizing estimates vary wildly. Market Research Future projects the cloud data warehouse sector growing from $36.3 billion in 2025 to $183 billion by 2035. Snowflake has cited a $342 billion TAM by 2028. Take your pick, but the opportunity is real.
What Comes Next
MotherDuck launched a European region on AWS eu-central-1 in September 2025, continuing its integration expansion. According to research firm Sacra, the company quietly raised an additional $33 million Series B extension in May 2025, pushing total funding to approximately $133 million.
Viviana Faga, the Felicis general partner who led the round, joined MotherDuck's board. The round announcement came bundled with 11 new integrations including Cube, Metabase, LlamaIndex, and Streamlit—the kind of ecosystem play that separates infrastructure wannabes from legitimate contenders.
The bet here is unconventional: that scale-down matters as much as scale-up, that hybrid execution beats pure cloud, that most companies are over-engineering their analytics stack. It's a thesis that requires both technical chops and market timing.
Top-tier VCs are backing Tigani's vision at a premium valuation. Whether that confidence proves justified depends on something venture capitalists hate to admit: execution details that won't reveal themselves for years.
For now, the BigQuery founder is building something deliberately smaller than what he left behind. That might be exactly what the market needs.
