Redefining spreadsheets with AI: David Kircos of Quadratic

What if your spreadsheet could write its own formulas, generate Python code, build charts, and explain its reasoning - all from a single plain-English prompt? That's what David Kircos built with Quadratic. The Boulder-based founder joins the show to break down why the future of data analysis isn't a black-box AI chat, but a transparent, auditable method of analysis that lives inside a spreadsheet anyone can open, edit, and trust.

From building iOS games at 15 to co-founding a neobank acquired by Tiffin Wealth, Kircos has spent his entire career at the intersection of software and finance. Quadratic is his most ambitious bet yet - a $5.5M-backed, nearly-100K-user AI spreadsheet that's doubling monthly and redefining what it means to work with data.

The Core Insight: Methods Over Answers

The fundamental problem with dropping your data into ChatGPT and asking for insights is that you get an answer - but you can't see the work. The connection between your data and the output is a black box. You can't verify it, you can't modify it, and if the underlying numbers change, nothing updates.

Quadratic is built around a different philosophy: AI should generate a method of analysis, not just an answer. When you prompt Quadratic, it writes Python code, builds formulas, creates charts - artifacts you can open, inspect, and rerun. Change a number in your data and every downstream calculation updates automatically, exactly the way a spreadsheet should. The AI accelerates the work; the spreadsheet provides the proof.

What Quadratic Actually Does

Quadratic started as an open-source, infinite-canvas spreadsheet built on WebAssembly and WebGL - high-performance, Python-native, designed for technical users who found Excel's formula language limiting. Over time the team discovered that technical and non-technical users increasingly needed to collaborate, and AI became the bridge between them.

Today's Quadratic lets you drag and drop CSVs, connect live databases, and - as of the episode's recording - import structured data directly from PDFs. The AI can add data, write formulas, generate Python code, build charts, move cells, and delete content. It has full awareness of everything in the sheet. Prompt it in plain English; get back transparent, editable, auditable work product.

Real-time multiplayer is also native - you can see other users' cursors and watch cell edits propagate live, which drives organic, product-led sharing. Someone builds an analysis in Quadratic, shares it with their team, and the team discovers the tool through the output.

Who's Using It and How

Early power users cluster around two profiles. Product managers connect Quadratic to their app's database and query user behavior to find product improvement signals - without needing to wait on a data team. Marketers connect Google Analytics or ad-platform exports and ask questions in plain English instead of wrestling with pivot tables.

The HubSpot use case Kircos walks through live on the show captures the promise well: export 250,000 prospects to CSV, drag it into Quadratic, and ask which contacts are most likely to convert in the next 30 days based on specific attributes. Quadratic doesn't guess - it writes Python code to reorganize the data precisely, and you can double-click into that code, verify the logic, and trust the output.

Traction and Growth

At time of recording, Quadratic was approaching 100,000 total signups, with 20,000 coming in the previous month alone - a real acceleration driven entirely by AI feature launches. The growth is word-of-mouth and influencer marketing: content creators focused on AI and data analytics tools showcase Quadratic to audiences that are highly receptive. No paid acquisition to speak of.

The product-led loop compounds this: a user builds an analysis, shares the spreadsheet with colleagues, and those colleagues discover Quadratic organically. The real-time multiplayer experience - seeing cursors and live edits - makes the product feel alive in a way that accelerates adoption within teams.

Funding and Team

Quadratic has raised $5.5M across rounds. The pre-seed was led by Catapult Ventures - a hardware-focused firm that nonetheless believed in the vision when the pitch was essentially "we're building a spreadsheet from scratch and it'll take over two years before we have a shippable product." Subsequent investors include Google Ventures (GV) and Betaworks.

Kircos is direct about the GV relationship: GV is a venture capital firm whose sole LP is Alphabet/Google, but it invests for financial return, not strategic pipeline. It's not a Google partnership. The team is 10 people, based in Boulder, Colorado.

Pricing Philosophy

The goal is a generous free tier - because most Excel and Google Sheets users already have their tool bundled into something they're paying for, and Quadratic needs to compete on value before it can compete on price. The Pro plan runs $18/user/month billed annually, justified primarily by AI inference costs. Enterprise tiers add identity and data security protections.

Kircos is candid that the model is still early and will evolve as the company learns how users engage with the AI features and where the real value concentration lies.

The Founder's Path: iOS Games to Neobank to Quadratic

Kircos started coding at 12 and was building iPhone games by 15 - his first app was the approximately 8,000th submission on the App Store in 2008. The realization that you could write software, let Apple handle distribution and monetization, and have money appear in your account for something you built set the trajectory for everything that followed.

After working at the Techstars accelerator in Boulder from 2015 to 2016, he co-founded Challenger Finance - a neobank designed to help employers sponsor emergency savings accounts for employees, addressing the widely cited stat that most Americans can't cover a $500 emergency. The company was ultimately acquired by Tiffin Wealth, which bundled the product into a broader employee benefits package. Kircos had already moved on to build Quadratic before the deal closed.

The origin of Quadratic traces back to his Techstars days: he was the only Python and SQL person in a room full of spreadsheet users, and their tools simply didn't interoperate. That friction became the product.

Tools & Resources

• Quadratic - AI-native spreadsheet that generates formulas, Python code, and charts from plain-English prompts; real-time multiplayer; PDF, CSV, and database imports (quadratic.ai)
• WebAssembly + WebGL - The performance foundation Quadratic is built on; enables high-speed, infinite-canvas spreadsheet rendering in the browser
• Apple ScreenTime API / Apple Vision Pro - Kircos loaded Quadratic on Vision Pro in its infinite-canvas days and got an early preview of spatial data visualization
• Catapult Ventures - Pre-seed lead; hardware-focused firm that believed in a two-year build-before-revenue thesis
• Google Ventures (GV) - Subsequent investor; financial LP is Alphabet/Google but invests independently for financial return
• Betaworks - Additional investor in Quadratic
• Techstars - Where Kircos worked from 2015–2016 before founding Challenger Finance; the environment that planted the seed for Quadratic
• Tiffin Wealth - Acquired Challenger Finance (Kircos's neobank) to bundle the emergency savings product with broader employee benefits

Key Frameworks from This Episode

Method Over Answer

The defining principle of Quadratic's AI design. Instead of asking an LLM for an answer and trusting the black box, Quadratic uses AI to generate the method - formulas, Python code, charts - that produces the answer. Users can inspect, modify, and rerun the method. If data changes, the method updates. Trust comes from transparency, not from faith in the AI.

Bridge Building for Collaboration

Quadratic's product evolution mirrors a broader workforce dynamic: technical users (data scientists, engineers) and non-technical users (PMs, marketers) need to work in the same tool. AI became the bridge - non-technical users can now prompt Python-level analysis without writing a line of code, while technical users retain full access to the underlying code editor.

Product-Led Virality Through Output

Quadratic's growth engine isn't ads or cold outreach - it's the spreadsheet itself. When a user builds an analysis and shares it, recipients discover the tool through the work product. Real-time multiplayer (live cursors, live cell edits) amplifies this by making the collaboration layer visible. The output markets the product.

Generous Free Tier as Table Stakes

When your primary competitor (Excel, Google Sheets) is bundled into tools people are already paying for, you cannot win on price alone. Quadratic's strategy is to make the free tier genuinely useful so the product earns its way into workflows before asking for payment. Upgrade triggers are natural: heavy AI inference costs and enterprise security requirements.

Patient Capital for Patient Builds

Kircos pitched Catapult Ventures on a two-year timeline before a shippable product and no near-term revenue path. Finding investors who believed in that thesis - rather than forcing the product into a faster but shallower shape - gave Quadratic room to build the technical foundation (WebAssembly, WebGL, open-source community) that now powers its AI features.

Founder-Problem Fit from Personal Frustration

Quadratic emerged directly from a problem Kircos lived: being the only Python person in a room full of spreadsheet users whose tools couldn't communicate. The best B2B product ideas often come from working inside the exact workflow the product will eventually serve. The frustration is the spec.

FAQ