Yasmin
Razavi

Eighteen months. That is how long Yasmin Razavi tracked cross-border payroll trends before she saw what most people in venture capital were still missing. The specific friction of paying a contractor in Lagos from a payroll system built for Ohio. The compliance nightmares that accrued at international borders. The sheer mechanical cost of moving money to someone in a different country who just did real work. She spent those months accumulating data points, building a mental model, until one day she recognized it: this was a category-scale problem, and nobody had yet built the infrastructure to fix it. That observation became her conviction in Deel, now one of the largest global payroll and compliance platforms in the world.

This is how Yasmin Razavi invests. She calls it "spending mental calories" - a phrase that sounds almost tedious until you understand what it means in practice. Not pattern matching. Not founder charisma. Not a quick gut check against a template. Deep, granular, time-consuming analysis of unit economics, market dynamics, and the specific mechanics of why a company is doing what it is doing. In a business where speed is celebrated and terms are competed over, she is deliberately slow where it matters.

The Anti-Gut Investor

Spark Capital operates from San Francisco and Boston, with roughly $4.2 billion in assets under management. It sits comfortably in the tier of firms that define categories - the early backer of Twitter, Tumblr, Warby Parker, and Coinbase. Yasmin joined the growth team in 2017, the year after she left Snap, where she had built the technology stack behind the app's monetization engine. She brought something unusual: a product manager's understanding of what enterprise software actually has to do to sell.

Her explicit rejection of gut-driven investing is not just a personal preference. It is a thesis about fairness. "I believe gut-driven investing often favors a certain type of entrepreneur," she has said publicly, "and disadvantages founders that are not performers, but who are nonetheless building solid businesses." The charismatic pitcher. The narrative genius. The person who can command a room at 8am on a Monday. Razavi is conscious of that bias in a way that most investors - who built careers on precisely that charisma-signal - are not. Her corrective is data, and it has paid out.

From Tehran to Toronto to the Board of Anthropic

She grew up in Tehran, left for the University of Toronto to study engineering, then pivoted to Harvard Business School for her MBA. The engineer-to-MBA path is well-worn in venture capital. What distinguishes Razavi's version is the stop in London at Index Ventures, one of Europe's most internationally minded VC firms, and then the detour through McKinsey, where she served financial institutions in the UK on product strategy. London, in particular, stuck. Her favorite city, she has said - and her invocation of the Dr. Johnson line ("When you tire of London, you tire of life") suggests someone who means it.

The Snap years - building the product layer for monetization - gave her something harder to acquire: firsthand knowledge of how advertising infrastructure scales, how developer systems integrate with commerce, and how a platform business actually makes money. It is the kind of operational texture that informs every portfolio decision at Spark. When she evaluates go-to-market strategy - which she considers the primary differentiator between good companies and great ones - she is drawing on experience, not theory.

Act One. Act Two. Act Three.

Razavi's investment mandate sits at a specific moment in a company's life: after product-market fit, before scale. The $5M to $15M check range, typically at Series B. The founders she backs have already proven they found something real. Her job is to help them figure out what to do with it. "I help the founders I work with out-execute their competitors in their act one," she has said, "and think through and capitalize on their act two and three."

Go-to-market is her obsession. Not as a buzzword - as a discipline with specific mechanics. When to hire the first account executive (too early, and you waste a great rep's time; too late, and you cede territory). How to restructure sales leadership at the $10M, $50M, and $100M revenue inflection points. How pricing and packaging change as the customer base evolves. How compensation structures create or destroy incentive alignment. She has written and spoken extensively on these questions, and her portfolio companies are the test bench.

Marqeta, the card-issuing infrastructure company, went public in 2021. Deel became one of the fastest-growing HR and payroll platforms in history, hitting unicorn status on the back of the remote work surge. Capitolis, operating in capital markets optimization. Earnin, the cash advance platform. Plaid. Coinbase. Affirm. Rapyd. The portfolio reads like a syllabus for the next decade of financial infrastructure.

The Anthropic Bet

In May 2023, Spark Capital led Anthropic's $450 million Series C. Yasmin Razavi joined Anthropic's Board of Directors. This was not a speculative swing into AI hype. By the time Spark moved, Razavi had already built a Kuhnian framework for what was happening in AI - framing it not as incremental progress, but as a paradigm shift following a decade of compounding work in deep learning, scaling laws, and reinforcement learning from human feedback.

What distinguished Anthropic for her was not just the technical pedigree - though the team, including founders from OpenAI, brought genuine research credibility. It was the explicit commitment to safety. "Getting it right is crucial," she wrote in announcing the investment. "Beyond their world-class technical expertise, the Anthropic team stood out to us in their resolute commitment to investing in AI safety and alignment." She described AI development as "more like brewing beer" - experimental, requiring iterative refinement - and positioned Constitutional AI, Anthropic's approach to making models behave according to stated principles, as the kind of alignment work that matters at civilizational scale.

The Qdrant investment, announced in January 2024, was the infrastructure complement. Qdrant builds the vector database technology that makes AI applications actually work at scale - the layer between raw model capability and real-world deployment. A $28M Series A, with Spark leading. Razavi was direct: she and the team were "thrilled to partner with the Qdrant team as they continue to build the most powerful vector search database and infrastructure." Read across Anthropic and Qdrant and the pattern is legible: she is building a position in the full AI stack, from model to memory.

She reads exclusively print fiction. No e-readers. She has a soft spot for London, where she lived during the McKinsey and Index years. The detail that sits most distinctly against the rest of the profile: someone whose professional life is about extracting signal from financial data, spending her personal hours in the un-analytical space of narrative. Perhaps that is not a contradiction. Perhaps understanding how stories work - how they arc, how they create inevitability in retrospect - is exactly the skill that lets a growth investor see where a company is going before the numbers show it.

Fintech, she has argued, is not a winner-take-all market. Regulatory complexity, local payment methods, and cross-border friction create openings for multiple billion-dollar players simultaneously. This is a thesis that runs directly counter to the zero-sum framing that dominates much of technology investing, and it has guided her to build a portfolio that includes companies competing in overlapping spaces without canceling each other out. Marqeta and Deel both touch the movement of money. They are not rivals. They are infrastructure at different layers of the same stack.