Garima Kapoor entered the storage business by way of economics, which may be the sensible direction. Data looks ethereal only until someone tries to move a petabyte of it. Then it acquires weight, rent, and a travel budget. Kapoor's decade at MinIO has been an extended study of that inconvenient material fact: information may be copied perfectly, but it never travels for free.
Her original motive was less polished than the origin myths founders are encouraged to manufacture. She wanted to build something of her own. She had studied business economics at the University of Delhi, completed a master's in economics at Banasthali Vidyapith, and earned a PhD in accounting and finance from Nirma University. Her dissertation dealt with the value of financial and non-financial disclosures in Indian pharmaceutical companies. It was a long way from object storage, at least by the shortest route.
Kapoor has said that the PhD's enduring value was the discipline of getting deep into a problem. The thesis topic stayed behind; the method travelled. It helped her cross from finance and angel investing into a technical field where she had no formal computer-science training. She jokes that the Bay Area eventually makes everyone a technologist. Silicon Valley has produced less charming forms of compulsory education.
A cloud wherever the bytes happen to be
MinIO began at the end of 2014 with Kapoor, Anand Babu Periasamy, and Harshavardhana. Their reasoning started with two observations. Data would grow exponentially. Much of it would be created outside the public cloud. Amazon had already established S3 as a familiar way to store unstructured and semi-structured data at scale, but companies still needed a comparable experience in private clouds, at the edge, and inside other infrastructure.
The data-gravity playbook
burst workloads
stays close
persistent scale
The distinction matters. Copying vast datasets between environments is slow and expensive. Kapoor's preferred response is to make applications portable, standardize on open formats and industry APIs, and let compute go to the data when the economics demand it. The public cloud remains useful, particularly for burst workloads and quick access to resources. Persistent workloads with constant reads and writes require a more deliberate calculation.
Data has gravity. The useful move is often to free the application, not relocate the planet.Kapoor's infrastructure thesis, paraphrased
MinIO was released as open-source software and spread through developers building S3-compatible applications. That bottom-up route produced recognition and credibility before the company built a full enterprise sales motion. Kapoor describes the eventual model as a bridge: community adoption from below, enterprise engagement from above. Neither side can be treated as decorative. Her advice to open-source founders is emphatic about honesty and transparency with the community.
This is where her finance background becomes more than a biographical surprise. Open source asks a company to hold two truths at once. The free product has to remain genuinely useful, and the commercial offering has to create a reason for organizations to pay. Abuse the first and trust evaporates. Ignore the second and so does the company. Kapoor's early role as CFO placed her directly on that narrow bridge.
One product, several discarded maps
Focus is easy to admire after it works. During the uncertain years it can look like stubbornness in a clean shirt. MinIO chose the horizontal problem of high-performance object storage and resisted becoming a miscellaneous cloud catalogue. Kapoor and her colleagues junked code, made ill-timed investments, and experimented with go-to-market approaches. The clean line visible in hindsight was drawn with an eraser nearby.
The company waited more than four years between its Series A and Series B. Kapoor explained that the intervening job was to build the commercial engine. Only when the team believed that engine could scale did it raise $103 million in the 2022 round, part of $126.3 million in reported capital, at a valuation of $1 billion. The sequence says something about her operating temperament. Capital is an accelerant; first make sure the vehicle has steering.
Early company life also supplied a less financial lesson in consequences. Someone abandoned a dog outside MinIO's office building. The team adopted her. The coworking space had rules against pets and children; complaints followed, and the startup moved into an independent office. Rent in Silicon Valley was not kind to seed-funded companies. Neither, apparently, was everyone prepared to share a corridor with a growing firm, a dog, and family life. The anecdote makes the beginnings feel appropriately improvised. Enterprise infrastructure can emerge from an office dispute with a surprisingly good cast.
Kapoor moved from CFO to COO, then became co-CEO in 2024. The titles trace MinIO's own shift from financing and formation to operational scale and market leadership. She has also invested in and advised emerging technology companies, with public profiles linking her to companies including Isovalent, Procurify, H2O.ai, Helpshift, Treasure Data, and Push Computing. In 2024, Inc. included her in its Female Founders 250.
Everyone watches the engine. She watches the loading dock.
AI gave MinIO's original thesis a fresh urgency. The market's attention has clustered around compute, particularly GPUs. Kapoor points to the system feeding those chips. Training and inference depend on enormous flows of unstructured data. If storage cannot keep pace, expensive compute sits underused. A spectacular engine is little consolation when the fuel arrives by teaspoon.
She sees object storage becoming central to AI infrastructure because it can scale while matching the way modern applications are built. Enterprises, meanwhile, are trying to unify data scattered across older systems, private clouds, public clouds, and new GPU-focused providers. Open formats and standard APIs are not philosophical ornaments in this account. They are negotiating leverage. They keep an application from becoming trapped by the location of its data.
Her 2026 conversations also reveal how she thinks AI changes the company itself. MinIO uses AI tools across engineering, sales, marketing, and product work. Teams build internal dashboards and reconsider which software should be bought and which can now be made. Kapoor evaluates token spending through opportunity cost: what output does it replace, accelerate, or make newly possible?
“The value has shifted back to people, back to your market access and customers.”Garima Kapoor on cheaper code in the AI era
That is a striking conclusion from the co-founder of a software company. If AI compresses the time required to build enterprise products, code alone offers less shelter. Customer relationships, distribution, credibility, and the ability to deliver durable value become the harder assets. Existing software companies, in her view, have a narrow window to reinvest and become AI-first before younger companies attack their markets.
The shift does not soften her financial rules. She still tells startups to watch cash, question every internal process, and keep revenue and spending in a relationship that permits survival. “Cash is king” is hardly embroidered-pillow wisdom, but its plainness is the point. A runway gives a company time to discover that its unfashionable storage product has become the loading dock for a fashionable revolution.
Curiosity before credentials
Kapoor's story is tempting to flatten into a tale about a non-engineer succeeding in technology. The useful part is more specific. She did not treat technical depth as optional. She treated the ability to acquire it as portable. Scientific reasoning from doctoral work, economic thinking about incentives and scarce resources, and operating discipline from finance became tools for examining a new domain.
Her public manner is unusually direct about what was planned and what was not. She wanted to found a company. The precise idea emerged through conversations and observation. MinIO's product found developers through open source. The commercial operation took experiments and mistakes. AI did not validate a perfect prophecy so much as intensify a stubborn problem the company had already chosen.
That leaves a practical founder's playbook. Choose a constraint likely to survive fashion. Give users something useful enough to travel on its own. Be honest with the community that carries it. Build the commercial machinery before feeding it a fortune. Keep the applications light on their feet, the data close to home, and the balance sheet capable of enduring a few wrong turns.
And if a stray dog appears outside the office, understand that every operating plan contains at least one undocumented dependency.