A cloud invoice is an architecture diagram written in dollars. Every line records a choice: an instance left running, a database sized for tomorrow, a backup kept for safety, a container multiplied by demand. Finance sees a number. Engineers see a system. For much of Asim Razzaq's career, those two views arrived in different rooms.
Razzaq had lived inside the machinery. At PayPal, his platform teams carried responsibility for core infrastructure behind logins and payments. He helped build the company's private cloud and developer platform, work that Yotascale's biography says supported multi-billion-dollar payments volume. He learned that scale makes small decisions loud. A reasonable technical choice, repeated across thousands of resources, becomes an operating expense with its own gravity.
The frustration was not that companies lacked data. Cloud providers generated plenty of it. The problem was translation. Bills were organized around accounts, services, and resource identifiers, while companies thought in products, teams, customers, and experiments. Untagged resources blurred ownership. Recommendations could look absurd when they ignored why a workload existed. A tool might suggest switching something off without understanding that the supposedly idle machine protected critical infrastructure.
Our ethos has been empowering engineering teams, because cloud computing can be expensive.Asim Razzaq
Before the cloud, there was Lahore
Razzaq grew up in Lahore, Pakistan, in a home where numbers were ordinary conversation. His father was a chartered accountant. Math came first, then physics. Computer science arrived in college and rearranged the plan. He has recalled simply falling in love with it.
His path to the United States followed a route visible among graduates of his competitive high school. Some went to Britain, some to America. Razzaq already had family in the United States and chose the University of Texas at Austin. He earned a bachelor's degree with honors in computer sciences and conducted undergraduate research in resource management and distributed computing. The subject now reads like an early clue: even before cloud computing became the commercial default, he was studying how shared systems allocate finite resources.
After university, he joined a startup as a lead architect and met the economics of data centers up close. He later moved through PayPal, eBay, and backup-software company Axcient. At eBay he directed engineering for the X.commerce platform. At Axcient he served as vice president of engineering. Then he returned to PayPal in 2013 as senior director of platform services engineering.
This was not a founder waiting for a problem that fit a fundraising story. It was an operator repeatedly running into the same knot: performance, availability, and cost could not be managed independently, yet the people closest to the architecture rarely received financial information in a form they trusted.
Conviction, assembled slowly
In 2015, Razzaq left the relative safety of an established engineering career. But he did not immediately turn his experience into a glossy certainty. He went out and spoke with data and infrastructure engineers at large companies. The question beneath those conversations was brutally useful: is this a market, or is it merely my problem?
Company after company described the same confusion. They liked the flexibility of the public cloud, but could not clearly explain where their money went. That repetition gave Razzaq the confidence to proceed. Yotascale was initially self-funded. Capital came after the interviews, then product work, then signs that customers were pulling the product into their operations.
The sequence distinguished the company. Yotascale raised $3.6 million in 2017 from Engineering Capital, Pelion Venture Partners, and angel investors. A Series A followed in 2018. In 2020, Felicis Ventures led a $13 million Series B, with Engineering Capital, Pelion, and Crosslink participating. At that point, total reported funding stood at $25 million.
Razzaq told TechCrunch he was not interested in collecting an oversized round for its own sake. He wanted enough money to move the company forward. The logic matched his product: spend should be connected to a purpose, not treated as an achievement by itself.
Razzaq wanted a name derived from yottabyte, a unit of digital storage. The domain was too expensive, so he dropped one T. The compromise became Yotascale.
Put the number beside the decision
Yotascale's thesis was organizational as much as technical. Cloud cost had often been handed to finance as a reporting problem. Razzaq treated it as a feedback problem. Engineers decide how applications are built and operated. If they can see cost alongside performance and availability, they can make a fuller trade-off while action is still possible.
That requires more than a lower total at the bottom of a dashboard. Costs need to be allocated according to the business: a product, feature, team, namespace, or customer. Shared resources need rules. An anomaly needs context. A forecast needs enough history to mean something. The useful unit is not simply a dollar. It is a dollar connected to an owner and a reason.
A company should not chew off more than it can digest. Start small, learn, and then scale up when ready.Asim Razzaq
The advice came during a discussion about managing AI and large-language-model costs, but it carries the shape of Razzaq's own company-building process. Experiment on a contained slice. Understand the architecture. Expand after the economics and behavior are visible. It is the same method, whether the system is a model deployment or a startup.
The engineer survives the title
Razzaq likes to call himself a "recovering engineer." The joke works because the recovery never quite happens. In 2022, he launched a short-form podcast called Always an Engineer. Its premise was that technical habits persist even when an engineer becomes a CEO. Early episodes covered the damage done by hero culture, how to motivate engineers, why estimation goes wrong, and why meetings irritate software teams.
The topics reveal a broader interest in systems made of people. Moving into management had taught him that individual technical strength does not automatically translate into leadership. A manager's output comes through other people. Delegation, context, and trust matter more than personally solving every difficult problem.
Former colleagues' public recommendations describe him as metrics-driven, clear, approachable, and attentive to mentoring. Those qualities sit together more comfortably than they may appear. Metrics tell a team what is happening. Approachability helps the team explain why. Neither works alone.
The modern smokestack
There is another consequence to idle infrastructure. Servers occupy buildings, draw electricity, and require cooling. Razzaq has called digital infrastructure "the modern day smokestack," an image that gives physical weight to something marketed as an airy cloud.
For him, the economic and environmental arguments meet at waste. A resource that consumes power without producing useful work is expensive in both senses. Yotascale's stated vision expanded accordingly: make cloud computing economically and environmentally viable for every organization. The mechanism remained grounded in daily practice. Show teams where resources go, help them understand what can change, and keep performance and availability in the same conversation.
In later product work, Yotascale applied generative AI to that interface. Yota Copilot let users ask natural-language questions about spend, trends, forecasts, and optimization opportunities. The technology changed; the job stayed consistent. Translate a complicated system into context someone can act on.
In 2025, Razzaq shared a ClickUp case study reporting a 20 percent reduction in cloud costs. He also said the company's engineering manager cut time spent on cloud-cost management from ten hours a week to twenty minutes after adopting Yota Copilot. The more interesting number may be the time returned. Saving money matters. Giving engineers back the hours to build is a second dividend.
A company becomes a chapter
Yotascale's live website now carries a short message: the Yotascale Platform is part of CoreStack, and customers are being welcomed to CoreStack's FinOps offering. It is a spare ending for a company story that began with long conversations about where cloud money went.
What remains distinctive is Razzaq's method. He followed the problem across jobs. He checked that other people felt it. He bootstrapped until conviction justified capital. He framed the buyer's spreadsheet as the user's engineering system. Even the company name records a practical constraint and an adjustment.
Razzaq has argued that entrepreneurship should be taught in elementary school, and that society should be more candid about both its value and its failure rate. His own lesson is less theatrical. Risk can be calculated without becoming timid. Ambition can coexist with evidence. A founder does not need to know everything at the start. He needs to ask a question precise enough that the market can answer.