The servers had moved. The financial questions had stayed behind. At LiveRamp, Sasha Kipervarg helped lead the shift to Google Cloud, a migration substantial enough to involve a cluster of 79,800 cores. Making that infrastructure work was an engineering undertaking. Explaining what it cost, and why, created another assignment.
It is a slightly awkward place to begin a founder’s story: after an achievement, with more work to do. But it explains Ternary rather well. Kipervarg and two LiveRamp colleagues, Patrick Raymond and Joshua Kwan, had encountered the gap between running a cloud and managing its economics. In 2020, they started a company to address it.
Kipervarg’s career has kept him close to the machinery. His public arguments keep returning to the people around it. Someone chooses a service. Someone approves a budget. Someone needs to know whether the product makes money. The difficult part is getting those people to recognize their own work in the same set of numbers.
The migration was the easy part
At RampUp in 2020, Kipervarg joined Raymond and Kwan on a migration panel moderated by Google’s Kelsey Hightower. The scale gives the discussion its weight: LiveRamp’s engineers had moved systems with 90 petabytes of disk space and 256 terabytes of RAM. These were working operators discussing a working enterprise, with all the obligations that implies.
The benefits included redundant Google data centers and capabilities for replicating data. Yet a successful technical change could still unsettle the budget. A FinOps account of LiveRamp’s containerization describes costs attracting finance’s attention after the move. Teams needed clearer responsibility, more consistent naming and better labels to allocate spending accurately.
That is an unglamorous list. It also tells you where the work was. A cluster can be functioning perfectly while the organization struggles to determine who is paying for it. The naming conventions become consequential because somebody must connect infrastructure consumption with the teams and products consuming it.
In his 2020 Cloud Native Insights appearance, Kipervarg described the change in purchasing power. Organizations had gone from planning server purchases over quarters to letting developers create resources through a command line. The financial consequences could arrive as quickly as the technical decisions.
His interest extended to unit economics: understanding what an application costs as people use it. That gives a spending discussion somewhere useful to go. Knowing the total is a beginning. Knowing the cost of delivering something to a customer brings the discussion closer to the business.
Before the cloud, a box of disks
Long before Ternary, Kipervarg was working through a more physical version of the cost problem. As an IT Infrastructure Architect at MokaFive, he investigated storage software that would let the company use inexpensive hardware of its own choosing. The requirements included compatibility with VMware and support for the company’s storage architecture.
A favorable blog post helped put Open-E on his radar. He researched the software, its price and its capabilities. The resulting implementation gave MokaFive flexibility in its hardware configuration while reducing costs. There is a pleasingly practical detail here: the person now selling clarity about cloud spending once had to work out which disks and software made sense together.
He subsequently served as Director of DevOps at Workday, then headed global cloud operations at LiveRamp. The titles changed, but the career stayed close to the business of making software infrastructure operate. By the time he became a founder, he had experience of the decisions his future customers would be trying to explain.
There was another kind of preparation. In 2018 he completed the Stanford Executive Program. His public reflection thanked LiveRamp for making room for the experience and his teammates for their support. He described hundreds of hours spent studying and playing with ideas. The note ended with a commitment to make something of it.
A question on the way home
The commercial possibility came into view through another practitioner. Kipervarg later recalled that Lukas Karlsson, of the Broad Institute, asked whether the questions he and his colleagues were pursuing were connected to building something themselves. They had been concentrating on their employer’s problem. Karlsson’s question widened the frame.
On the flight home, Kipervarg discussed the possibility with Raymond. Perhaps, once they had solved the problem in front of them, there was something more to build. It is a modest origin scene for a software company: a question, then a conversation between colleagues. No theatrical revelation required.
Ternary began with Kipervarg as CEO, Raymond leading product and Kwan leading technology. Google Cloud supplied both their initial focus and their own platform. In a later reflection on that choice, Kipervarg described an enduring attachment to the ecosystem even as Ternary expanded support to other clouds.
Early coverage of the launch put the team at ten employees and initial funding at $1.6 million. The product forecast spending, monitored unusual changes and modeled discounts. Partners including SADA, Cloudwerx and Katana1 were part of its route to customers. The founders were turning operating experience into something other organizations could buy.
Give the bill a common language
FinOps provides a framework for the collaboration Kipervarg wanted. The FinOps Foundation organizes a community around managing the value of technology, with education, practices and standards. Its work helps give people in different roles a vocabulary for a problem that crosses their reporting lines.
Kipervarg has carried that conversation into public discussions. A 2023 session with Ken Cheney covered accountability, tagging workloads, identifying untagged resources and making reports useful for planning. These are the small disciplines that allow a large bill to become intelligible. Their usefulness depends on repetition, long after the introductory workshop ends.
For an engineer, a resource label can identify a workload. For finance, the same spending needs a place in a budget. Putting those views together reduces the translation required before anyone can act. Otherwise, a recommendation can sit patiently on a dashboard, impeccably presented and wholly unemployed.
In October 2023, Jump Capital led Ternary’s $12 million Series A. Its investment argument emphasized collaboration across engineering, finance and FinOps. It also highlighted the ability to combine cost sources and give different stakeholders information relevant to their responsibilities.
The announced development plans reached toward customers’ existing systems, including financial ledgers, forecasting tools and ticketing workflows. The practical ambition was to connect an observation about spending with the processes needed to do something about it. Finding an opportunity and following it through are separate jobs; enterprise software must survive both.
A ledger with more doors
By January 2026, Ternary was explicitly reaching beyond native cloud bills. Its Bring Your Own Data announcement described ingesting technology spending and business information together, including SaaS, AI, data-center and revenue data. The underlying approach used FOCUS, the FinOps Open Cost and Usage Specification, alongside support for other data formats.
For Kipervarg, that expansion puts the earlier cost question in a larger setting. A business may deliver one product using several clouds, outside software and its own facilities. Understanding only one provider’s charges leaves an incomplete account of what that product costs to deliver.
On the FinOps Weekly podcast in March 2026, he discussed the discipline’s broader remit. He acknowledged having questioned whether expansion beyond cloud would dilute the mission. Seeing what practitioners were being asked to manage helped change his view. He now emphasized the financial questions executives needed answered about products, margins and future investment.
That willingness to reconsider is more revealing than a prediction delivered with perfect confidence. It places his company’s direction inside an argument he has actually had with himself. The ambition is substantial, but its starting point remains recognizable: people need a better account of the value their technology creates.
His September 2026 writing adds another complication. AI can make enterprise information easier to ask about. Financial decisions still require dependable underlying data and calculations that can be traced. Ternary’s current offering advertises access to its ledger through Model Context Protocol, allowing an AI assistant to query the system.
The interface can become conversational. The number still has to carry its meaning with it. Which costs belong to a product? Which period does the answer cover? Kipervarg’s recent emphasis puts the groundwork beneath those questions ahead of the convenience of asking them aloud.
Leave room to change your mind
His approach to leadership makes room for similar uncertainty. He values changing course when new information warrants it and has spoken about learning the importance of advisors. In a public testimonial for coach Sebastien Dumont, he praised questions that helped him work out answers himself.
“Success relies on relationships and trust.”
Sasha Kipervarg
In Pacifica, California, the interests extend beyond infrastructure. He makes Neapolitan pizza and explores modular synthesis, including older analog instruments. He has also described taking up snowboarding. These are useful details because an executive biography can otherwise make a life sound like one continuous meeting.
Kipervarg’s professional story has a more persistent question running through it. When people gain a new way to build, how do they understand what they are spending? He has approached it through storage, cloud operations and now a company of his own. The migration ends. The need to make sense of the bill keeps finding new forms.