Amiram Shachar

Running Military Data Centers in Israel

Like many Israeli founders, he started in the army. He told Intel Capital: I was lucky to be in the unit responsible for the military's cloud infrastructure. It was a really fun role at the age of 18 to be installing and virtualizing servers, operating datacenters and modernizing applications. The ModernCTO podcast credits him with introducing VMware virtualization in the Israel Defense Forces for the first time. After the army he worked days as a director of DevOps, moving a company's data center to the cloud, and studied computer science at night.

A Final Project That Became Spot

His last-semester project became a company. On the ModernCTO podcast he described himself as a lazy student, trying to get the highest grade possible in my final project, running a project about something that the dean has no idea about what I'm doing. So I chose cloud infrastructure, and then it turned out to be a company. The idea came from his army work: In the military I was always trying to find ways to save costs on energy and server space in the data center, so why not do it in the public cloud? This was the genesis of Spot. He founded Spot, first called Spotinst, in Tel Aviv, Israel, in 2015. It used automation and machine learning to run workloads on the cheapest available cloud capacity.

Investors doubted it. He told the Founder StorieZ newsletter that venture capitalists warned the cloud providers would shut him down. So on one hand, like very smart people, VCs tell me my business is really bad. And on the other hand, you have customers with pains. So who do you listen to?

Moving to San Francisco

In early 2016 his board told him to wait until Spot had 100 customers before moving to the United States. He did not wait. So on that night, I actually booked my flight ticket to go to the US. Everyone said no, and I said, no, that's the right thing to do right now. His wife quit her job at another Israeli tech company to come with him. We just took two suitcases and came to San Francisco, he said, meaning San Francisco, California. NetApp bought Spot in 2020, for about $450 million according to TechCrunch, and more than 1,500 organizations were using it by early 2021.

Founding Upwind After NetApp

Upwind grew out of what happened after the sale. NetApp's security team scanned the cloud systems his team had built and sent long lists of findings. He told TechCrunch in 2026: The security team would scan our environment and report issues, but they lacked critical context. His team could explain every finding because they ran the systems. He told Salesforce Ventures: We actually had the opportunity to build a security company but with a DevOps mindset.

He has written that the company began on a family trip. Omri Casspi and I were on a family trip to San Diego when I told him about the idea to start a company. That weekend Casspi introduced him to Gili Raanan of the venture firm Cyberstarts, who took 45 minutes to make a decision. The same weekend he met Saam Motamedi of Greylock and other early backers, and Bessemer Venture Partners later backed the Series B. The early sales calls were humbling. Our first 100 or 200 calls were not customers who bought from us, he said. Customers told the team they wanted several kinds of cloud protection in one product, so Upwind built broadly even when advisers urged a narrow start.

He is candid about the doubt. Three years ago, we would spend hours asking ourselves if we were heading in the right direction, and 80% of the time, it felt like we weren’t, he told TechCrunch, after a $250 million Series B in 2026 that valued Upwind at $1.5 billion.

Watching Cloud Systems From the Inside

Upwind's approach, which it calls runtime security, uses signals from inside running systems, such as network requests and API traffic, to decide which risks need attention now. Many cloud security tools scan from the outside without installing software. He argues that is not enough for containers, short-lived serverless code and AI agents that call each other: Inside-out isn’t an advanced option; it’s the only way to solve the next generation of problems.