For most software companies, COGS is dominated by cloud costs. Across the market, these can account for anywhere between 10 and 20 percent of a company’s revenue. In compute-heavy applications, such as AI and machine learning, this percentage can be significantly higher.

And as every company becomes a software company, getting control of these costs is vital.

In the days of on-prem, provisioning new infrastructure was slow and deliberate — you’d have to plan out your needs, order new hardware, and install it in your data centre. But at least your costs would be fixed.

Now, you may think that you have total control and transparency (“this compute unit is charged at $0.02”). But choosing the right infrastructure isn’t straightforward, doing it manually isn’t scalable or optimal, and linking spend back to a particular app or product is far from clear-cut.

Moreover, as your computers become invisible, so too do your costs. It might be frictionless and instantaneous to add more compute power, yet true control means not just having clarity on costs, but combining that with effective utilisation.

Meanwhile, multicloud continues to grow, bolstered by workloads that are best suited to different companies’ requirements, and because diversifying mitigates business risk and improves operational resilience. Performance optimisation — that “effective utilisation” piece of the puzzle — thus demands a neutral third party — one that is agnostic to the underlying infrastructure and works in different environments.

Enter Granulate.

Granulate’s software agent can boost performance and slash cost of cloud deployments by up to 60 percent. It studies application-specific resource consumption patterns to infer continuous and real-time optimisations at operating-system level.

In simple terms: it looks at how applications are using an operating system, and then optimises the operating system so those applications can run more efficiently — which includes things like improved memory allocation, thread scheduling or CPU optimisation.

Think of the leap from just dealing with your energy bill, to setting up a smart system in your house that allocates energy to specific “events” on an extremely granular level — say, automatically generating heat and light in specific rooms as you move through the house, or only turning on appliances when you use them, thereby driving the overall efficiency of your home.

What we find particularly exciting about Granulate is that the performance gains it enables its users to achieve stand completely orthogonal to other optimization techniques they might have in place. And crucially, customers can deploy their solution with zero code changes and across all environments.

This truly is a new frontier in resource optimization. It goes far beyond what even the best DevOps person in the world could do manually, continuously making a vast number of optimisations in real-time.

Of course, impressive and world-changing software is invariably matched by impressive and world-changing founders. Asaf Ezra and Tal Saiag, CEO and CTO respectively, are exactly this. Israeli defence forces veterans, they founded Granulate just two years ago, in 2018, having identified a huge problem with a vast addressable market.

This is reinforced by the strong initial customer adoption the team has achieved, with Mobileye, Appsflyer, Rappi and Hike on the list, a strong foothold in the US and off-the-charts net revenue retention. And when it comes to that team, Asaf and Tal have brought together some of the finest AI and performance engineering talent in the world.

In this virtual world, we still haven’t met any of them! But even across Zoom, their drive, tenacity, wit and passion is boundless.

Dawn has worked with several founders and teams that embody the standard associated with Israeli tech startups: inimitable products and approaches. We first invested in Minute Media in 2014, and Access Fintech last year, too.

Cloud has won. CapitalOne’s November announcement that it’s moving fully to cloud was just the latest nail in the coffin for on-premise data centres. This is now how enterprises build, operate and grow. The category-defining companies of tomorrow are those that can make cloud bricks and mortar more efficient, and at scale — in other words, the businesses that transform the infrastructure upon which we all rely.

If infrastructure 1.0 was buying and managing on-prem and racking servers, and infrastructure 2.0 was moving to public cloud platforms like AWS, Azure and GCP, then infrastructure 3.0 is all about managing those platforms — introducing robust governance, optimising performance, controlling costs, working seamlessly with different cloud platforms, and operating efficiently at scale.

Granulate’s team has recognised this vast and accelerating trend, and we are delighted to be supporting them as they continue as the vanguard of this huge opportunity. If you want to join this fantastic team, they are hiring!

Follow Evgenia on LinkedIn and Twitter