How MilkStraw AI is redefining cloud optimisation

Biography

Jawad Shreim is the CEO and co-founder of MilkStraw AI, a UAE-based cloud optimisation startup. This week, MilkStraw AI raised a $600k pre-seed round led by Flat6Labs, with participation from Angel Spark, Beyond Capital, and other angel investors.

MilkStraw AI optimises AWS cloud costs by leveraging long-term contracts. It provides clear insights into cost sources and offers actionable recommendations for further optimisation.

What inspired you to launch MilkStraw AI, and how did your background influence the company’s focus on cloud cost optimisation?

Before starting MilkStraw, I was working for a hedge fund for a few years, primarily focused on trading stocks, analysis, and risk management. During that time, a lot of models were being trained. While I wasn’t directly involved in monitoring cloud costs, I noticed how challenging it was for the team. Cloud expenses were extremely unpredictable. You couldn't accurately estimate costs, especially when training models. It could spike to $80,000 and then drop back to $20,000.

This unpredictability was a major pain point. Towards the end of my time at the hedge fund, I kept thinking about how interesting it would be if this problem could be solved differently. I didn’t want to approach it the same way as others. Typically, companies assign a few engineers to work on optimisation for months, and while they might reduce costs by 15%, I felt it could be done better.

So, I started studying cloud infrastructure more deeply. Around that time, I met my current co-founder, Anas Abdullah, who is a cloud expert. We began discussing the issue and exploring ways to improve cost management. We realised that if a solution required heavy engineering work, it would naturally take a long time. So, we shifted our focus to understanding cloud billing instead.

Cloud providers generally offer two payment models: a pay-as-you-go model, where you pay by the hour, or a commitment model, where you commit to a one- to three-year term in exchange for a significant discount. The price difference is substantial—around 50 to 75% savings. Despite this, neither Anas nor I had ever made such commitments before, and we wondered why.

We dug deeper and realised the reason: risk. Committing to a specific server for up to three years can feel like a gamble. If the server isn’t fully utilised, who bears the cost? Most companies avoid long-term commitments unless they’re certain of their resource needs.

Identifying this gap, we saw an opportunity. Cloud providers don’t care who pays them, as long as they’re paid. So, we thought—what if we committed to the contracts ourselves? We approached a few friends and said, "You can make a reservation, and if you don't need it, we'll take over the commitment."

The premise was simple, and it worked. AWS, Google Cloud, and Azure all have the same requirement: someone needs to pay. We started buying and moving contracts around based on actual usage. After a few months, we decided to stop the trial as a test. The reaction was eye-opening—customers had saved so much money during the trial that they didn’t want us to stop.

That’s when we realised there was a real need for this service. Customers genuinely didn’t want us to leave because of the value we provided. And that's how we knew we had something worth building with Milk Straw.

Cloud cost optimisation is a competitive space. What convinced you there was still room for a new player like MilkStraw AI? Was it driven by a regional gap or a broader market need?

The premise isn't just about cloud services but B2B in general. Most B2B software today is built to be functional but boring. I believe it's essential to remember that you're still interacting with humans, so a great user experience and interface should be a priority. Unfortunately, many B2B platforms have a similar, uninspired design and offer poor support, making the entire experience frustrating.

When we looked at cloud optimisation tools, we saw the same pattern—clunky, difficult to work with, and challenging to get support for. So, the question for us wasn't just whether there was space in the market but if we could stand out completely. Our first goal was to create a beautiful UI/UX while delivering impactful, fast optimisation results. Today, the average customer completes their optimisation in just 10 minutes. From onboarding to leaving the website, the experience is so smooth that customers often tell us it feels too good to be true.

That's our focus—providing such a great experience that customers wonder why they'd switch to another, more complicated solution. We're part of a broader trend where smaller startups like MilkStraw are disrupting traditional industries by prioritising usability and efficiency. Whether it's in email, task management, or cloud services, companies like ours are redefining expectations by building tools for modern, fast-growing startups rather than older, traditional businesses. We're creating something that's both enjoyable and impactful.

Many cloud optimisation tools claim to use AI but often rely on basic automation rather than true machine learning. Can you clarify the specific role AI plays in your platform and how advanced the automation really is?

From a customer's point of view, the entire process takes minutes, and they don't have to worry about anything else afterwards. Initially, we handled optimisation manually, but as we scaled past managing thousands of contracts, we needed a more sophisticated approach. That's where my finance background came in. We began treating server reservations like a hedge fund portfolio, monitoring their utilisation and trading contracts in real-time based on demand.

Predictive models now play a key role. Rather than waiting for underutilised resources to become a problem, our AI predicts usage patterns up to a month in advance. If we see a server is likely to be underutilised, we proactively match it with another customer in need, ensuring efficiency across the board.

How does MilkStraw AI balance automation with control, ensuring businesses maintain visibility and decision-making power while costs are being optimised in the background?

Yes, so another critical pain point we noticed was visibility. Many customers struggle to understand their own cloud costs. We're integrating LLM technology to allow them to ask simple questions like, 'How much did my databases cost today?' or 'Can you predict my end-of-month expenses?' This level of transparency empowers them to make better decisions beyond our recommendations.

Ultimately, we're working towards making cloud cost management intuitive and fully integrated. Whether it's daily summaries, reports, or a native MacOS app, we want to provide customers with complete control and clarity over their infrastructure while keeping the experience seamless and effective.

On the business model side of things, is it subscription-based, or how exactly does it work from the customer's perspective?

On the business model side, we believe our best selling point is our performance-based billing. We're so confident in our product that if we save you money, we make money. If you come in and use our services, such as monitoring, you don't have to pay us anything upfront. However, when we successfully save you money, we take a percentage of those savings as our fee. This model ensures alignment between our success and the customer's success.

Congrats again on securing your pre-Seed funding! What made Flat6Labs and the participating investors the right strategic partners for this round?

This was my first time raising in the region. I moved from Silicon Valley two years ago, and my perception of pre-seed funding was based on that context—an idea, a strong background, and a small ticket to get started. However, I quickly learned that pre-seed here often requires a fully built product.

By luck, I met Mohanad Elhartany while he was working at another fund. He was incredibly supportive and promised to help us secure funding. Though his fund didn’t move forward, he later joined Flat6Labs Abu Dhabi and reached out to encourage us to apply. We had no product or dashboard yet, just the core idea, but they believed in us.

Mohanad was instrumental – he championed our vision and remains one of my closest friends and supporters. After our pitch, they made a decision on the spot to invest, which is quite rare in this region. Since then, we've continued building and expanding, supported by a fantastic team of investors, including some of our own customers who were so impressed they decided to invest. Their confidence has been a huge driver for our growth.

Since launching in April 2024, MilkStraw AI has already secured prominent clients like Magnitt, Voltlines, and Floranow. What has driven such rapid adoption?

This is one thing I’m quite proud of. We've done zero marketing to date—everything has been through network effects. One customer brings in another, and referrals have been our strongest driver. A great example is Tapper AI. I met their founder the same day we got funded, and we became friends. He loved what we were doing and later mentioned us to the CTO of Magnitt. She checked us out, and after a bit of back and forth, she became a customer.

It’s been this organic network effect. People hear about us, try the product, and the experience speaks for itself. Referrals have become our biggest growth engine because people genuinely love the product and want to share it. We’re prioritising the product experience over aggressive outreach, and so far, it’s been working extremely well.

Cloud infrastructure optimisation often requires deep integration with platforms like AWS, Azure, and GCP. How do you balance neutrality while partnering with these major providers, especially when they offer their own cost optimisation tools?

Absolutely, we do plan to expand but we're pacing ourselves carefully. Originally our goal was to expand to Google Cloud in the first quarter of next year and Azure in the following quarter. 

We realised that even if we focused only on AWS, we have years of work ahead to fully optimise the experience. AWS remains our primary focus because it still holds 33% of the global market. So rather than spreading thin across multiple platforms, our strategy is to perfect the AWS experience first—delivering a seamless, optimizsd product—and then expand.

Customer feedback has also reinforced this approach. Most of our clients are on AWS, and very few request Google Cloud or Azure support. We want to ensure we deliver the best possible product on one platform before scaling to others. It's about prioritising quality and user experience over speed of expansion.

If a competitor with similar technology but deeper resources emerged tomorrow, what would make MilkStraw AI the stronger choice for businesses?

I would actually welcome it. It would further validate the work we’re doing. Beyond customer validation, a competitor's presence would highlight our unique strengths: our focus on creating a beautifully crafted product, our incredibly talented team, and the strong network effect we've already established. Our team’s proactive initiative and commitment to delivering a top-tier experience is something I emphasise to investors—our people are our strongest asset. Additionally, our network of thousands of managed contracts gives us a substantial head start. Simply having more resources doesn't equate to delivering a better product. Our attention to detail, from UI/UX design to customer interaction, sets us apart. We genuinely care about delivering exceptional value, and our customers can feel that connection when working with us.