Strong Compute raises $10.9m to help build the future of cloud computing

Machine learning tech start-up Strong Compute has raised just under $11 million (US$7.8 million) in seed funding for its purpose-built cloud computing AI business, which founder Ben Sand hopes will uncover ‘limitless’ opportunities in an industry estimated to be worth $2 trillion by 2030.

The round received support from a total of 30 funds and angel investors, including the likes of Sequoia India, Folklore Ventures, Blackbird and Skip Capital, as well as Y Combinator, Starburst Ventures and founders and engineers from companies like Cruise, Waymo, Open AI, SpaceX and Virgin Galactic.

Sand acknowledges that looking to build the future of cloud computing is “intellectually meaty stuff” for his Sydney-based team but believes they are motivated by being able to see the impact they’re making in real-time.

“I think everyone has the sense that AI is pretty significant, but people perhaps aren't quite aware of just how significant it will be,” Sand told Business News Australia.

“We certainly see it transforming absolutely every industry in a way that will make pretty much anything we deal with today almost unrecognisable in a short space of time.”

“When you look at a company like NVIDIA, they've gone from being a $10 billion company to somewhere between a half a trillion and a trillion-dollar company in less than 10 years, and that's because their chips, which were originally used for gaming, are perfect for these matrix multiplications,” he added.

The entrepreneur says that typically AI today refers to what's called neural networks, which is software that is essentially trying to simulate parts of the brain. At the core, the way this is done mathematically is with what's known as matrix multiplication, which he describes as a fancy way of saying a spreadsheet multiplied by a spreadsheet.

“When people are developing AI, what they want to do is train their AI, which means you take a large pile of data - in our case, we mostly work with images - and then you feed it through these spreadsheets (neural networks),” he said.  

“The idea is that the neural network at the end of the day will be reasonably good, or at least good enough, at guessing the content or some characteristic of the image.”

One example he mentions is self-driving cars, where a robot is fed images of people enough times that they are confident the technology can correctly identify a person to deem it safe. Another example is an MRI scan, where the AI would need to identify if there was cancer present at least as many times as a radiologist looking at the same information.   

The process of training and devising experiments to test specific kinds of neural network designs typically takes a lot of time. Strong Compute is looking to solve this problem by working on developing these networks faster.

Sand is comforted that he and his team are on the right path, as he notes that the investors who have come in to support the business are well-connected to the market and are experiencing the same issues.

“We’ve gone through rounds of technical grilling from people who've been at Open AI or worked at the leading self-driving companies, and for them to go, ‘We've seen this at the highest level, and we think you've got something here, it’s quite rewarding to have that validation,” the Strong Compute founder said.

“It feels like the industry's properly taking off here in many ways; I think the unicorn counts in Australia are up significantly over the last 12 months.

“It just means that we've got people here with real experience that have helped build that, and they're coming on the journey with us, and we're pleased about that too.”

The seed funding means that Sand can hire more deep learning engineers, doubling down on AI in Australia, and it also opens up some longer-range research and development projects.

“We had initially gone out with a smaller raise target, but the interest was enough that we can bring forward some of the bigger projects that we're looking to do, which we're excited about as well,” he said.

“That will be some really hardcore math stuff that doesn't necessarily result in a normal two-week sprint cycle but can give some serious value over the longer term.”

Folklore partner Tanisha Banaszczyk believes the industry has only just scratched the surface regarding what machine learning and AI can do.

“We love working with founders who have long-range ambition and visions that will endure across generations. Having invested in autonomous driving, we know how important speed to market is – and see the impact Strong Compute can have on this market with its purpose-built platform, deep understanding of the $500 billion market and world-class team,” she said.

Sequoia India principal Anandamoy Roychowdhary believes machine learning will unlock a new level of productivity in multiple $100 billion industries via the deployment of larger and larger models with significant complexity.

“We were very impressed with the work that Strong Compute is doing to dramatically accelerate both performance and productivity of ML (machine learning) teams everywhere, and Sequoia Capital India is excited to partner with them to help bring this to the world,” he said.

For Sand, who previously co-founded augmented reality hardware and software Meta (not the same company as the recently renamed Facebook parent), the raise endorses his decision to bring a lot of the working culture and processes he picked up during his time in Silicon Valley back to Australia.

Having been previously attracted to moonshot companies, he is excited about creating an “epic” company that might one day compete with the big players in the industry like Amazon, Tesla and Google, and isn’t weighed down by the pressure that brings.

“I've always found that if you're not working on the biggest thing you can imagine doing, you tend to spend most of your time wishing you were. I like to take on the biggest challenge that I can find and work through it and then upgrade to the next size,” he said.  

“You don't want to bite off more than you can chew, but if you can't see an unlimited long range on the paths you’re on, you have to wonder, why bother?”