When we started ArrayFire, we had 4 other startups in our market that provided role model information for us:
- Peakstream – they had the most similar mission to ArrayFire but had entered and exited the market before we even started, namely they sought to make parallel computing easier through a library with a simple vectorized interface (acquired by Google in 2006)
- Rapidmind – they also sought to make parallel computing easy by providing a software tool that could optimize all the parallel computing components in a system, including multi-core, GPU, FPGA, and other components (acquired by Intel in 2009)
- Interactive Supercomputing – they paved the way for using parallel computing across clusters of CPU resources through simple programming constructs (acquired by Microsoft in 2009)
- Cilk – they provided an extension to C/C++ that enables utilization of parallel computing resources (acquired by Intel in 2009)
In each of those cases, the companies had raised $10+ million. To my knowledge, none of those exits provided the founders of those companies with a significant financial windfall. My understanding is that while they were all 8-figure exits, they did little more than return the VCs their original investment capital. I may be wrong about that and would love to be corrected, but that is my understanding.
Due to that information, it was straightforward for us to make the decision to remain bootstrapped. Our burn and development resources have been less than those companies due to the absence of VC money. But now after 7 years, it is my understanding that we have more revenue and traction that any of those 4 predecessors. That is an exciting milestone for me looking back at those times when I viewed those 4 startups as huge role models for our business.
I’m always excited about the annual Supercomputing Conference in November. This year I’m even more excited because there is a meetup about HPC startups. HPC and technical computing is an exciting market and is ripe for disruption in so many ways. I look forward to the future!
What are your thoughts on HPC startup role models?