The panel shared many insights into starting HPC companies. Here are some of the themes I repeated for this audience:
- The history of raising venture capital for this market, especially for middleware software tools companies, is not good. We are successful merely because we bootstrapped our way to success. If we would have raised VC, it would not have worked out.
- Anyone contemplating leaving a BigCo to do a startup needs to realize there are only 3 things required for success: competencen, working on solving an important problem that people care about (which often requires pivots over time), and persistence. With those 3 ingredients, anyone can become successful in time (with a lot of persistence).
- If your first target customers are in HPC (like the national labs), you are going to struggle. The main players in HPC are actually very slow to adopt anything from small companies. It is a terribly conservative market with long sales cycles.
- The future of scientific and technical computing is going to be in a browser. Installed software is dying.
What other advice would you give to the startup HPC community?