asking for 10m$ upfront to develop custom asic is considered 'the proper way' ? after asking pre-payments for hardware that was way too late too market ... think not.
Yeah, I think for Sumo, it's somewhat too late... But overall, I think this is not a bad concept. Granted, I am not a fan of the idea to keep the Algo and plans secret at the beginning, but I think having a specialized algo with specialized hardware right from the beginning is in theory the ideal setup. In practice, you'll obviously have a whole bunch of problems on the production side, with companies being more effective than other and all that stuff.