As someone trying to raise funding for a semiconductor startup in "silicon" valley, the biggest problem I see is lack of information. Most investors think that

As someone trying to raise funding for a semiconductor startup in "silicon" valley, the biggest problem I see is lack of information. Most investors think that the minimum amount needed for a potentially (still very low chance) successful Semi startup is $100 million, which is nonsense carried over from the last century. I recently talked with a name partner at a very large (universally known) VC in the valley, and they thought that for a 5-10% chance at success, a fabless semi startup needed $100 million minimum, and to increase that to 30% you needed $250 million and hundreds of engineers. The only reason these ideas are prevalent throughout the venture world is because the market incumbents (Intel, AMD, NVIDIA, etc) want the rest of the world to think that they are the only viable companies so that new competitors do not form. Over the past 10 years the semiconductor IP (and standard cell) ecosystem has made huge progress, saving significant time and money on the design and implementation of parts of a chip that are not the startup's "secret sauce"... FPGAs and their tools have only gotten better, allowing anyone to test a new chip/component design with little more than a laptop and a relatively inexpensive FPGA. The fact that the foundries of the world have been able to lower the cost of multiproject wafer fabrication runs, making test of 50-100 chips of real silicon cost as little as $500,000 using a ~4 year old process to $2,500,000 for the latest (20nm) process. The fact that these fabs want to be producing at capacity means that the cost for getting to a full fabrication run (tens to hundreds of thousands of chips) can cost as little as $5,000,000 and goes up to about $15,000,000 depending on the fab and the process. Realistically, a small startup of ~3 people should be able to prove something on an FPGA (and show good simulations for a produced chip) for ~$100,000-$250,000. ~5-10 people and $2 to $5 million should be able to produce actual chips, and depending on the process node, be able to produce a couple thousand chips and have revenue. No modern fabless semi startup that should exist that has the right team should need more than $15 million to be able to get to revenue and hopefully profitability. I've always believed and thought that it was a silicon valley idea that innovation happens with small teams with what are thought of as crazy ideas... and if they aren't able to do it with 1/10th of the capital and 1/10th of the number of people, and the potential to make something 10x better, then it is not a worthy startup.