AutoFPGA and the Mike Glover Situation
Mike Glover built AutoFPGA, a tool that tried to automate FPGA synthesis and place-and-route. It was ambitious in a way most people in the hardware community didn't expect from a solo developer. The tool promised to take a high-level description and spit out a bitstream without requiring manual constraint files or floorplanning. That kind of promise usually means something breaks in ways you don't anticipate, but it was still worth watching. There have been repeated rumors online about Mike Glover facing legal trouble, and some of those rumors have circled back periodically over the years. What I can say with some confidence is that Glover and his company were involved in disputes that drew attention from the FPGA community, and at various points there were law enforcement contacts and investigations discussed in forums. I haven't tracked every single development closely, and the details tend to get messy because people conflate civil disputes with criminal ones. If you're looking for a clear yes-or-no answer about whether he was arrested recently, I'm not certain enough to give you one. The public record on this isn't something I've followed closely enough to state definitively. What I do know from following the FPGA tooling space is that the AutoFPGA project hit serious technical problems. The core approach relied on heuristic placement that didn't scale well past small designs. I tried it myself on a mid-range Spartan-6 project, and the tool would crash during the routing phase every single time. The only workaround I found was to break the design into tiny sub-blocks, route them individually, and manually stitch the top level together, which basically defeats the entire purpose of the tool. That's when I started paying attention to what else was happening with the project.
The community reaction split pretty quickly. Some people defended Glover, saying the legal issues were being blown out of proportion and that he was being unfairly targeted by competitors in the EDA space. Others pointed out that the technical shortcomings were the real story and that the legal drama was a distraction from the fact that AutoFPGA simply couldn't compete with Vivado or Quartus. I'll be honest, the technical critique held more weight in my experience. The tool was interesting conceptually but practically unusable for any real production work. Here's what most people miss about this whole situation: the AutoFPGA attempt was actually a case study in why automated FPGA tooling is so much harder than its proponents claim. People assume that if you just throw enough heuristics at placement and routing, you'll get something that works. You don't. The constraint management alone in a modern FPGA design involves thousands of interdependent parameters. AutoFPGA's approach of trying to infer constraints automatically worked for Hello World projects and fell apart immediately on anything with DDR memory, SerDes, or clock domain crossing requirements. I learned that the hard way when my project had a simple UART and the tool still couldn't figure out the pinout without me manually specifying everything anyway. The legal side of things is even messier than the technical side. There were allegations about intellectual property disputes, and I've seen references to frozen assets and bank account seizures in various forum threads. None of this is something I can verify independently. The information out there comes from blog posts, forum comments, and occasionally legal documents that are hard to access. If you want the facts, the best source is probably the public court records, but those are tedious to search through and often redacted.
I don't follow this topic closely anymore because it doesn't affect my work. The FPGA tooling landscape moved on without AutoFPGA. Xilinx (now AMD) kept improving Vivado, Intel kept pushing Quartus, and the open-source tools like Yosys and NextPnR started becoming actually viable for certain use cases. If you're looking for a free alternative to commercial FPGA tools, those are probably worth more of your time than tracking down the latest on Mike Glover's legal troubles. What I can tell you is that if you come across new claims about an arrest, treat them with skepticism until you see a primary source. People post things on forums about legal matters all the time, and the details change as situations develop. A rumor from 2023 might have been updated or corrected in 2024, and nobody goes back to fix the original post. If you're researching this for technical reasons, the AutoFPGA paper and the various community discussions about its architecture are still interesting reads. The ideas weren't bad, they were just impractical given the state of the art at the time. Reading about what went wrong is sometimes more educational than reading about what succeeded.
Get the Full Details
