Tech companies are incredibly profitable. Fixed costs are really low. Employees are probably the biggest cost. This is in stark contrast to the likes of, say, C

Tech companies are incredibly profitable. Fixed costs are really low. Employees are probably the biggest cost. This is in stark contrast to the likes of, say, Comcast, Spectrum or Verizon that have massive capex budgets. Building a broadband network is an interesting business. I like to describe it as a hyper-local national business. I saw this because every municipality has different rules (eg compare getting pole access in Louisville vs Nashville vs Austin), the ground is different (one area might have a lot of limestone in the ground and make trenching slow, difficult and expensive) and so on. Getting something resolved might be a case of whoever is the construction manager in a given area knowing the right person to call that might otherwise have 11 people standing around for 2 days. This is something that national ISPs are exceedingly good at because they've spent decades doing it. I mean they go further than that by trying to keep competitors out through regulatory blocks and so forth but this doesn't change the fact that this is a core competency. It is not nor was it ever Google's core competency. Google is at its heart a technical company that solves technical problems. Making search or Maps or GMail worth with petabytes or even exabytes of data is something Google is exceedingly good at. This is an entirely different business model and core competency. I don't know this for a fact but I SUSPECT that Google's executives went into Fiber with the naive belief that (a lot of hand-waving here) they were Google and could do anything and somehow they could reduce the cost per household connected from a more normal $1000-3000 to <$500. This might be true if it was a software problem. Digging up trenches and stringing cables to poles is not a software problem. What's more Google has never really had to control costs in the same sense that a large capital project would have to. Worse, Google's culture that it'll ship when it ships (which is great from a software engineering POV) doesn't work when you need to plan for lighting up networks and areas and there are a lot of moving parts to make that happen. It's really a case of what gets measured gets done. Add to this that there is uncertainty over what the future of fixed broadband is going to be. Will it be wireless or wired? That's the ultimate question. If it's wireless and you've just spent $50B building a national wired network then... woops. If it's wired and you've just invested $25B in building wireless infrastructure then... woops. I suspect it's one of these cases where doing nothing is worse than doing something but I suspect Google's board was scared off by the large capital costs involved in doing anything and were too risk-averse to be wrong. Disclaimer: Ex-Fiber Xoogler.