Why is it so hard to design an open-source smartphone?

Because of all the secrecy in phone hardware.

IIRC the development budget of the iPhone was around 2 billions and half (did a brief search on google and most sites confirm it was around that estimate), rebuilding a somewhat modern mobile phone (say, circa 2015 specs) without using proprietary components from Qualcomm, Samsung, or even a cheapo chinese foundry like Allwinner would be probably even more expensive.

It’s basically a scale problem. Modern smartphones are insanely complex: custom chips, radios, antennas, battery/thermal management, and tight OS-hardware integration. Doing software only is feasible for a small team, but building a full open-source phone from scratch would need hundreds of engineers and billions in investment - which is why almost everyone focuses on OS mods instead.

The correct answer is anti-competition; but not by who you think.

You absolutely can produce an open source smartphone. Pine64 has done it. ubuntu and purism have done it. Nokia was doing it, woot n900. Many chinese brands do it. fairphone kinda?

They can build fully open hardware and we have loads of OS that work just fine.

But why then does no mobility provider at all offer them in their phone offerings. Even if they were on page 34. Tmobile usa, everyone in canada for sure, has AOSP and KaiOS for dumb flip phones. But no open options? If you bring an open phone to them, you'll be imei capped and isolated and separately tracked.

But the answer comes when you also realize ulefone or umidigi are also not offered. It's about who is in control. I highly recommend your next phone is imported from a different region than you are in.

What I see as the challenge for open-source is organization and incentives. It’s hard to find good people who want to volunteer their time to create the less fun and glamorous parts of the puzzle. There needs to be someone directing the talent pool and incentives to do hard and otherwise unrewarding work.

I think this is the same thing that holds back desktop Linux. There aren’t a lot of people willing to do that last 10% to make it really great for the average user. Instead we have a dozen different package managers, that average users simply don’t care about.

I mean think of all the things a smartphone has to do nowadays, there's so much complexity

- The hardware is complex. It requires an immense amount of resources, skill, and money to reverse engineer or design.

- Hardware components are often closed and/or documentation is confidential.

- Manufacturing is hard, manufacturers are busy with bigger customers, and you're at the lowest priority to manufacturers.

> So I'm wondering why can't we have a project that starts from scratch and build the electronics and the OS?

That was kind of what OpenMoko was ... the end result was pretty lackluster IMHO. Starting from a finished, proven hardware design is probably worthwhile to avoid shipping a phone that can't actually connect to real networks, and can't actually make phone calls if it does connect.