I worked for a solar co in the US for a while, 10yrs ago or so. It was a very interesting time, the hardware (solar panels, inverters) part of our Cost of Goods Sold fell considerably over the course of a few years. I'll suggest a few reasons, as I saw them:
1) Larger-scale, fully-automated manufacturing facilities were coming online. Some mfrs were reputable and high-quality, some not so much. But the abundance of supply caused these vendors to compete for business on price/warranty/etc terms, which drove down our cost to build systems.
2) Lots of developers were also bidding for projects, and utilities forced solar/wind developers to compete to offer lower prices for their power.
3) Government/utility subsidies helped get manufacturing facilities up and running, and made marginal projects more affordable to build/buy. Some incentive programs were well-designed, some were not. The markets that offered the best total return to investors got done first -- the lowest-hanging fruit for solar was in areas with relatively good sun, expensive grid power, and perhaps some extra tax/etc subsidies. Same idea for wind (or any infrastructure or real estate investment, really).
4) Attractive project margins encouraged lots of developers to enter the market, which increased competition and resulted in innovation in business models and processes.
5) Manufacturers scaled up further to satisfy increased demand, resulting in lower hardware costs.
6) Larger investors started angling for a slice of the action. VCs were looking for unicorns, MBA students were looking for jobs. Bankers wanted to make loans and do IPO/M+A work. Big renewable projects/portfolios of >$100M are pretty much the minimum interesting size deal for big project financiers, small deals don't move the needle for them. In fact, if you can put $1B+ to work in a responsible manner, that's even better. Banks are risk averse and renewables were a new/unproven (and thus risky) asset class, and in the early days it was hard to get any sort of financing, at any rate. As more projects demonstrated good ROI, more banks did the work to understand the (hardware/warranty/operational/etc) risks and became willing to lend in this sector. The IRS issued clear guidance on some tax issues which further reduced uncertainty (=risk) for would-be investors. As more financiers entered the market, they competed for deals and the cost of capital for high-quality developers/borrowers fell by several %.
7) Over time, the subsidy programs faded away, but (as intended) during that time hardware costs fell considerably. As developers got more experience, they also got better at reducing ancillary costs (sales, development, legal, etc).
And so it goes, a nice virtuous cycle of competition, some subsidies to help a nascent industry, decent deal economics for investors and early adopters, and manufacturing scale-up.