You can't build a product company like this. Products require lots of development before any revenue arrives. Underdeveloped capital markets (VCs, stock exchanges etc.) is why Europe's software industry is mostly services. The difference between SV and anywhere else is access to risk-tolerant money. SV is not primarily an IT cluster, it's a VC cluster.
Sure you can, you just bootstrap through monetising the ground work. For example, if you want a product company you can start out as a freelance design agency. The bulk of the initial design work will be for clients, but you can be involved with low volume manufacturing on the side. As the company grows, not only do you have the funds to take on progressively more ambitious product manufacturing projects, you've also built up a team of designers that can help push these projects forward. You're going to need designers right? Why not get them making you money from day one.
To give an alternative example, let's say you want to manufacture mobile phones. A good way to start small would be to manufacture accessories for existing phones. This is within the grasp of anyone within the Western world, given enough drive to do it. The revenue and knowledge you can gain doing so allows for increasingly sophisticated accessories to be built until the point where you have your own manufacturing premises (or have access to one at low rates, due to the volume of products you're pushing). The step from this point to manufacturing mobile phones is not that great.
Does this require more patience than burning through VC money? Sure, but it's also lower risk. Building a company gradually from the ground up means you get a better sense of what it takes to run a successful business (or in other words, experience is the best teacher).
No, because you're making a service company. Culture eats strategy for breakfast. The org chart you need for product and the one you need for services are different. The necessary mindset of the employees are different.
Not only that, but the buyers of product and services are often different as well.
Sure, building up a huge services company lets you squirrel off some funds for product development, but how often do you see this in practice? The only example I can think of is IBM, and they're moving in the other direction.
To counter your hypothetical "can't be done", here's a real example of a product company that's grown a single person selling electronics kits from their college dorm room to a product company making $33 million in revenue less than 10 years later:
A similar example being this product company, which started out in 2012 with two people and a laser cutter, growing to a product company that employs 30+ people within the space of 5 years:
thanks @ZenoArrow ... additionally, it's over $42m and we have a free and paid for service, http://www.adafruit.io and a subscription service, ADABOX, with thousands of customers, http://www.adabox.com ... no loans, no VC, 114+ people, USA manufacturer of open-source hardware in NYC, USA.
So they started out as hardware product companies.. and remained hardware product companies. I don't see any examples of software services companies transitioning to product.
Who said anything about software service companies? I said design agency, which is not the same thing as a software service company. Sure, they use software to do their job, but that's not the most important part of what they do.
As for examples of design consultancies morphing into manufacturing companies, Cambridge Audio is one such example:
