Yes, but...

carbonated and bottled*, and the insulin must be carefully extracted and isolated from the bateria using a combination of centrifugation, cell lysis, chemical treatments, more centrifugation and so on. 

The result is that there is an inevitable cost to produce a unit of output, which is very strongly affected by the cost of the inputs, running costs of the environment, and cost of processing the product. With a lot of these costs already at or near the limits of what can be achieved**, the final variable is the organism. Which is why so much effort is expended in the breeding of dairy cows, selection of strains of beer yeast, and design and transformation of genetically-engineered microorganisms. Industrial biotechnology, as an aside, is both the majority of biotech today that isn't GMOs, and is overwhelmingingly tied to the microbe + bioreactor approach.

This is where your biotechnologist comes into the picture, because this is the person working on the organism side of the equation. And, unfortunately, the discipline tends to select for a certain type of entrepreneur. Specifically, the type that develops a gee-whiz new technology for producing something (often coming out of an academic environment), and then looks for a way to make a business out of it. Which is, unfortunately, completely backwards.

Good entrepreneurs identify a problem that sounds in money, then solve it***, ideally by inventing as little as possible and removing risk wherever they can. And a lot of the problems that biotechnology is trying to solve are already solved in one form or another by existing industries. The video mentions a few products, some of which I've seen play out in my own practice (albeit from the sidelines): biofuels, food proteins (bulk proteins and mammalian muscle cell cultures), cosmetics, other proteins (peptides and enzymes) and drugs (monoclonal antibodies). A lot of these are either harvested from nature (fuels, cosmetics, food proteins, some other proteins) or can be at least partially made synthetically (fuels, cosmetics, peptides) using chemical processes. And the result is some pretty brutal math. For a fully scaled-up fermentation setup, the video author estimates a cost of around $100 per litre to run the fermenter itself, with the cheapest known technologies (e.g. beer brewing) running at something like $5-10 / l. As for what you can get for that cost using current industrial biotech methods, biofuels math out to around $0.1/l, food proteins around $1/l, cosmetics $5/l, other proteins $100/l (not including the cost of the extensive processing needed to purify the product) and drugs $1000/l. By way of comparison, I should also add that milk costs something like $0.4/l to produce, while eggs are in the region of $0.6-0.7/l.

The core issue is that the inputs and environments for industrial biotech processes just do not play nicely with the prices that their products can charge on the open market. This, combined with the sorts of people starting companies in this field of technology, results in a constant conveyor-belt of invention-heavy start-ups trying to break into markets where they realistically have no hope of competing with existing production technologies. The solution is to reverse the process and find problems to solve that only industrial biotech can solve, where there is a market with strong demand, and where the relatively high price of production can be justified. Which is why so many of the surviving companies are making biologics rather than biofuels.

On a final note, the author also mentions the 'pickaxe and shovel' companies, which tend to be more successful than the pure industrial biotech companies because they are selling specialised tools and materials into an existing market (biotech startups burning VC funding). This isn't an inherently dishonourable (hah) business to run, and good luck to anyone that can find a steady, well-paying niche out there. But bear in mind that the era of easy start-up money (which never really existed in South Africa to begin with) seems to be nearing an end, and that you need to have a plan for what to do with all the pickaxes and shovels when the gold prospectors move on.

* I know that some beers are naturally carbonated, some are not filtered or pasteurised and so on. Bear with me here.

** This is also a bit of a simplification. Most biotech systems which use bioreactors (those stainless steel brewing tuns or fermentors) work in batch mode, where the reactor is loaded, runs, and is then unloaded again for the next run. For various reasons this is inherently less efficient than a system which runs in continuous mode - constantly taking in input materials at one end and constantly outputting a stream of product and waste materials through the other. Unfortunately, for various reasons this is very hard to do with most cell culture systems, which is why brewing still runs in batch mode despite beer being a mass-produced product with the accompanying razor-thin production margins (with a few very interesting exceptions).

*** Not all entrepreneurs, or even all successful entrepreneurs. Plenty of great products and successful companies have been spun out from an inventive new technology looking for a market. But pointing to those winners is a bit like pointing to someone who won the lottery to justify blowing your life savings on scratchcards.