Is GIS industry always ruled by two options only, namely Esri products and OSGeo projects?
I disagree that there are only two options in the GIS industry on a number of levels. The first is that there are many other well established commercial GIS offerings other than ESRI SmallWorld, Bentley, AutoDesk, ERDAS, MapInfo, Integraph and Idrisi spring to mind without thinking too hard. You say they have a market share "well below" ESRI. ESRI admittedly has 40% of the market share, but that still leaves 60% for everybody else and Integraph has more than half of ESRI's market share - which is a substantial proportion and revenue.
Secondly, to state that ESRI is the market leader displays a bias towards 'traditional desktop' GIS. Broaden your field a little and a company called MDA outstrips ESRI. You can debate whether Googl;e is a 'real' GIS but it certainly provides basic functionality to the masses and arguably has infinitely greater penetration into the home-user market than ESRI can ever claim. We haven't even touched on NASA, Trimble and the rest of the GPS market, all of whom are part of the wider GIS industry.
Thirdly, it is hardly logical to lump all FOSS software into a single camp anymore than it is to say that ESRI it would be to say that the only commercial GIS is related to ESRI. You might make a case for lumping all software under development by OsGeo together (though I would dispute the validity of that too). Even then, this would still leave many 'indie' FOSSGIS products un accounted for.
You mention Oracle. Oracle spatial is one of the biggest and most common GIS databases available. No, it is not 'a GIS' but it is part of the GIS industry. Of course the giants like MicroSoft and Google are interested in GIS. Why else would they develop Bing Maps (Microsoft) and Google Earth? However, when it comes to fully-featured hard-core GIS, they are decades behind the likes of ESRI, Integraph and GRASS (for instance). This is not their market and they wisely are sticking to what they know, which is software for the masses.
Far from being a bad thing, I believe it is a very positive sign of a young, healthy and growing industry that there are in fact so many companies carving out names for themselves alongside a very vibrant FOSS community. To have a couple of giants like MicroSoft controlling GIS development would be stifling. You really would have the scenario of only one or two products to choose from.
ESRI's been around for a long time, and essentially helped invent the term "GIS". There are other big players, but they often come from a different angle (i.e. AutoCAD Map 3D, or Intergraph/Microstation). Increasingly all these different dominant players in the maps/drafting/design world are starting to overlap and come together, but they still hold their own niches.
There are some smaller players that have made decent attempts (i.e. Manifold), but as someone who works for a big company, intertia is hard to overcome, and big companies want to go with the de Facto "standard", so they by Microsoft and ESRI (who have partnered in various ways), rather than looking at Linux/FOSS/smaller companies.
Could the industry benefit from an ecosystem of smaller companies? Certainly. Is it more likely that FOSS will help provide the necessary pressure on ESRI? I suspect so...
If the question is to help determine how to get a start-up GIS software company off the ground, I'd suggest looking for other angles into the market than trying to go straight up against the like of ESRI. Gaining a clear understanding of pinch points in other IS/IT workflows, and working to leverage off of one of those pinch points is likely more productive than trying to make a desktop software to compete with the big boys (no one is going to want to learn another package on top of what they've already learned unless it's free, or very obviously helps them past a major pinch point in their workflow, no matter how much better that software is than other commercially available packages).
To answer the question: Why no industry standard product from any established software giants?
The problem is you appear to be begging the question. There is an industry standard product from an established software giant. They're called ESRI and being founded in 1969 they easily predate Microsoft (1975), Google (1998), Oracle(1977), and Apple(1976). The reason none of those corporations offer a full GIS is because, its outside of their remit. Each of them dominates their field, just like ESRI does GIS, and each of them has smaller-scale proprietary and open source competitors.
Microsoft is first and foremost a OS company, though they also make office productivity software. While they have their fingers in a lot of pies, most of those efforts are typically lackluster and have a dismally small market share. Exceptions are Bing and Xbox where they effectively used their money to buy a share; both of these things supports their core products however.
Google is an advertising company that happens to do well in Search. They too branch a lot but everything they do has a web-element.
Oracle is a database company. They have a database extension to their GIS but why would they offer a full GIS? Again, its outside of their remit.
ESRI are a GIS company. They don't offer a database solution, or a web-search solution or an Operating System, just like those companies don't offer full desktop GISes. It just happens that GIS is a more niche field which is probably why ESRI are "only" worth about 1% of the other companies listed here.