For the sake of argument let's define a great software company as having at least one dominant product in a lucrative market, and great for a product as being both the dominant brand in its market and financially meaningful to its owner. I'm using "software company" in a very loose sense here to mean any company who main interaction with customers is through software - this encompasses e-commerce and digital media - and I don't care about greatness from a quality perspective.
If you only look at software, Google is all search; Facebook is all Facebook; Amazon is all e-commerce; and Apple, Oracle, Intel, HP, and IBM don't exist.
It's tempting to say that products like Maps and YouTube are meaningful to Google, but it's not that hard to imagine the company without them. Look at how the story of Apple dropping Maps and YouTube from the iPhone has played out. Meanwhile Facebook has designed its product to encompass every possible product they could build (unclear what's going on with Instagram, which they didn't build). Etc.
Is it because software lasts forever, perpetuating the same company cultures and business lines? Are software companies doomed to the innovator's dilemma? We don't have a long enough history of software companies to say, but I wouldn't be surprised. When you can't separate the company from the product, you're trapped by your brand. Kodak didn't fail because they didn't know how to do digital, they failed because Kodak had become so associated with film that no one could conceptualize their brand as standing for anything else. Does that sound familiar for Google and social? Sure Google is still killing it on earnings, but remember that with every disruptive technology the incumbents continue to look invincible until the moment they fall.
Compare that with hardware: hardware goes obsolete much faster and has more inflexible requirements than software (e.g. form factor, cost of production), meaning a company has to put out many more hardware products at a time than software products. Apple probably managed to consolidate hardware SKUs more than any other company ever, but they still have five major hardware lines (iPod, iPad, iPhone, MacBook Air, MacBook Pro) to three major software lines (iTunes, iOS, and OSX).
Another factor is that a single software "product" can actually encapsulate many different functionalities and price points (think SaaS, or even Microsoft Office tiers) without having to become a separate product to the market. There has been some experimentation with this on the hardware side, with Oracle and others renting servers on-premise based on how much of its resources you consume, but I'm not sure how this has worked out with the rise of the cloud (the cloud is the transformation of hardware into software; from the customer's perspective it is a digital abstraction of hardware).
The one example of a single corporate brand being extended to more than one wholly separate great software product is Microsoft with Office and Windows (and DOS and Basic before that). Exceptions might be IAC, which does it by running the products as wholly separate businesses within a single corporate structure; and companies like 37signals that have made a conscious decision not to put their company name on their products and so have not come to have their brand identified with any one of them. I'd love to get more examples or counter examples.
It's also interesting to consider what this means if software is indeed "eating the world." No has paid much attention to the dichotomy between Apple becoming the world's most revered and valuable company and their scorn for software businesses at the very moment that software is considered the thing. Either this hypothesis has to be wrong, software companies have to become capable of creating more than one product, or they are going to keep getting smaller as disruptive innovation cycle times speed up and there is only so much gain to be made on each turn.