There are any number of organisations and organisational functions which are essentially risk management. Some of those risks are upside (new products / markets), some are downside (loss, fraud, moat-building, etc.).
You'd expect that a VC of all people would appreciate the fact that in high-tech spaces, many new ventures are required to find one unicorn, with absolute failure rates of about 90%,[1] and Angellist giving 1:40 odds (2.5%) likelihood of seed-stage startups reaching US$1b valuation ("unicorn").[2]
Google has hit paydirt once with advertising, and spent most of its subsequent life either reinforcing that status (with varying levels of legality) or seeking out alternative hits, with ... stunningly little success. But then, nobody else has had any real success there either, save Facebook.
There are also the roles within organisations which are principally risk mitigation. I had the insight a few years back that ops is essentially a risk-management role.[3] In one of my earlier mentions of this on HN someone responded that that is in fact how Google treats its SRE role, and is a point emphasised in its fairly well-known book on the topic.
Another point that's become glaringly clear over the past few years is that efficiency and resilience are contrasting points on the optimisation space. I'm not sure if that's one-dimensional or multi (think "fast, cheap, good"), but it does have multiple optima depending on the specified criteria.
And ... in some spaces, notably pharmaceuticals (very large search space, very high search costs, long-term exploration), a business model which has emerged is to spin out new-molecule / treatment development indepenently of the majors (Pfizer, J&J, Roche, Novartis, Merck, AstrZeneca, GSK, etc.), either as controlled subsidiaries or independent ventures, and then aquire the successes. Contrast other R&D-heavy models such as 3M, Edison Mfg. Co., DuPont, etc. There's also the hit-factory / tentpole model of film, music,[4] and print (a small number of hits supports a much larger mid-list). I'd argue that much of the start-up and established-firm tech space has come to resemble these models, for much the same underlying dynamic logic.
--------------------------------
Notes:
1. https://explodingtopics.com/blog/startup-failure-stats> https://www.luisazhou.com/blog/startup-failure-statistics/>
2. https://web.archive.org/web/20210716175009/https://www.angel...>
3. https://news.ycombinator.com/item?id=27701929>, https://news.ycombinator.com/item?id=29152832>, and https://archive.is/HknlO>
4. I've discussed this and referenced Charles Perrow's excellent treatment here: https://news.ycombinator.com/item?id=31254795>. Perrow's discussion is in Complex Organizations (1972, 1985), pp. 186--187. https://archive.org/search.php?query=Complex+organizations+:...>