Instead of asking why one hasn't happened yet, maybe we should instead ask why national-security experts keep warning about them.
One possible answer is that national-security experts are ignorant. I get the sense that "national" security experts have very little expertise in "cyber" security. That's why I include a brief resume at the top of this article, I've actually broken into a power grid and found 0days in critical power grid products (specifically, the ABB implementation of ICCP on AIX -- it's rather an obvious buffer-overflow, *cough* ASN.1 *cough*, I don't know if they ever fixed it).
Another possibility is that they are fear mongering in order to support their agenda. That's the problem with "experts", they get their expertise by being employed to achieve some goal. The ones who know most about an issue are simultaneously the ones most biased. They have every incentive to make people be afraid, and little incentive to tell the truth.
The most likely answer, though, is simply because they can. Anybody can warn of "digital 9/11" and be taken seriously, regardless of expertise. It's always the Morally Right thing to say. You never have to back it up with evidence. Conversely, those who say the opposite don't get the same level of press, and are frequently challenged to defend their abnormal stance.
Indeed, that's how this article by
The Atlantic works. It's entire premise is that the national security experts are still "right" even though their predictions haven't happened, and it's reality that's somehow "wrong".
Now let's consider the original question.
One good answer in the article is that terrorists want attacks that "cause certain types of fear and terror, that garner certain media attention, that galvanize followers". Blowing something up causes more fear in the target population than deleting some data.
But something similar is true of the terrorists themselves, that they prefer violence. In other words, what motivates terrorists, the ends or the means? It is it the need to achieve a political goal? Or is it simply about looking for an excuse to commit violence?
I suspect that it's the later issue. It's not that terrorists are violent so much as violent people are attracted to terrorism. This can explain a lot, such as why they have such poor op-sec and encryption, as
I've written about before. They enjoy learning how to shoot guns and trigger bombs, but they don't enjoy learning how to use a computer correctly.
I've explored the cyber Islamic dark web and come to a couple conclusions about it. The primary motivation of these hackers is gay porn. A frequent initiation rite to gain access to these forums is to post pictures of your, well, equipment. Such things are repressed in their native countries and societies, so hacking becomes a necessary skill in order to get it.
It's hard for us to understand their motivations. From our western perspective, we'd think gay young men would be on our side, motivated to fight against their own governments in defense of gay rights, in order to achieve marriage equality. None of them want that, as far as I can tell. Their goal is to get married and have children. Sure, they want gay sex and intimate relationships with men, but they also want a subservient wife who manages the household, and the deep family ties that come with spawning progeny. Thus, their motivation is still to defend the umma (the whole community of Muslims bound together by ties of religion) against the West, not pursue personal rights.
The point is, when asking why terrorists do and don't do types of attacks, their own personal motivations are probably pretty darn important.
Another explanation in that article is simply because Islamic hackers aren't good enough. This requires a more sophisticated discussion of what skills they need. As
The Atlantic says in their article:
The most powerful likely barrier, though, is also the simplest. For all the Islamic State’s much-vaunted technical sophistication, the skills needed to tweet and edit videos are a far cry from those needed to hack.
It's indeed not just "editing videos". Most hacker attacks you read about use un-sophisticated means like phishing. They are only believed to be sophisticated because people get confused by the results they achieve with the means with which they do it. For example, much of the DNC hack which had important consequences for our election was done simply by phishing the password from people like John Podesta.
A convincing cyber terrorism attack, such as causing a power black out, would take different skills -- much rarer skills. I refer to my pentests above. The techniques used were all painfully simple, such as SQL injection from the Internet, but at the same time, it's a much rarer skill. No matter how simple we think SQL injection is, it takes a different skillset than phishing. It takes people more interested in things like math. By the time such skills are acquired, they get gainfully employed at a technical job and no longer have free time to pursue the Struggle. Phishing skills won't land you a high paying job, but web programming (which you need for SQL injection) will.
Lastly, I want to address the complexity of the problem.
The Atlantic quotes Robert M. Lee of Dragos, a well-respected technical expert in this area, but I don't think they get the quote right. He points out the complexity of the power grid. What he means is not complex as in
hard but complex as in
diverse. There's 10,000 different companies involved in power production, long haul, distribution to homes, and so forth. Every state is different, every city is different, and even within cities there may be multiple small companies involved.
What this means is that while hacking any one of these entities would be easy, it'd only cause a small-scale effect. To cause big-scale effects would require a much larger hacking campaign, of a lot of targets, over a long period of time. Chances are high that before you hacked enough for a convincing terror effect, they'd catch on to you, and take moves to stop you. Thus while any individual target is easy, the campaign as a whole is complex.
In the end, if your goal is to cause major power blackouts, your best bet is to bomb power lines and distribution centers, rather than hack them.
Conclusion
I'm not sure if I have any better answers, just more complex perspectives.
I think there are lots of warning from so-called "experts" who aren't qualified to make such warnings, that the press errs on the side of giving such warnings credibility instead of challenging them.
I think mostly the reason why cyberterrorism doesn't happen is that which motivates violent people is different than what which motivates technical people, pulling apart the groups who would want to commit cyberterrorism from those who can.
At least for power grid blackouts, while small attacks would be easy, the ones large enough to grab people's attention would be difficult, due to our power grid's diversity.
