On a recent business mission across ASEAN, a fellow delegate asked me a question I haven’t stopped thinking about. With cyber incidents seemingly stacking up around the world, was this the opening move of a global digital war? Was a single actor, armed with highly advanced artificial intelligence, finally making its move on the West?
That conversation happened before this week. Then, within roughly 48 hours, Microsoft 365 went down across more than 70 countries, the United States grounded every departing flight for over four hours, and reports of a nationwide weather service outage spread faster than the agency could deny them.
So, was my fellow delegate right?
My honest answer: I don’t know. And I would take an educated guess that neither does anyone else.
Only the people behind an attack ever truly know their motives. But having read widely on the crises that have shaped us, my instinct is that most “perfect storms” are exactly that: unrelated failures colliding in the same window of time.
So far, this week’s evidence points the same way. Microsoft says its outage was a technical fault linked to a network upgrade, not a cyber attack and not AI. US officials have pointed to new equipment overwhelming decades-old air traffic telecommunications. The weather service says its radars never went down at all. Three events, no established link.
I also struggle to see the motive for a single state actor. Read together, Chris Miller’s Chip War and Dan Wang’s Breakneck left me with a clear picture: the world’s major economies are bound together by supply chains, trade and ambition. A rival that depends on your markets gains little from your collapse.
Perhaps historians will look back on this decade and call it a world war of a different kind. There is conflict in the Middle East, tension involving Russia and China, an AI arms race and deep political division in the United States, so I understand why people reach for that frame.
But here’s what I told my peer: it doesn’t matter.
An organisation that can’t send an invoice doesn’t need to know who caused the outage in order to recover from it. The impact is identical either way: you can’t reach the systems you rely on, and you can’t do the work that brings in revenue. At a household level, the same pattern shows up as higher fuel and food prices, or worse. Water security is no longer a problem we can attribute to “somewhere else”. The Panama Canal, which carries a significant share of world trade, is cutting shipping because of drought for the second time in three years. We wrote about this recently.
What the Microsoft outage taught us (again)
In the interest of transparency: Fixinc runs on Microsoft 365, and we were affected by the recent outage. We’ve been here before, so we had simple, tested workarounds ready.
The single most effective mitigation for this kind of event is a secondary communications channel that is genuinely independent of your corporate system. It needs to be built, tested and known by your people before you need it.
Most organisations don’t have one. I ask executives this regularly, and I genuinely can’t remember the last time someone said yes.
The excuse is wearing thin. This was Microsoft’s second major multi-service outage this year. In July 2024, a faulty CrowdStrike update crashed around 8.5 million Windows devices worldwide, grounding flights and disrupting hospitals. If that didn’t move a backup channel up your priority list, what will?
We haven’t finished learning from COVID-19 either
The same complacency shows up outside technology. Many organisations now treat remote working as their pandemic plan. But moving laptops out of the office was never the hard part. The mental health and financial impacts were, and few plans address them.
Crises no longer queue politely
My colleagues and I in the resilience profession have said for years that crises don’t arrive one at a time. A traveller in the US this week could have lost their email and their flight on the same day, while scrolling past false reports that weather data had gone dark. The events were unrelated, but they were felt all at once. Concurrency is now the baseline we have to plan for, and misinformation travels with it.
The AI question every board should be asking
Online speculation was quick to blame AI for both outages. Microsoft and the FAA’s AI contractor have denied it, and nothing yet says otherwise. I won’t claim what the evidence doesn’t show.
But the context deserves attention. The FAA rolled out a new AI-supported traffic management tool nationwide just days before the ground stop. Microsoft is asking tens of thousands of staff to consider leaving while it spends heavily on AI. Neither has been linked to this week. Early findings suggest the FAA’s failure came from new technology being bolted onto equipment more than twenty years old. That isn’t AI, but it’s the same pattern: change moving faster than the systems underneath can absorb. Together, they raise a question every leadership team should ask:
are we moving faster on AI than our ability to understand, govern and recover the systems underneath it?
I regularly see organisations inject AI into processes that have run well for years, expecting improvement, and getting disruption, slower delivery and unexpected cost instead. Our UK partners at Informed Solutions often find that their AI reviews recommend fixing the underlying process before introducing the technology at all. Customer service is a good example. AI can surface a knowledge base in seconds, but it can’t replace the initiative and imagination that make a great service representative.
The real risk isn’t whether AI is smart or dumb. It’s that access is universal, enthusiasm is high and governance lags. Decisions get made that an experienced engineer or product manager would have challenged. And when the people who understood how a legacy system actually works have moved on, recovery gets slower. That’s a resilience problem, not a technology one.
One great attack, or many small decisions?
My bet is on the latter: many ordinary decisions, under-governed, landing in the same funnel at the same time. Whichever it turns out to be, the response is the same.
Act now. Prepare with calm and with purpose. Preparation isn’t scaremongering or an admission of weakness. We should be proud to say, “We’re in a crisis, and we’re managing it.” The alternative is being misled, or having no plan at all.
No investor would knowingly back a business that can’t recover from disruption. Yet very few ask to see that plan tested. They should.
So, back to that question on the mission. Is this a digital world war? History may decide. But the leaders who come through this decade well won’t be the ones who guessed the attacker correctly. They’ll be the ones who could still reach their people when the screens went dark.
If Microsoft went down again next month (over Christmas), how would you reach your team in the first hour?