Rubio Says US Is ‘Not Under Systemic Attack’ as a Week of Outages Hits Microsoft, Banks, the FAA, the NSA and Apple
The United States is not under attack. That was the message Secretary of State Marco Rubio delivered on Tuesday, after six days in which the systems Americans depend on most failed one after another: Microsoft’s email and workplace software on Thursday, the nation’s air traffic system on Friday, the biggest banks on Monday, a classified National Security Agency network on Tuesday, and then Apple — iCloud, the App Store, Apple Pay and Apple Music — on Tuesday afternoon.
“We are not under systemic attack,” Rubio told reporters in Manila, where he was travelling with President Donald Trump for an economic conference. “These events, one has nothing to do with the other, they are not inter-related.”
Rubio said investigators had found “zero evidence” connecting the disruptions to hostile actors. “Every indicator we have is that every single one of these events has an explanation or a cause that has nothing to do with a cyber criminal, has nothing to do with a hostile actor or another country,” he said, according to Reuters.
The Director of National Intelligence, Jay Clayton, gave the same assessment separately, telling CNBC that the outages were “not a cyberattack” and that patching and upgrading ageing systems can itself cause disruptions, according to ABC News.
President Trump played the episode down in Manila, telling reporters that things “like that happen” and that “it’s fine”, according to Reuters.
But the reassurances have not settled the argument. The failures arrived so close together, and touched so many parts of daily life at once — work email, flights, bank accounts, phone payments, a spy agency’s own internal network — that lawmakers from both parties are demanding briefings and investigations, cybersecurity experts are warning that the pattern itself is the problem, and, according to Axios, the past week brought the highest number of weekly outages in 12 months, based on data from the outage-tracking site Downdetector.
Each incident, taken on its own, has a mundane explanation. Taken together, they have become a stress test of a country that runs on software — and the test has not gone well.
Thursday: Microsoft goes down
The sequence began on Thursday, 12 November, when Microsoft suffered a major worldwide outage affecting its widely used services, including Outlook email and Teams, the workplace messaging platform used by millions of businesses.
Microsoft subsequently attributed the disruption to a combination of software defects, increased network traffic and reduced server capacity, according to the technology publication Intelligent CIO. Other reporting described the cause as a network capacity upgrade that had encountered technical problems. Either way, the company ruled out a cyberattack.
For many workers, the outage was the first sign of what the week would become: a morning in which email would not send, meetings would not load, and the tools of the working day simply stopped.
Friday: the skies close
The following day brought the most consequential failure of the week. On Friday, 13 November, the Federal Aviation Administration ordered a nationwide ground stop, suspending departing flights across the United States for more than four hours — the longest nationwide halt to American air traffic since the attacks of 11 September 2001.
About 20,000 flights were cancelled between Friday and Sunday, according to CBC News. Passengers were stranded for hours, and in some cases for days; earlier reporting put the toll at around 7,000 cancellations on Friday alone, with some travellers facing delays of up to 60 hours.
The cause, according to Reuters, was a newly installed Juniper router from the defence contractor L3Harris at an FAA radio site in El Dorado, Arkansas. The router overloaded a key communications system, firing off 1,000 connection attempts every four to five seconds and destabilising communications and radar at 17 of the FAA’s 22 en route air traffic control centres. A second site, in Amarillo, Texas, was also affected. The FAA has paused the rollout of the new routers — 3,648 are due to be converted, of which 400 have been installed and 69 are operational — pending testing at an L3Harris laboratory in Florida.
Transportation Secretary Sean Duffy, who stood beside the Homeland Security Secretary at a White House briefing on Friday afternoon, insisted there was never any risk to life, and said the episode showed that the air traffic control system needed to be rebuilt rather than patched. He has called for 50 billion dollars of upgrades to the system.
Thirteen Democratic senators, led by Maria Cantwell, have urged the FAA to keep the router deployment paused and to give the Senate a technical briefing by 23 November. Senator Josh Hawley, a Republican, called the router explanation “a little hard to believe” and has pressed for a full Senate briefing. The initial explanations given by the Department of Transportation and by L3Harris differed, and the full root cause remains under review.
The ground stop also produced the week’s most damaging political sideshow: it emerged on Tuesday that the Homeland Security Secretary, Markwayne Mullin, had flown on a government aircraft on Saturday to a college football game in Louisiana, less than a day after the airspace reopened, sitting in the private box of a shipbuilding executive whose company holds billions of dollars in Department of Homeland Security contracts. Mullin told CNBC on Tuesday morning: “We still gotta live our life at the same time.” The department says the secretary is required by law to travel on government aircraft and pays personally for non-official travel.
Monday: the banks falter
On Monday, 16 November, the failures reached Americans’ money. Customers of JPMorgan Chase, Bank of America and Wells Fargo reported trouble reaching their accounts online, and the money-transfer platform Zelle also experienced disruptions, according to reports cited by Intelligent CIO and other outlets.
JPMorgan attributed its problems to a software bug in on-premises hardware and said the issue was “not cyber-related”. The bank said the fault then spread through parts of the financial system. The same day, United Airlines suffered a technology failure that delayed about 150 flights.
For customers locked out of their accounts, the reassurance that no crime had taken place was cold comfort. The money was still there; the access to it was not.
Tuesday: the NSA, then Apple
Tuesday, 17 November, brought two failures that, more than any others, fed the suspicion that something larger was happening.
First, a classified communications network used by employees of the National Security Agency went down. The agency said on Tuesday that it had “experienced temporary degradation” during scheduled routine maintenance, according to Reuters. The White House said it saw “no evidence that this was a cyberattack”. The network carries classified communications inside the agency; its failure, even briefly, was the kind of event that would normally pass unnoticed by the public — and it did not pass unnoticed this time, because it arrived in the middle of a week of failures everywhere else.
Then, shortly before noon Eastern time, Apple began to fail. Reports of problems with Apple Pay spiked at around 11:40 a.m. Eastern, according to Downdetector, peaking at about 2,300 reports from users in the United States. Reports of iCloud problems surged just before noon and peaked at roughly 1,800 to 2,100 reports. The App Store, Apple Music, Apple Support and AppleCare were also affected, and some users reported being unexpectedly signed out of their Apple accounts on iPhones, iPads and Macs without warning.
Apple’s own System Status page confirmed problems with iCloud account sign-in, the App Store, Apple Music and AppleCare on Device, saying the services “may be slow or unavailable” and that all iCloud users were affected. By mid-afternoon, Apple said an internal update was to blame. “Many of our systems are back up and running after an internal update caused some outages this morning,” the company said in a statement reported by Bloomberg and USA Today. “We’re working to get the remaining systems online as soon as possible.” Apple told Bloomberg the problems were not caused by a cyberattack.
By 5:24 p.m. Eastern, Apple’s System Status page showed all affected consumer services functioning normally, according to USA Today. Some developer and enterprise services took longer to recover. The brokerage Charles Schwab also reported a glitch on Tuesday, with more than 1,000 user reports on Downdetector.
That was five weekdays, five failures — and the week was not over.
The wider picture: it was not only America
The American outages were the most visible part of a wider run of technology failures and cyber incidents, according to Reuters. In the same period, governments took websites offline to address vulnerabilities in the PHP programming language, which powers a large share of the world’s websites. A North Korean-linked group was blamed for a 6.3 billion dollar hack of the Binance cryptocurrency exchange. And food cooling systems in France and elsewhere in Europe were hacked in attacks the French government attributed to Iran.
None of those incidents has been linked to the American outages. But their arrival in the same week sharpened the sense, among the public and in Congress, that the world’s digital infrastructure is under strain from several directions at once.
That sense was not helped by the rumours. Unfounded claims spread on social media in recent days that a worldwide blackout of up to 90 days was coming, with one widely shared post warning of “no electricity, no internet, no banking”. Some pointed to a decades-old episode of the television show The Simpsons as supposed evidence. The claims have no basis, according to reporting by The Sun, which noted that a planned FEMA emergency alert exercise was sent through television and radio rather than to phones, which is why many Americans did not receive it on their handsets.
The rumours were false. The outages were real. The gap between the two is part of the story: when official explanations arrive slowly and in fragments, speculation fills the space.
Analysis: Why It Matters
Five failures, five causes — coincidence or pattern?
The official position is unambiguous: these were five separate accidents with five separate causes. A Microsoft capacity upgrade, a faulty FAA router, a JPMorgan software bug, scheduled NSA maintenance, an Apple internal update. Rubio’s formulation — “one has nothing to do with the other” — is, on the evidence available, accurate in the narrow sense. No public evidence links any of the incidents to a hostile actor, and each of the organisations involved has given a technical explanation that fits the facts as reported.
But the narrow sense is not the whole story, and the people paid to think about this for a living are saying so. Cynthia Kaiser, a former senior FBI cyber official who is now a senior vice president at the cybersecurity firm Halcyon, told Reuters that “while we’re still waiting to understand the exact cause behind today’s NSA Net outage, the recent string of major disruptions across industries highlights how deeply interconnected our digital infrastructure has become, and how a single technical failure or software update can cascade with outsized effects across critical systems.”
That is the pattern that matters. The incidents do not need to be connected to each other to be connected to the same underlying condition: a small number of systems, run by a small number of companies, on which almost everything depends. When Microsoft’s email fails, millions of businesses cannot work. When the FAA’s routers fail, the skies close. When JPMorgan’s software fails, people cannot reach their money. When Apple’s update fails, people cannot pay for things with their phones. The concentration is the vulnerability, whether or not anyone attacks it.
The update is the new single point of failure
Look closely at the week’s causes and a second pattern appears: three of the five failures were caused, directly or indirectly, by the very processes meant to keep systems safe and current. Microsoft’s outage followed a network capacity upgrade. Apple’s followed an internal update. The FAA’s followed the installation of a new router as part of a modernisation programme. The NSA’s failure came during scheduled routine maintenance.
This is the paradox of modern infrastructure. Systems are so complex, and so interdependent, that changing them is itself dangerous — and not changing them is more dangerous still. Old software accumulates flaws; new software introduces them. The FAA’s air traffic control system is a case in point: parts of it are decades old, built on copper wire and equipment that vendors no longer support, and replacing it piece by piece means touching a live system that cannot be switched off. Duffy’s call for 50 billion dollars of upgrades is an admission of how far behind the replacement has fallen; the El Dorado router is evidence of how risky catching up can be.
The industry learned a version of this lesson in July 2024, when a faulty update from the cybersecurity company CrowdStrike disabled about 8.5 million Windows computers worldwide, grounding flights, halting hospital systems and shutting down banks. The lesson then was that a single bad update, pushed to a single widely used piece of software, could paralyse the world. The lesson of this week is that the same thing can happen without any single bad update at all — five ordinary ones, in five different places, in five days.
AI is finding the flaws faster than humans can fix them
There is a third thread running through the week, and it may prove the most important. Kaiser told Reuters that “we are in an era where AI is accelerating the discovery of legacy software flaws, forcing organizations to race to patch vulnerabilities before adversaries can exploit them, which inevitably creates friction and disruptions.”
The logic is straightforward. Artificial intelligence tools can now find long-hidden bugs in widely used software far faster than human researchers could. That is good news for defenders — but it also means that organisations are discovering, all at once, how much of their software is vulnerable, and are being forced to patch it at a speed their systems were never designed for. Every patch is a change; every change is a risk. The PHP vulnerabilities that forced governments to take websites offline this week are an example: severe flaws in a language that powers a huge share of the web, made worse, in the words of one expert, by large language models that make such flaws “much easier to surface and weaponize”.
Dr Daniel Gardham, a lecturer at the Surrey Centre for Cyber Security, put the dilemma plainly to Reuters: “Finding vulnerabilities is becoming faster and easier, but fixing them across the range of systems that rely on this software is much harder. How well organizations manage that gap will determine whether AI makes the internet safer or more vulnerable.”
Senator Mark Warner, the senior Democrat on the Senate Intelligence Committee, warned during the week that artificial intelligence was outpacing cybersecurity defences. The week’s outages suggest he has a point — not because AI caused them, but because the race to keep up with AI-driven vulnerability discovery is itself producing the churn that causes them.
One unconfirmed report, cited by the Washington Sun and attributed to three unnamed sources, claimed that Wall Street executives and the NSA had told members of Congress the disruptions came from attempts to fix gaps in cyber defences that AI models had identified. That claim has not been verified by a second source and should be treated with caution. But it points in the same direction as the on-the-record evidence: the maintenance burden is growing faster than the capacity to carry it out safely.
When everything fails at once, trust is the casualty
The most durable damage from a week like this may not be measured in cancelled flights or failed payments. It may be measured in trust.
Consider what an ordinary American experienced. On Thursday, work email stopped. On Friday, flights were grounded. On Monday, the bank would not open. On Tuesday, the phone would not pay for lunch, and the agency that spies on the rest of the world could not keep its own network running. Each time, the explanation was the same shape: a technical fault, not an attack, nothing to worry about. And each time, the explanation arrived hours after the failure, in fragments, from different organisations, with different details.
That is how conspiracy theories are born — not from nothing, but from a vacuum. The viral claim of a 90-day global blackout was baseless, but it spread because, for five days, the evidence of people’s own eyes told them that the systems were failing. Rubio’s “not under systemic attack” was meant to close the argument. For many people, it will have sounded like a statement about the wrong question. The public was not only asking whether someone had attacked America. It was asking whether America could keep its own systems running. On that question, the week’s evidence is less reassuring.
The experts Reuters spoke to were careful to draw the same distinction. “Even when an outage isn’t caused by a cyberattack, these moments of operational distraction leave dangerous gaps open for threat actors and opportunists to take advantage of,” Kaiser said, warning that people should remain on high alert against social engineering and unsolicited messages during such episodes. Nic Adams, the co-founder of the cyberkinetics company 0rcus, advised people to check an affected company’s instructions by opening its app or typing its web address themselves, and to treat any urgent message telling them to “secure your account” with extra suspicion. Scammers, he noted, move fastest when the headlines are confusing.
The test is what Washington does next
The week will be judged not by what officials said about it, but by what they do about it. There are concrete tests ahead, and they arrive quickly.
The first is the FAA. The router investigation is still open, the Senate wants a technical briefing by 23 November, and the question of whether the modernisation programme should continue, pause or be redesigned has not been answered. The initial explanations from the Department of Transportation and from L3Harris differed, and until they are reconciled, the public cannot know whether the ground stop was a one-off fault or a symptom of a programme that is being run too fast.
The second is transparency. Microsoft, JPMorgan, Apple and the NSA have each given a cause, but none has yet published a full account of what went wrong, what was affected and what will change. The companies that run the world’s infrastructure are, for the most part, accountable to their shareholders first and to the public second. A week in which five of them failed in five days is an argument for treating their reliability as a public matter, not only a commercial one.
The third is investment. Duffy’s 50 billion dollar figure for air traffic control modernisation is a measure of the scale of the problem in one sector alone. The same arithmetic applies to the power grid, the banking system, the cloud and the networks that tie them together. The United States has spent decades building digital infrastructure and comparatively little time maintaining it. The bill for that neglect is now arriving in instalments — a router here, an update there, a ground stop, a locked bank account — and the instalments are getting closer together.
Rubio is probably right that no one attacked America this week. The harder question is whether America, left to itself, is attacking its own foundations one software update at a time. That is the question the next outage — and there will be a next one — will answer.
What to Watch Next
- The FAA briefing: Thirteen senators have asked for a technical briefing on the router failure by 23 November. Watch for whether the FAA pauses the wider router rollout, and whether the explanations from the Department of Transportation and L3Harris are reconciled.
- The full post-mortems: Microsoft, Apple, JPMorgan Chase and the NSA have each given a short cause. Watch for detailed public reports on what failed, how many users were affected and what will change — and for whether Congress demands them formally.
- Congressional hearings: Lawmakers from both parties are already demanding more answers on the outages. Watch for hearings in the Senate Commerce and Intelligence committees, and for whether the focus stays on each incident or turns to the concentration of critical services in a handful of companies.
- The Downdetector trend: Axios reported the past week brought the highest number of weekly outages in 12 months. Watch whether the rate falls back, or whether the autumn of 2026 marks a lasting step up in failure frequency.
- Apple’s remaining services: Consumer services were restored by Tuesday evening, but some developer and enterprise services took longer. Watch for Apple’s final status update and for any further detail on which internal update caused the failure.
- The AI patching race: Watch for further disclosures of vulnerabilities found with AI assistance — and for whether the pace of emergency patching produces more outages of its own in the weeks ahead.
- The next test of public trust: The 90-day blackout rumour was false, but it spread fast. Watch how officials communicate the next failure, whenever it comes: speed, detail and candour will decide whether the public believes the explanation or the rumour.
Sources
- Reuters — “Multiple tech systems have been crashing. Here is how to protect yourself” (17 November 2026)
- USA Today — “Apple services restored after outage impacted thousands” (17 November 2026)
- Intelligent CIO — “Technology failures expose the dangerous fragility of global digital infrastructure” (18 November 2026)
- Benzinga via TradingView — “‘Not Under Systemic Attack’: Marco Rubio Says ‘Zero Evidence’ Microsoft, JPMorgan and FAA Outages Due to Hostile Actors” (18 November 2026)
- 9to5Mac — “iCloud, App Store, and other Apple services hit by ongoing outage” (17 November 2026)
- Engadget — “iCloud and other Apple services are down in an apparent outage” (17 November 2026)