On Thursday, 12 November, a critical communications system
underpinning modern capitalism began to fail. Outlook and other components of
Microsoft 365 became inaccessible or unreliable across much of the world, with
Teams and Copilot functions also affected. Government agencies, health
services, courts, police forces, and private corporations relied on a single
American technology corporation to restore communications over which they had
virtually no direct control.
The immediate cause was technical. Microsoft later said the
incident began with an earlier network capacity upgrade and became apparent
after a subsequent network change. The company rejected suggestions that a
cyberattack or artificial intelligence agents were responsible. The outage's
significance is heightened by the fact that it did not require intervention by
a hostile state, criminal organisation, or sophisticated computer attack. A
failure within a single corporation's infrastructure was sufficient to disrupt
institutions across multiple countries.
The outage revealed a fundamental contradiction in the organisation
of contemporary technological society. Communications systems that serve
hundreds of millions of people and are increasingly essential to hospitals,
government departments, and other public services have been consolidated within
a small number of privately owned corporations. While technology has become
more international and social in character, ownership remains private, and
strategic decisions are ultimately driven by profitability, competition, and
shareholder interests.
Problems became apparent on Thursday afternoon when users
reported being unable to access Exchange Online and Outlook. Microsoft
initially described the disruption as affecting the Americas, before
acknowledging that users in other regions were also impacted. Outlook was
central to the failure, but other Microsoft 365 services were disrupted as
well, along with external organisations whose systems relied on Microsoft's
infrastructure.
Microsoft's explanation evolved as engineers worked to
understand and contain the failure. Servers became unexpectedly unavailable, increasing
the load on the remaining infrastructure. Initial restoration efforts achieved
only limited success, as recovered capacity was insufficient to achieve
immediate stability. Ultimately, Microsoft attributed the incident to a problem
introduced during an earlier network capacity upgrade, which became apparent
after a subsequent network change.
Currently, there is no evidence that Microsoft's job cuts or
AI investments directly caused the outage. Determining the precise chain of
technical and organisational responsibility requires significantly more
evidence than is currently available. However, this does not negate the broader
economic issue. The central question is not whether a specific redundant
engineer could have prevented this failure, but rather what occurs when
essential infrastructure is managed according to the investment priorities of
competing corporations.
The breakdown occurred less than three years after the July
2024 CrowdStrike incident, which demonstrated the vulnerability of extreme
concentration in global computing. A faulty security update crashed millions of
Windows computers and disrupted airlines, hospitals, banks, and businesses
worldwide. Although governments and corporations pledged to learn from this
event, the concentration of technological infrastructure that allowed a single
defect to have global consequences remained fundamentally unaddressed.
The Microsoft outage should not be regarded merely as
another instance of the CrowdStrike event. The immediate technical causes
differed. However, both incidents share the same underlying social structure:
large portions of global communications and computing infrastructure are
concentrated among a small number of corporations. As a result, a decision or
error that once affected only a single organisation can now propagate across
countries and continents.
Cloud computing has intensified this contradiction.
Centralisation brings enormous technical advantages. Computing resources can be
distributed internationally, immense quantities of information can be processed
and transmitted almost instantaneously, and sophisticated applications can be made
available to organisations that cannot build such systems independently. These
are genuine achievements representing an extraordinary development of
humanity's productive capabilities.
Under private ownership, however, technological integration
results in corporate dependence. Governments, which are responsible for
administering the lives of millions, increasingly operate as customers of
Microsoft, Amazon, and a few other technology companies. Hospitals, police
forces, courts, councils, and entire government departments therefore rely on
infrastructure whose fundamental operation is beyond their direct control.
The most significant consequences of the November outage did
not occur among office workers inconvenienced by the temporary loss of email,
but rather in areas where public administration had become dependent on
Microsoft's systems. Health organisations, police forces, government
institutions, and judicial bodies were affected internationally. The
interruption revealed how deeply ordinary communications software has become
embedded in essential social services.
It also exposed the increasingly artificial distinction
between "front-line" and "back-office" services. Hospitals
depend upon administrative communications to coordinate staff and information.
Police forces use digital systems to receive and distribute reports. Courts
depend upon them for official communications. Referrals, appointments, records,
and innumerable administrative processes depend on systems that seem secondary until
they stop working.
When these systems fail, their associated work does not
vanish. The burden shifts directly to workers. Nurses, administrative staff,
call handlers, doctors, police personnel, and IT engineers must set up
alternative arrangements, revert to telephones or manual methods, address
accumulated backlogs, and reassure members of the public who are unsure whether
their communications have been received.
This additional labour represents one of the largely
invisible social costs of corporate technology failures. Microsoft's financial
accounts do not adequately reflect these costs because much of the work is
performed elsewhere. Public-sector employees, local IT departments, and users
absorb the disruption while corporate engineers focus on repairing the central
infrastructure. As a result, the economic damage from the failure is socialised,
even though ownership and control remain private.
The outage also took place amid an unprecedented
international investment race in artificial intelligence. Microsoft is one of
the principal corporations allocating substantial resources to data centres,
computing capacity, and AI infrastructure, competing with Amazon, Google, Meta,
and other technology giants for a dominant position in what investors
anticipate will become a central market in the coming decades.
It would be inappropriate to interpret this coincidence as
an unsupported explanation for the outage. No evidence currently indicates that
AI expenditure caused the November failure, just as Microsoft's workforce
reductions cannot automatically be blamed for a specific network malfunction.
The significance lies in the priorities established by an economic system where
investment is allocated based on anticipated profitability rather than
democratically determined social needs.
Technology analyst Rupert Goodwins raised this broader issue
during the disruption, contending that Microsoft's intense focus on artificial
intelligence risked marginalising mature services such as email and Teams. He
also highlighted the departure of experienced engineers and the risk of losing
accumulated institutional knowledge. While this argument does not establish the
technical origin of the November failure, it sheds light on the corporate
environment in which essential infrastructure is maintained.
For investors, artificial intelligence represents a
potentially vast new field for capital accumulation. Outlook and Exchange are
mature services whose operation may generate limited enthusiasm in financial
markets. However, from a societal perspective, providing reliable
communications to hospitals, ambulance services, courts, and government
departments is likely more important than developing the next commercially
promising AI application.
The market cannot resolve this contradiction because it
generates it. Capital flows towards anticipated returns. Labour simultaneously
appears as an expense to be reduced. Experienced engineers, administrators and
support workers remain costs on corporate balance sheets even when the networks
they maintain have become indispensable to the functioning of modern society.
The institutions of the capitalist state already acknowledge
the extraordinary degree of concentration involved. Britain's Competition and
Markets Authority concluded that Amazon and Microsoft have significant market
power in cloud services. It also identified obstacles to switching providers or
operating effectively across multiple clouds, including interoperability issues
and Microsoft's software licensing practices.
Notably, the Competition and Markets Authority (CMA)
describes cloud computing as "core infrastructure for the UK
economy." This characterisation raises a more fundamental question
than the regulator may intend. If cloud computing constitutes core national
infrastructure, and if hospitals, government institutions, and businesses
cannot function effectively without it, we must ask why its organisation,
staffing, and investment priorities should remain the private concern of
Microsoft, Amazon, and other technology corporations.
The conventional answer is increased competition. Regulators
propose making it easier for customers to move between cloud providers by
lowering barriers to transferring data, increasing interoperability, and
challenging restrictive licensing arrangements. Some of these changes could
undoubtedly improve resilience and make individual organisations less dependent
upon a particular provider. But competition between monopolies does not
constitute democratic ownership. Dividing an organisation's dependency between
Microsoft and Amazon may reduce the consequences of one supplier's failure, but
society remains a customer, not the owner, of its essential infrastructure. The
basic relationship therefore persists.
Perhaps the most politically revealing feature of the
November failure was accordingly the position occupied by governments
themselves. Ministers could investigate the effects upon public services,
engineers could cooperate with Microsoft, and officials could communicate the
corporation's assurances that the event was not a cyberattack. But governments
themselves could not repair Microsoft's infrastructure.
They had to wait for Microsoft to diagnose its network,
implement repairs, and explain the failure. Governments commanding intelligence
services, armed forces, and enormous administrative bureaucracies have
nevertheless allowed essential components of their civilian communications
systems to depend on infrastructure controlled by private technology
corporations.
This dependence partly explains the growing discussion of
"digital sovereignty", particularly within Europe. Yet reducing
European dependence upon American technology companies is not equivalent to
establishing democratic control over technology. Replacing Microsoft with
nationally favoured European corporations would merely reorganise private
technological dependency along geographical lines.
Under conditions of increasing protectionism, economic
nationalism and geopolitical conflict, the demand for technological sovereignty
can readily become part of the struggle between competing capitalist states.
Workers have no fundamental interest in choosing between American, European or
Chinese technology monopolies. The question facing them is who owns and
controls the infrastructure itself.
Government policy nevertheless remains confined to
regulation. Every major technological breakdown produces the same declarations.
Resilience must be strengthened, updates tested more carefully, contingency
arrangements improved, suppliers diversified and lessons learned.
Investigations are promised, and regulators consider modifications to existing
market structures. Then another failure occurs.
The problem is not that technical failures can be
permanently eliminated. Any sufficiently complex computing system will
eventually experience hardware problems, programming errors, configuration
mistakes or unforeseen interactions. Public ownership would no more abolish
computer failures than public ownership of transport could abolish mechanical
breakdowns.
The decisive question is who controls the conditions within
which those failures take place. Who determines staffing levels, maintenance
budgets and investment priorities? Who decides whether experienced engineers
are retained? Who determines the level of redundancy built into critical
systems? And who possesses the information necessary to investigate failures
affecting millions of people?
Under the existing system, these questions are ultimately
subordinate to corporate property. Microsoft's engineers may possess immense
technical expertise, but they do not determine Microsoft's overall economic
strategy. Decisions involving capital expenditure, employment and corporate
priorities are conditioned by competition for markets, profitability and
shareholder returns.
The contradiction therefore is not between competent
engineers and unreliable technology. It lies between an immensely sophisticated
collective productive apparatus and the social relations that control it.
Digital networks are created and maintained through the labour of enormous
numbers of people, yet decisive control over those networks remains
concentrated within a tiny number of corporations.
The November breakdown belongs to a wider series of
technological infrastructure failures. CrowdStrike in July 2024 demonstrated
the international consequences of distributing a defective update across
millions of computers. Amazon Web Services subsequently experienced major
disruption, while Microsoft has suffered repeated service problems. The
immediate technical explanations have varied considerably, but the underlying
social relationship has remained remarkably consistent.
In each case, innumerable organisations depended upon
technological infrastructure concentrated within a handful of companies. When
something failed, customers depended on those companies to diagnose their
systems, repair their infrastructure, and explain their failures. Corporate
ownership survived while workers and the wider population absorbed the
consequences.
This is why describing such events simply as
"outages" is inadequate. The word isolates the immediate technical
fault from the economic structure that gives it extraordinary social
significance. A server malfunction is a technological event. Hundreds of
millions of people becoming dependent upon the functioning of a relatively
small number of privately controlled systems is a social relationship.
Cloud computing itself represents an extraordinary form of
socialised production. Engineers, programmers, electricians, construction
workers, semiconductor workers, scientists and telecommunications workers
across numerous countries contribute to systems no individual corporation could
have created independently of the accumulated development of modern science and
technology.
The internet emerged from decades of scientific research and
enormous public investment. Contemporary computing rests upon generations of
work in physics, mathematics, telecommunications, semiconductor manufacturing
and software engineering. Yet the resulting infrastructure confronts society as
the private property of corporations.
Microsoft 365 expresses this contradiction starkly.
Hospitals, courts, governments and businesses can communicate through
interconnected networks spanning the globe. The technology demonstrates
humanity's extraordinary capacity to integrate productive and intellectual
activity internationally. Yet a network change within one corporation can
simultaneously interrupt those communications across national boundaries.
Capitalism has therefore created technological
interdependence without creating democratic control over interconnected
technology. Modern production and communication have taken on an increasingly
international and social character, while ownership of essential infrastructure
remains private and highly concentrated.
More careful testing, redundant systems, open standards and
improved interoperability are all worthwhile. Governments, hospitals and other
essential organisations plainly require reliable contingency arrangements when
centralised systems fail. But these measures cannot answer the political
question exposed by the events of 12 November.
Why should infrastructure upon which hospitals, courts,
emergency services and governments depend be privately owned at all?
Microsoft's technology is not the enemy, nor are the
engineers who created and maintain it. On the contrary, the global
communications systems operated by Microsoft and its competitors embody an
extraordinary technical and social achievement. They show what internationally
coordinated human labour and scientific development can produce.
The obstacle is subordinating this social achievement to
private ownership. Essential digital infrastructure should instead be treated
as a public utility, publicly owned and democratically controlled by the
workers who build and maintain it together with the population that depends
upon it.
Investment decisions could then be guided by social needs
rather than corporate rivalry. Reliability, adequate staffing, redundancy,
security and the preservation of accumulated engineering knowledge could become
explicit social priorities rather than costs continually weighed against profit
margins and investor expectations.
Such organisation cannot ultimately be achieved simply by
nationalising one corporation within an otherwise capitalist economy. Digital
infrastructure is already international. The workers producing it are
international, the networks themselves cross borders, and the populations
depending upon them extend across continents. The rational organisation of such
infrastructure must therefore also assume an international character.
The November 2026 Microsoft outage demonstrated in
concentrated form one of the central contradictions of contemporary capitalism.
Humanity has created communications technology capable of integrating society
on an unprecedented scale, yet control of that technology is increasingly
concentrated within corporations accountable principally to private capital.
The solution is neither nostalgia for the world before cloud
computing nor the nationalist project of replacing American monopolies with
European or other national competitors. It is to bring ownership into
conformity with the technology's already social character.
Infrastructure society depends on should be collectively
owned and democratically controlled. Until we resolve that fundamental
contradiction, another massive technological failure will not be an
inexplicable aberration. It will remain an ever-present possibility built into
a system that has placed the increasingly interconnected infrastructure of
social life under private monopoly control.
Bibliography
Al Jazeera. Heba Habib, “Microsoft hit by global outages:
What happened and why?” 13 November 2026.
Competition and Markets Authority. Cloud Market
Investigation: Final Decision Report. London: CMA, 2025.
Competition and Markets Authority. Update on Cloud
Actions: Covering March–October 2026. London: CMA, 16 October 2026.
Competition and Markets Authority. Microsoft's Business
Software Ecosystem: Strategic Market Status Investigation. 2026.
CrowdStrike. Preliminary Post Incident Review: Windows
Sensor Content Update. 2024.
Forbes. Kate O'Flaherty, “Microsoft Apologises for Outage
and Reveals Cause.” 13 November 2026, updated 14 November 2026.
Marx, Karl. Capital: A Critique of Political Economy,
Volume I. London: Penguin/New Left Review, 1976.
Marx, Karl. Grundrisse: Foundations of the Critique of
Political Economy. London: Penguin/New Left Review, 1973.
Braverman, Harry. Labour and Monopoly Capital: The
Degradation of Work in the Twentieth Century. New York: Monthly Review
Press, 1974.
Mandel, Ernest. Late Capitalism. London: New Left
Books, 1975.
Noble, David F. Forces of Production: A Social History of
Industrial Automation. New York: Oxford University Press, 1984.
Srnicek, Nick. Platform Capitalism. Cambridge:
Polity, 2017.
Wu, Tim. The Master Switch: The Rise and Fall of
Information Empires. New York: Alfred A. Knopf, 2010.
US House Committee on Homeland Security. Committee
materials on the July 2024 CrowdStrike global IT outage. September 2024.
