Europe’s Old Technology Companies Are Becoming the Unexpected Winners of the AI Boom

The artificial-intelligence boom was expected to reward the glamorous laboratories building the world’s smartest models.
An unexpected group is also beginning to win:
Europe’s established software, consulting and cloud companies.
SAP, Capgemini, Sopra Steria and OVHcloud have reported stronger demand, faster growth or improved financial outlooks as large organisations move from experimenting with AI to deploying it inside real operations. �
Reuters
The reason is becoming clear.
Accessing an AI model is relatively easy.
Connecting that model safely to decades of company software, permissions, data and regulations is much harder.
The Experimentation Era Is Ending
During the first wave of generative AI, many companies ran small pilots.
Employees used chatbots to:
Draft emails
Summarise documents
Generate marketing ideas
Test code
Produce meeting notes
These experiments were useful but often disconnected from the organisation’s core systems.
The next stage is more difficult.
Businesses want AI to work inside:
Finance
Procurement
Human resources
Supply chains
Healthcare records
Manufacturing
Customer databases
Defence systems
Critical infrastructure
These environments contain sensitive data, specialist software and complicated rules.
A chatbot cannot simply be dropped into the middle and given unlimited access.
Why Older Technology Companies Have an Advantage
Established technology and consulting firms already understand the messy reality of large organisations.
They know that enterprise systems are rarely clean.
A global company may contain:
Software written decades ago
Several cloud providers
Custom internal applications
Inconsistent databases
Complex staff permissions
Different national regulations
Acquired companies using incompatible systems
AI must operate across this fragmentation without exposing confidential information or breaking important processes.
Companies such as SAP and Capgemini have spent years helping organisations integrate difficult technology.
AI creates a new and valuable version of the same problem.
The Numbers Behind the Shift
SAP reported a 26% increase at constant currencies in its cloud backlog, reaching €22.9 billion.
Capgemini raised its annual growth target after bookings increased by 9.2%, while Sopra Steria improved its outlook after organic growth accelerated to 5.3%.
OVHcloud’s public-cloud revenue rose by 20.2% during its third quarter. �
Reuters
These figures suggest that companies are beginning to spend real money on the infrastructure and professional work required to make AI operational.
The model may attract the headlines.
Integration collects the invoice.
Why One AI Model Will Not Run Everything
Large organisations are unlikely to rely on one model for every task.
They may use:
One model for coding
Another for customer communication
A private model for confidential data
A specialist model for medicine or engineering
A low-cost model for routine automation
A powerful frontier model for complex reasoning
The challenge becomes orchestration.
Which model receives the task?
What information may it access?
Who verifies the result?
How is the decision recorded?
How does the system comply with local law?
Companies capable of answering these questions may capture more practical value than the laboratories competing over benchmark scores.
Europe’s Sovereign-Cloud Advantage
European organisations are increasingly concerned about where sensitive data is stored and which foreign laws may apply.
This is especially important in:
Defence
Aerospace
Healthcare
Government
Critical infrastructure
Airbus plans to run around 70 critical applications on Scaleway, a French-controlled cloud provider, by the end of 2028. It is also working with French AI company Mistral. �
Reuters
The attraction is not merely European patriotism.
Companies want greater control over security, jurisdiction and regulatory compliance.
European cloud firms may therefore benefit from the demand for AI environments protected from certain forms of extraterritorial access.
Why the Story Went Viral
The narrative of the AI industry has focused on disruption.
New companies would supposedly replace slow established businesses.
This story suggests that old experience still matters.
The winners may not only be the companies inventing intelligence.
They may be the companies that understand how to place intelligence inside banks, hospitals, factories and governments without causing chaos.
The boring plumbing of AI may become one of its most profitable layers.
The Psychology of Enterprise Adoption
Businesses often become excited by demonstrations.
A model writes a beautiful report, produces code or answers a difficult question.
But adoption stalls when employees ask practical questions:
Can we trust it?
Can it access the correct data?
Will it expose customer information?
Who approves its actions?
What happens when it is wrong?
Can auditors reconstruct the decision?
Will the system work with existing software?
Innovation enters an organisation emotionally through excitement.
It survives operationally through trust.
Mary Chuks’ Perspective
This story proves the difference between having AI and building an AI system.
Anyone can subscribe to a model.
The business value comes from designing the workflow around it.
That includes:
Data
Governance
Human approval
Security
Measurement
Integration
Accountability
The next major AI opportunity is not necessarily to build another chatbot.
It is to help organisations make several forms of intelligence work together.
That is exactly where PublisherAI 360, Practical AI 360 and multi-model orchestration thinking belong: the application layer where businesses pay for outcomes rather than model hype.
Practical Takeaways for Businesses
Map the workflow before selecting a model.
Use different models for different risk levels.
Keep sensitive data inside controlled environments.
Define human-approval points.
Measure results against business outcomes.
Maintain logs and audit trails.
Train staff before expanding access.
Avoid locking the entire organisation into one provider.
Integrate gradually rather than automating everything at once.
Treat governance as part of the product—not paperwork added later.
Conclusion
Europe may not currently dominate the frontier-model race.
It possesses something equally valuable: companies that know how complex organisations actually work.
The AI boom is moving from impressive demonstrations to difficult deployment.
And deployment is where established expertise becomes profitable again.
Original Source and Further Reading
Primary analysis: Reuters, Leo Marchandon, “Europe’s Established Tech Firms Emerge as Unexpected AI Winners,” published 5 August 2026. �
Reuters