David Brooks
The €100M reality check: What France's ban actually costs taxpayers
France's government projects €25 million in annual savings by ditching Teams and Zoom licenses. The press releases tout "digital sovereignty" and "data protection." What they don't mention: the real cost of migration will likely hit €80-100 million in year one alone.
I've covered enough government tech projects to recognize the pattern. The €25M figure only accounts for software licenses. It ignores training, data migration, system integration, and productivity loss. Germany's similar sovereign cloud transition in 2022 cost 4 times the projected savings in the first year. Three years later, 40% of German agencies still use hybrid solutions because the sovereign platforms couldn't deliver enterprise functionality.
Let's break down the hidden costs France isn't advertising:
Training 2.5 million employees: If each worker needs just 2 hours of training at an estimated €30/hour labor cost, that's €150 million in training expenses alone.
Data migration: Years of chat histories, shared files, meeting recordings. Microsoft charges for bulk export, and the technical process requires specialized teams. Conservative estimate: €10-15 million.
System integration: Corporate email, calendars, ERPs, CRMs. Each custom integration costs between €50K and €500K depending on complexity. With hundreds of ministries and agencies, this scales rapidly to €20-30 million.
Productivity loss during transition: Germany reported a 15% productivity drop in the first 6 months of their sovereign cloud migration. Translated to salaries, that represents hundreds of millions in lost time.
The elephant in the room is that Tchap, France's existing sovereign messaging platform launched in 2019, has a 23% adoption rate after 5 years. 77% of potential government users actively avoid it. Forcing 2.5 million more people onto platforms that their colleagues have been rejecting for half a decade isn't a digital sovereignty strategy—it's institutional denial.
Tchap's 5-year failure: Why 77% of government employees refuse to use it
500,000 active users out of 2.2 million potential. That's Tchap's real adoption rate after five years in production: 23%. The GitHub repository has 87 stars, 23 contributors, and the last commit was 4 months ago. Open issues: video call crashes, phantom notifications, broken device sync.
Capterra: 3.2 out of 5 stars, 47 reviews. Recurring complaints: "No stable group video calls," "Missing basic integrations," "Unintuitive interface." Microsoft Teams, by comparison, has 4.5 out of 5 stars with over 15,000 reviews. The gap isn't marketing—it's technical maturity.
When a product has been in market for 5 years and only captures 23% of its captive audience (government employees who have no alternative), the problem isn't communication. It's the product itself. Forcing 2 million more people to use a tool their colleagues have been avoiding for years isn't digital sovereignty—it's ignoring user reality.
Why does the resistance persist? Because productivity doesn't care about flags. A civil servant coordinating a video conference with 15 different ministries can't afford to lose 20 minutes troubleshooting technical issues that Teams solves in 3 clicks. And if the government hasn't closed that gap in 5 years with Tchap, what makes anyone think Osmose will succeed in under 12 months?
12 urgent job openings: The Zoom replacement that doesn't exist yet
On LinkedIn, France's DINUM has 12 open positions for Osmose: 7 DevOps roles, 3 Product Managers, 2 Security Engineers. This isn't expansion of a mature product—it's emergency hiring to build something that should already be finished if they're really migrating 2.5 million users before year-end.
I haven't had direct access to Osmose's internal metrics, so I'm relying on public signals: urgent LinkedIn postings, private repository, absence from Gartner analysis. After years covering government tech projects, this pattern is clear: they hire urgently when they should be in maintenance mode. It's the clearest signal that something isn't ready.
Osmose's GitHub repository is private, which is already concerning for a platform marketed as "sovereign and transparent." There's no public documentation on its ability to scale. Tchap, which does have a public repo, never managed to support 500K concurrent users without latency issues. Osmose needs to handle 5 times that capacity, with high-quality video calls, smooth screen sharing, and meeting recordings.
Zoom handles 300 million daily users without breaking a sweat. Teams supports 280 million. Osmose doesn't even appear in Gartner's Magic Quadrant for Enterprise Communication. Element (based on the Matrix protocol) does appear, in the "Niche Players" category. That indicates the most mature European sovereign tools are barely considered enterprise-ready by analysts.
The inevitable analogy: Healthcare.gov, the Obama administration's health portal in 2013. Launched with fanfare, collapsed on day one due to capacity issues, and cost $2.1 billion to fix. Osmose faces a similar risk: if 2.5 million employees attempt to connect simultaneously in January 2027 and the platform crashes, the reputational damage (and technical rescue cost) will be massive.
The Cloud Act excuse: Legal cover for industrial protectionism
The US Cloud Act is a real legal risk, but France is using a sledgehammer to kill a fly.
| Aspect | Cloud Act (US) | GDPR (EU) |
|---|---|---|
| Jurisdiction | Extraterritorial (applies to US companies globally) | Territorial (applies to EU citizen data) |
| Notification requirement | Doesn't require notifying data subject | Requires notification except national security exceptions |
| Sensitive data protection | National security prevails | Fundamental rights prevail |
| Recourse mechanisms | Limited for non-US citizens | Extensive for EU citizens |
The Clarifying Lawful Overseas Use of Data Act (Cloud Act) of 2018 allows the US government to request data stored by American companies, regardless of where the servers physically reside. This directly conflicts with European GDPR, which prohibits data transfers without adequate safeguards.
The EDPB warned in 2025: the Cloud Act can violate GDPR "in certain cases." But there's a nuance France omits: Microsoft offers EU data residency with government non-access contracts. Teams can be configured so no data leaves European servers, and Microsoft has legally committed to challenge US government orders that violate GDPR.
So why didn't France negotiate these safeguards instead of banning the platforms? Because digital sovereignty isn't just technical—it's industrial. This move benefits OVHcloud, Scaleway, and other French cloud providers competing against AWS and Azure. The Cloud Act is the perfect legal argument to justify a decision that also responds to industrial policy.
I don't criticize the intent: protecting sensitive state data is legitimate. But deploy an alternative that matches or exceeds Teams/Zoom functionality first. Otherwise, this is protectionism disguised as cybersecurity.
When sovereign clouds collide: Europe's looming interoperability crisis
OVHcloud published a post celebrating the measure. Scaleway, France's second-largest cloud provider, did the same. And they're right to celebrate: this is a massive opportunity to capture government contracts that previously went to Microsoft Azure or AWS. But there's a risk nobody's discussing publicly: Europe's digital balkanization.
If France develops Tchap/Osmose, Germany has its Sovereign Cloud, Italy launches its own alternative, and Spain follows suit, how do these systems collaborate with each other? The Matrix protocol (open-source, federated) is a possible solution, but it requires all countries to adopt the same standard. France has already demonstrated it prefers proprietary national solutions over open European standards.
The nightmare scenario: a French official can't video call a German colleague because Osmose and Sovereign Cloud aren't interoperable. Both countries have invested hundreds of millions in their sovereign platforms, so neither will back down.
Neither will back down.
The result: loss of efficiency in cross-border collaboration precisely when Europe needs more coordination (defense, energy, migration).
The European Commission has tried to prevent this with initiatives like GAIA-X (a pan-European sovereign cloud project), but it has advanced slowly due to governance disagreements. France has historically been the country most reluctant to cede control over digital infrastructure to supranational bodies.
I support European digital sovereignty, but it must be European, not national. If each country builds its own digital Tower of Babel, the only winner will be the United States, which will continue to have globally interoperable platforms while Europe fragments into incompatible walled gardens.
The opportunity for European SaaS startups is real: government contracts, subsidies, market protection. But if they don't coordinate under common open standards, this movement won't strengthen Europe—it will just make it slower and more expensive.
The bottom line: Forcing adoption without fixing the product
After years covering the enterprise sector, this pattern repeats: governments deciding by decree what should be decided by organic adoption. Tchap has been in market for 5 years and hasn't convinced even a quarter of its captive audience. Osmose hasn't even demonstrated it can scale.
Digital sovereignty is a legitimate goal. But the tool to achieve it isn't prohibition—it's investment in alternatives that truly compete. If Tchap and Osmose were objectively better than Teams and Zoom, public employees would adopt them voluntarily. They're not doing that. And forcing them won't change product quality; it will only generate resistance and productivity loss.
What France needs isn't a forced migration plan. It needs a product development plan that prioritizes usability, stability, and interoperability. And if that takes 3 more years, then wait 3 more years. Because launching an immature platform to 2.5 million users and crossing your fingers isn't digital strategy—it's technical roulette with public funds.
