The office market is hollowing out
The post-pandemic era has brought a wave of under-utilized office buildings. With hybrid and remote work becoming the norm, many companies are not returning to full occupancy. For example, one report shows that as of Q2 2025 the U.S. office vacancy rate was roughly 20.6%. Vibe Sparking AI.
With fewer tenants and lower demand, many office buildings have become what real-estate professionals call “stranded assets” — structures built for a use that’s no longer in full demand.
The rise of data-centers and digital infrastructure
Meanwhile, demand for data centres — facilities that house large-scale computing, storage, AI workloads, cloud infrastructure — is booming.
- According to McKinsey & Company, global capital expenditure on data-center infrastructure (excluding IT hardware) is projected to exceed US $1.7 trillion by 2030 as AI, edge-computing and high-performance computing scale. McKinsey & Company
- The U.S. Census Bureau reports that employment in data-center establishments rose from 306,000 in 2016 to 501,000 in 2023 — a ~60 % jump. Census.gov
- The investment direction is clear: capital is shifting from “cubicles to racks”. Vibe Sparking AI
So: offices are losing relevance; data centers are gaining ground.
Office-to-data-center conversions: Repurposing the built-environment
Given the twin dynamics above, many developers and real-estate investors are pivoting. Instead of letting office buildings sit idle, they’re converting them into data centers. So much for that worry about the cause of the supposed upcoming ‘market crash’. Some key facts:
- A specialist article on “Office Parks to Cloud Parks” highlights how vacant office parks are being retro-fitted into data-center campuses. Data Centers
- In Germany, the real-estate firm Aroundtown announced plans to convert under-utilized offices into data centrer. Reuters
- An article (“Data centres are finding new homes in old buildings”) shows examples around the world of older offices or warehouses being converted to data-center use. jll.com.mx
Why it makes sense
- Location & infrastructure: Many offices already have power, water, fibre connectivity, and are situated near transportation or workforce clusters — which helps for data-centre purposes. Data Centers
- Lower cost, faster deployment: Repurposing an existing building is often faster (12-24 months) vs a greenfield build (36-48 months) and can reduce capital cost.
- Higher revenue per square foot: One developer noted that converting to high-density compute can drive “$400 of monthly revenue per square foot” versus far less for conventional office tenancy. Zak Kostura
The practical hurdles
Of course, retrofitting offices into data centers is not trivial:
- Offices often need structural reinforcement, upgraded power and cooling infrastructure, and fiber/cable interconnects. Data Centers
- Regulatory and utility approval can be challenging: e.g., approvals for power supply can delay transformations. heise online
- Not every office building will be a candidate — some buildings lack sufficient floor loads, cooling, or power connectivity. (See developer comments in industry forums.) Reddit
What this means for white-collar jobs
Here is where the “AI is literally taking white-collar jobs” piece comes in — and yes, it’s partly metaphorical but with real consequences.
1. Automation and AI substitution
Research shows that AI and automation are increasingly able to perform many tasks that were once the preserve of office professionals:
- A recent bibliometric analysis found that between now and 2029, the U.S. may lose more than one million jobs in office and administrative support occupations due to automation/AI substitution. arXiv
- Industry commentary points out that AI is first replacing outsourced/off-shore roles and then moving into internal white-collar roles. Vibe Sparking AI
2. The physical manifestation: Less office space needed
As more work is done digitally, remotely, or via AI-enabled systems, the need for large concentrations of people in office buildings declines.
- High office vacancies mean fewer workers are physically present.
- At the same time, the rising need for compute (and thus data-centers) means space that once held people now holds machines.
So there is a layering of trends: Fewer white-collar jobs in physical offices → more vacant space → conversion to non-people-centric infrastructure (data centers) supporting AI and compute. This is almost symbolic of a structural shift from “people doing work in offices” to “machines doing work in facilities”.
3. The job-creation mismatch
One angle worth noting: Data centers, while they require big build-outs and peaks of construction employment, often employ far fewer people once operational compared to what an office full of workers would house.
- A Reddit thread notes one data-center development as “a job-creation bust” compared to typical expectations. Reddit
- Thus, you could argue: As offices vanish (and white-collar jobs decline or move remote/automated), the replacement asset (data center) doesn’t create equivalent numbers of human jobs — reinforcing the sense of “machines replacing human labor”.
4. Wider implications
- Real-estate value shifts: Properties suited for people (offices) lose value; those suited for machines (data centers) rise in relevance.
- Skills pivot: The workforce needs to move from traditional office-based roles to more technical skills (AI oversight, data-center operations, infrastructure management).
- Geographic/structural change: Data centres impose different demands on infrastructure (power, cooling, fiber), which affects locations and communities differently than offices did.
Why this matters — and what to watch
Why it matters
- For workers: If your role is process-driven, administrative, or hybrid/remote friendly, it may be at higher risk of automation or relocation away from traditional office geography.
- For real-estate investors: Office portfolios may carry rising risk of vacancy/stranded assets, while data-center-ready properties or conversions may offer upside.
- For cities & communities: Office-centric downtowns may see long-term decline unless repurposed; and data-center siting raises infrastructure, energy and zoning questions.
- For policy-makers: Issues around workforce transition, energy use (data centers consume lots of power), and urban land use need more attention.
What to watch
- Vacancy rates in major office markets: Continue to monitor how fast offices empty out or convert. E.g., some U.S. markets are already showing > 18 % vacancy. CBRE
- Number and scale of office-to-data-center conversions: How many old office buildings are being repurposed? What is the geographic spread?
- AI substitution in white-collar labor: How many roles are being automated? Which job categories are most impacted? The research indicates office & administrative support are next.
- Data-center employment vs job creation: How many lasting jobs do these new data centers provide compared to traditional offices? Is the community benefit meaningful?
- Infrastructure/energy constraints: Data centers impose heavy loads for power and cooling. Can local grids handle this? For example, one developer noted power approvals were the biggest hurdle. heise online
Conclusion
In summary:
- The traditional office market is under pressure — fewer people working onsite, more vacancies.
- Meanwhile, demand for data-centers — driven by AI, cloud, digital services — is surging.
- This combination is leading to a physical repurposing of office space: Empty offices being converted into data centers.
- In parallel, AI and automation are quietly but steadily eating into white-collar roles, especially those that are process-heavy or remote-friendly.
- The result? The landscape of “work” is shifting. Not only are jobs changing, but the very places where work happens (offices) are being replaced by physical infrastructure for machines.
- That means for many white-collar workers: it’s time to rethink skills, roles, and how value is created — because the machines are already moving into our spaces.