The facility is Meta's first in Oklahoma, its 28th in the US, and part of a broader push to build AI-optimized infrastructure across the country. For Tulsa, the implications go beyond construction jobs.
Meta broke ground this week on a new AI-optimized data center in Tulsa, by revealing the company's first facility in the state, its 28th across the US, and its 32nd globally. The project will support roughly 1,000 construction jobs at peak build-out and around 100 permanent operational roles once it's running.
Those numbers sound modest for a tech giant spending billions on infrastructure. But the real story is what happens around a facility like this — to the local power grid, the workforce pipeline, the regional tech economy, and the communities that suddenly find themselves adjacent to one of the largest AI buildouts in the world.
What Meta Is Actually Building
Meta's Tulsa facility is designed specifically for AI workloads. That distinction matters. Traditional data centers primarily handle storage and content delivery. AI-optimized facilities are built around dense GPU clusters that demand significantly more power and cooling per square foot. They're engineered for the kind of sustained compute that trains and runs large language models, recommendation systems, and the inference engines behind products like Meta AI.
The Tulsa campus fits into a broader infrastructure strategy Meta has been telegraphing for months. Meta framed its data center investments as critical to US economic power and national security in a March newsroom post, noting that the company has spent 15 years building relationships with data center communities through workforce development, water stewardship, and philanthropic programs.
For the Tulsa project specifically, Meta says it will invest a modest $25 million in local infrastructure improvements. That includes partnering on a workforce development program and assisting low-income Oklahomans with electricity and water bills, the company announced.
The bill-assistance commitment is worth noting. It's a tacit acknowledgment that large-scale data centers can strain and interupt local utilities. Meta has previously stated that it pays the full costs of its data centers' energy and water use so those costs aren't passed on to consumers, and that it funds new and upgraded infrastructure, in its March data center overview.
The Regional Tech Ecosystem Effect
A hyperscale data center doesn't just create jobs inside the building, it reshapes the economic geography around it.
When a company like Meta commits to a multi-year, multi-billion-dollar facility, it sends a signal to the broader tech supply chain. Networking equipment vendors, fiber providers, electrical contractors, HVAC specialists, and security firms all follow. Secondary service businesses, from commercial real estate to restaurants, tend to cluster around these sites.
For Tulsa specifically, the timing is notable. The city has spent the past several years actively courting tech investment, most visibly through programs like Tulsa Remote, which paid remote workers up to $10,000 to relocate. A Meta data center adds a different kind of anchor: not remote workers choosing a lifestyle, but physical infrastructure that locks in long-term capital expenditure and creates demand for specialized technical skills.
The 100 permanent jobs Meta cites will likely skew toward facilities management, electrical engineering, network operations, and physical security. These aren't software engineering roles, but they're skilled positions that pay well above regional medians. More importantly, they create a local talent base familiar with hyperscale infrastructure, the kind of workforce that attracts additional data center operators to the same region, research on data center site selection shows.
Let's examine the facts - Oklahoma's energy profile is part of the draw. The state has significant wind generation capacity and relatively low electricity costs compared to coastal markets. For AI workloads that can consume tens of megawatts continuously, power economics are often the single biggest factor in site selection.
The Uncomfortable Questions
Meta's data center ambitions don't exist in a vacuum. The company is simultaneously navigating serious questions about how it handles the data that flows through these facilities.
Facebook users opting into "cloud processing" have been inadvertently giving Meta AI access to their entire camera roll, The Verge found, a practice that raises pointed questions about consent and data boundaries. When Meta builds more AI infrastructure, it's building the capacity to process more personal data at greater scale. The Tulsa facility will be part of that pipeline.
There are also labor and oversight concerns. A Svenska Dagbladet investigation revealed that behind Meta's AI-powered smart glasses lies a hidden workforce of human reviewers who see intimate, often disturbing content captured by users — including bank details, nudity, and scenes where subjects appear unaware they're being recorded.
Workers described feeling uneasy about the work. These are the human costs embedded in the AI supply chain that a gleaming new data center represents.
And Meta's broader relationship with user trust remains strained. As we covered earlier this year, a Los Angeles jury awarded $6 million to a young woman who sued Meta and Google over childhood social media addiction, finding that both companies acted with "malice, oppression, or fraud." Hundreds of similar cases are pending. More AI infrastructure means more AI-driven engagement features, which means the same design incentives that led to those lawsuits will be operating at even greater scale.
These tensions don't invalidate the economic benefits of the Tulsa project. But they do complicate the narrative. A $25 million community investment is real. So is the question of what data, collected from whom, under what terms, will flow through the servers inside.
What Comes Next
Meta now operates 32 data centers globally, with 28 in the US. The Tulsa facility adds to a footprint that already spans multiple states and represents billions in cumulative investment.
The competitive context is straightforward: every major AI company is racing to secure power, land, and permits for new compute capacity. The companies that build fastest will have the most flexibility to train next-generation models and serve inference at scale. Meta's open-model strategy makes this particularly important — if Llama models are going to compete with closed alternatives from OpenAI and Google, Meta needs the infrastructure to iterate quickly.
For Tulsa, the next few years will test whether a hyperscale data center can genuinely catalyze a regional tech ecosystem or whether it remains an isolated industrial facility with limited local spillover. The workforce development programs and infrastructure investments Meta has committed to will be the early indicators.
