Palantir Got $3.9 Million to Watch Federal Workers. It's Just the Start.
The same company that built ICE's deportation tracking system is now monitoring federal employees' return to the office.
A no-bid contract to track USDA employees’ return to office shows how workplace surveillance becomes federal infrastructure: quietly, contract by contract, agency by agency.
“Real-time analytics to optimize space utilization and employee seat assignments.” — USDA contract language, via The Register.
Three things to know:
USDA gave Palantir a no-bid contract, starting at $3.9 million, to track federal employees’ return-to-office compliance.
The contract requires the capability for continuous behavioral monitoring, not just attendance, and reporting confirms the same plan extends to Social Security and Veterans Affairs.
The GAO already found this kind of monitoring increases worker stress and anxiety, while the agencies meant to guard against that harm have been rolling back their own oversight since January 2025.
I read that sentence in an actual government contract; the bureaucratic flatness of it is what got me. “Employee seat assignments.” Not a euphemism written to obscure something. Just the plain language of a tool that watches who shows up, where they sit, and whether they’re following the rules.
The contract is real. It was first reported in March 2026, when The Lever and then The Register surfaced the sole-source justification USDA had posted to SAM.gov. The award itself started on May 1, 2026. The Department of Agriculture obligated an initial $3.9 million to build a tool tracking employees’ return to the office. The original justification cited a potential ceiling of $75 million, but the federal spending disclosures for the actual award show a confirmed growth path that tops out at closer to $13.3 million for this fiscal year, which runs through September 30, 2026. Worth being precise about which number is which: $75 million is what USDA’s justification document floated as a ceiling, $13.3 million is what the obligated spending data actually shows so far. No competitive bidding. No other vendors considered. The justification was that only Palantir could deliver it fast enough.
That alone would be a story about one agency and one compliance tool. It isn’t one story. It’s the opening move.
Where the Line Actually Is
Employers have always tracked attendance. Badge swipes, calendar systems, and a manager noticing an empty desk- none of that is new, and none of it is inherently sinister. If “real-time analytics on space utilization” just meant counting occupied desks so an agency could right-size its office footprint, that’s a boring facilities question, not a surveillance story. Taxpayers have a legitimate interest in knowing whether employees mandated back to the office are actually showing up.
That is not what this contract describes. A badge swipe tells you whether someone is in the building. The USDA solicitation requires the capability for “continuous compliance monitoring” with automated alerts “upon detection of “threats or anomalies,” language confirmed independently by two outlets quoting the contract notice directly down to the same slightly garbled grammar, a strong sign it’s lifted verbatim rather than paraphrased. Here’s the distinction worth being precise about: that confirms the contract requires this capability to exist. It does not yet confirm, based on current reporting, that the system is actively running this way on USDA employees today. A contracted requirement and a deployed system are not the same claim, and I’m not going to blur them.
Even as a required capability, though, it’s a categorically different tool than a badge swipe. That is not attendance. That is the contracted capacity for an ongoing behavioral assessment, built by a company whose core product is anomaly detection and pattern flagging, not facilities management. The distinction isn’t “tracking is bad.” It’s between a binary record of presence and a system built to continuously score behavior, generated by a firm whose entire business model is finding what looks unusual in someone’s activity. One tells your employer where you are. The other is built to tell your employer what it thinks you’re doing, all the time, without you ever learning what tripped the flag.
The tool doesn’t need to fire anyone to change behavior. Knowing the capability exists is enough.
The Plan Was Always Bigger Than One Agency
The USDA return-to-office tool sits within a much larger contract umbrella, a $300 million “One Farmer, One File” initiative that consolidates farmer data across agencies and retires roughly 1,000 contractors in favor of automation. The workforce tracking piece isn’t a standalone product. It’s a feature riding alongside a much bigger data consolidation effort.
And it doesn’t stop at the USDA’s door. Reporting confirms that the same contract also kicks off a stated plan to track workers at the Social Security Administration and the Department of Veterans Affairs. I want to be precise about what’s documented here versus what’s a structural inference. Documented: USDA has the contract, and reporting confirms that SSA and VA are next. Inference, but not a wild one: once a federal vendor builds a tool that works at one agency, GSA contract-sharing rules make it trivial to roll the same tool out everywhere else. That’s not paranoia. That’s how federal procurement actually works. If this tool proves itself at USDA, there’s no structural reason it stops at three agencies. The federal workforce is roughly 2 million people.
This isn’t Palantir’s only federal expansion this year, either. The company also won a $30 million ICE contract to build a system tracking undocumented immigrants and self-deportations, its largest single award from that agency since 2011. And 10 Democratic lawmakers have demanded answers from CEO Alex Karp about Palantir’s alleged role in building a searchable “mega-database” layered atop every IRS database, raising the possibility that tax records could become accessible across agencies far beyond tax administration. The total federal contract value this year has crossed $900 million. The company itself is projecting a more than 60 percent jump in revenue for 2026, driven substantially by government work.
Different agencies. Different stated purposes. Same vendor, same underlying architecture, same year.
Running This Through the Four-Part Test
I built a framework for this site because I kept seeing the same pattern across completely different industries, and I needed a way to test whether something is actually coercive or just unpleasant. The test has four parts: a voluntary trade, a genuine benefit received, data weaponized back as a financial or material instrument against the same person, and an exit cost engineered to be prohibitive.
Federal employees pass the first two parts cleanly. Taking a government job is voluntary, and the paycheck is a genuine benefit. The third part is where the distinction above earns its keep: whether the capability is fully active yet. A system built to generate a continuously updated anomaly record is a much sharper instrument than an attendance sheet, and it would be compliance officers, not the employee, deciding what that record means. The fourth part, exit cost, is what separates this from a corporate badge-swipe system. You can quit a private employer over workplace surveillance and look for another job. Quitting a federal position in a labor market already strained by agency cuts and hiring freezes is a materially different decision. The cost of leaving isn’t abstract. It’s a household budget.
There’s a mental health dimension to this worth naming directly rather than just implying. The Government Accountability Office reviewed 122 studies on digital workplace surveillance and found that it can increase stress, anxiety, and depression, and that it risks employees’ physical health and safety by pushing them to move faster to meet productivity metrics. Eight of those studies found that constant surveillance negatively affects mental health specifically, with continuously monitored workers describing feeling anxious and demoralized. Researchers studying electronic monitoring call this stress proliferation: the damage often comes less from the monitoring itself than from the secondary stressors it sets off afterward, the second-guessing, the anticipatory dread, the sense of being evaluated by a process you can’t see or appeal. Layer that onto a workforce that can’t easily leave, and the exit cost from the fourth part of the test stops being just financial. It becomes a captive population for a documented mental health risk, with nowhere else to go.
That same GAO report contains one more detail worth sitting with. Several federal agencies, including the NLRB, the Department of Labor, the EEOC, and the Consumer Financial Protection Bureau, had previously issued guidance meant to protect workers from exactly these kinds of digital surveillance harms. Since January 2025, GAO has found that those agencies have either rescinded that guidance outright or are reviewing it to align with the current administration’s priorities. The same government that is now contracting for a tool capable of continuous behavioral monitoring of its own employees is, in parallel, walking back the rules that would have governed how such a tool is used.
What I’m Watching Next
I don’t think anyone designed this as a single coordinated plan in a back room. I think a contracting officer at USDA had a return-to-office mandate to enforce, a vendor with existing federal relationships and a promise of fast turnaround, and a procurement process that rewards exactly that combination. Nobody had to conspire. The incentive structure did the rest. That’s the same mechanism I keep finding everywhere else on this site, just wearing a different agency seal this time.
What I’m watching for now: whether reporting or a records request confirms the continuous-monitoring capability is actually running, not just specified in the contract; whether the SSA and VA contracts materialize on the timeline reported; whether GAO or congressional oversight gets a real look at the tool before it scales further; and whether “compliance monitoring” quietly expands past return-to-office into something with sharper teeth.
What You Can Do
If you’re a federal employee, ask your union representative directly whether your agency has any pending Palantir contract for workforce monitoring, RTO compliance, or “space utilization” tools. Get it in writing if you can.
Search USAspending.gov for “Palantir” and your agency name. The contract data is public.
File or support a FOIA request asking whether the continuous-monitoring capability described in the contract is actually deployed, since the contract’s requirements and what’s currently running may not be the same thing.
Contact your representatives on the House Oversight Committee, which is actively investigating Palantir’s federal contracts, and ask them to specifically request the contract terms for the USDA, SSA, and VA.
Follow The American Prospect and Federal News Network for ongoing reporting. Both have stayed on this story closer than most outlets.
If your household includes a federal employee, have the conversation now about what “continuous compliance monitoring” might mean for them before it’s fully built out, not after.