Having spent time working in UK healthcare tech, I never understood why everyone was lining up to throw buckets of money at Palantir. Quite apart from being obviously evil and so on, none of their solutions were actually very good.
Unfortunately, it's hard to escape the feeling that friends in high places, some lobbying and some er... reciprocal back scratching might have been instrumental.
See also senior staff at NHS England (or Digitial? can't remember) handing massive NHS compute contracts to AWS, and then leaving the civil service to become... an AWS employee.
The revolving door as it's known. That's part of it. Another is simply the lack of in-house talent, largely due to poor pay and conditions. It's a self-fulfilling prophecy to a certain extent. I'd love to work for NHS digital and make a difference, but all the interesting work is contracted out, so they can't keep the staff who are capable of building themselves. Also the recruitment process is terrible.
They’ve built a platform and sales pipeline optimized for selling data consulting into highly bureaucratic tech hostile orgs with data privacy concerns. All these factors apply equally to public health programs and the military, so it’s no surprise that they see success in both areas.
In the UK, long term sick leave with a mental health diagnosis is a way to be unemployed (ie. not working) but not have the disadvantages of that. There's the money difference: 530 pounds per month for sick leave, 338 or 425 pounds per month for unemployment. On long-term sick leave you get all advantages job seekers get (ie. "Universal Credit"), PLUS others (support for not being able to work, ESA, and support for extra living expenses due to long-term sickness, PIP). So if you don't want to work, long-term sick leave has many advantages, plus it's a big cost to employers. You don't have to look for work in long-term sick leave. In fact, nothing is expected at all (other than medical evidence).
2.8 million people are long-term ill, not participating in the workforce, not being economically productive at all, at least half due to being diagnosed with mental illness. The issue with this is that this is happening with full support of government employees, and even the courts cooperate to a lesser extent.
What the government is trying to do, in other words, is trying to kick people off, uh, let's say "mental disability", force them to work. And they need to do this without relying on government workers, because they often side with the people on sick leave.
Hence, Palantir. Being hated is a feature here, not a bug.
1. they aim to deliver product company margins with a consulting-heavy model.
2. your software purchase funds a cadre of "free" FDEs and deployment strategists who customize your install, build a bunch of data pipes/transforms, and talk to people to figure out what all the data means.
This could be a good deal if your tech org is not very competent at integrating your data, or if you have a sudden, short-term need. In the longer term, it's probably cheaper and more effective to develop a competent tech team, modernize the source data systems, and roll your own integration -- but that also requires leaders with long-term vision who are resistant to external hype and pressure.
I never understood why nation states pay outside companies for this stuff. You need the expertise to actually evaluate what you're getting anyway. Incentives are in no way aligned. At the state level you have the scale to do it in house.
If a senior government employee can get a very expensive Palantir contract approved, they have a good chance of a much better paid job at Palantir in the future:
Buy: you need expertise in contracts and knowing what you need.
Build: you need expertise in contracts, knowing what you need and also software development.
It's obviously easier to buy than build, especially for civil service roles where they can't attract the best developers due to political/ideological constraints.
Because "nation states" are not one making decision either. It's done by one specific career bureaucrat or group of them and even best of people who work on such positions usually choose it because of job security and stability.
Spending 10x more on IBM or Palantir can't get them fired, but trying to build something in-house their organization don't have competence for can get them fired.
And this is even if you don't take lobbying or corruption into account.
> It's done by one specific career bureaucrat or group
Almost all governments have a legally defined public procurement framework. If this is overridden, it's pretty much always by elected politicians, not by regular government employees.
Plausible deniability. "We paid £5 million for consultants who recommended this system, it's not our fault it turned out to be a steaming pile of crap that wasted £20 million and took 3 years".
if Palantir (or other consultancies) are friends with government decision makers (especially in the US) then spending more on a service is a feature, not a bug.
reads like Salesforce to me, ugh! Enterprises are paying so much to blatantly vendor-lock in themselves using hundreds of "Salesforce engineers". It's baffling to me.
I owned Salesforce setup with 4 engineers and 500+ licenses. I don‘t see how could I replace our SF setup with an in-house product on the same budget within reasonable timeline. We won local competition within a few years, because our sales could use good CRM from day 1 and our competitor, according to the rumors I heard, could not calculate properly sales agent commission. Vendor lock-in is not always a stupid thing. Sometimes it‘s the bet that wins you a market.
Zoom out a little though.
I've always felt the main reason
That most companies use Salesforce
Is that most companies use Salesforce.
I'll give you an example.
At a previous employer,
We used Google Analytics.
We paid for Google Analytics.
I feel positive that as a mid size company,
We shouldn't have paid for Google Analytics.
The free product with 50 events in GA4 should be plenty for us.
But why do we use Google Analytics in the first place?
Because everyone uses Google Analytics.
I agree that sometimes Salesforce might be a good idea. However, it should be a part of an overall strategy, not just because everyone does it. This kind of deliberate tooling strategy is difficult though because the way Google Analytics or Salesforce works from what I understand is make marketing folks feel they are specialized in Google Analytics or Salesforce so they feel like they have to keep using it or their skill will become useless.
It is like resume driven development but for the whole business.
I think it‘s kind of a common knowledge now that Salesforce is very expensive, so it is not a go-to choice for most startups/no-CRM-experience people. You are more likely to start with Hubspot today than with anything else, but those low-effort CRMs are also quite easy to migrate from. Google Analytics too, so it’s not exactly a „lock-in“. The lock-in happens when you struggle with your current setup or risks associated with it become unacceptable, but do not have the budget and a competent team or external partner to execute the migration.
„Everyone does that“ is definitely part of decision-making process almost everywhere, but I personally have not seen companies where it’s just a cargo cult rather than a reasonable strategic choice. The obvious benefits are that it’s easier to find implementation partners, the costs are predictable and your users may already know the system, so you won’t have unnecessary friction in your ops.
1. Palantir isn't selling consulting as much as Palantir is selling the confidence you get from buying a name brand. It's the same as paying for McKinsey to provide justification to do what you already want to do.
2. Palantir actually has some good core tech. An in house team can probably do a better job just because the incentives are better aligned, but they'll be starting from behind and have to catch up.
3. LLMs aren't at a level to replace a team of FDEs. Maybe in a couple of years. The role requires too much understanding of the human systems, and too much initiative to keep the ball rolling/acknowledge and deal with real problems.
"In a 2023 blog post, external, Palantir described the challenge of combining data from multiple government systems containing tens of thousands of visa applications and hundreds of thousands of accommodation offers."
This is the kind of thing GDS and other Civil Service departments build all the time, its a completely standard kind of challenge that a small team of Devs (+ supporting staff) from a departments DDAT department does day in and day out.
The output will be open source by default and use existing standards.
Yeah, this exactly. "multiple government systems", "tens of thousands", "hundreds of thousands" is the typical "part-time allocation for four people in an office" government project. This should have a budget in the low hundreds of thousands of £ at most.
Hundreds of thousands of documents is small enough that you can feasibly run a pen and paper office handling them. Especially since most of them do not cross-reference eachother (family applications do, but unrelated families have no such links).
That America's brightest tech minds can't solve this problem is embarrassing. (Never mind the baggage of giving a foreign, potentially adversarial nation access to something as sensitive as residency and visa information.)
I assumed that when the GP said the UK was "giving a foreign, potentially adversarial nation access" the GP meant that the US is that "foreign, potentially adversarial nation"
I can't believe that in our timeline Europe has to think like this, but here we are.
My parents took in a Ukrainian family as part of this scheme, and I knew many others who did. They all matched with each other through Facebook groups set up for this purpose. I don't know anyone who was matched automatically by the Palantir thing
There's not really enough info to know if this is just a coin toss or something more. "Company tries to roll its own system and [saves / loses] money" is just a common story, one way or the other.
For context, the Homes for Ukraine refugee scheme cost 2-3 billion as of 2023. I can't seem to find an updated cost. This cost (from the article) was Palantir working for free for the first 6 months (could they have beat that, time wise?), then awarded 4.5m and 5.5m for two more 12 month terms, and now they're transitioning to something home-grown instead.
> The MHCLG [
Ministry of Housing, Communities and Local Government] said it initially needed a system which could be ready within days but, in seeking a "steadier service", later created an updated platform to meet the programme's longer-term needs and bring down costs.
I basically agree with the MHCLG's reasoning here. It's always worth at least experimenting to see if you can roll your own.
I worked on a small part of one of these back in around 2013 ( specifically managing beds ).
You were talking about a team of 5 cranking this out in about 2-3 months with some longer term part time involvement, with an annual cost of less than 1m and those people mostly all dellivering several product lines ( so actual cost is half or a quater ).
> "Company tries to roll its own system and [saves / loses] money" is just a common story, one way or the other.
Governments build these kinds of systems ("collect data from a bunch of internal systems and show some public forms and have some internal processes for handling form submissions") all the time. When I worked for a local municipality, we built something like this every other month.
> There's not really enough info to know if this is just a coin toss or something more.
The difference is always having one or two devs who care. Every successful software project I've ever seen has had a few devs who care way more than is healthy
Developing a replacement system is still going to cost a hell of a lot. It's not like if you dropped palatir then we'd suddenly have a free drop-in replacement and everyone can have their fiver back
You pay money to Palantir that money essentially escapes the economy, you develop a sovereign solution yes you pay millions even more but that goes into corporations and people actually living in the country, paying taxes and spending their coins here.
Honestly, a web app to match people looking for and willing to provide accommodation is completely different to dealing with the health data of an entire country that is pre-existing in different formats. The first is essentially a CRUD app that should never have been given to Palantir.
Palantir is not just analyzing data, but, it is increasingly wired into operational decisions like deportations, policing, health-data access, military targeting and public-sector workflows.
Tjheir "ELITE" guide says that during "special operations" normal safeguards may need to be turned off.
Palantir's Maven Smart System ha grown into a Pentagon program of record with 20,000+ active users. "Human in the loop" may become "human rubber stamp" when the number and speed of AI recommendations exceed real human review capacity.
A Palantir-backed program reportedly operated secretly from city council members, defense attorneys, and the public.
Vendor lock-in issue: once a system becomes embedded in agency workflows, switching vendors becomes politically and operationally hard and they are trying their best to achieve this. The Army's $10 billion enterprise agreement consolidating many contracts into one Palantir platform is the cleanest example of institutional dependence.
---
tldr;
The accountability chain is broken: when harm happens, the agency blames the tool, the vendor blames the customer, the operator blames policy, and the model blames the data.
---
Also, I won't share the full report link since whenever I share something like that here, I get banned/flagged for a day.
1. ICE awarded Palantir a reported $30 million contract for ImmigrationOS, described as a platform to support immigration lifecycle operations, including enforcement prioritization and self-deportation tracking.
2. Palantir’s Maven Smart System was designated a Pentagon ‘program of record’ in March 2026, with 20,000+ active military users and a contract ceiling that grew from $480 million to $1.3 billion.
3. The US Army’s $10 billion enterprise agreement consolidates 75 separate contracts into one Palantir platform.
4. The Maven Smart System has 20,000+ military users across 35+ military tools.
5. The UK NHS Federated Data Platform, valued at £330 million ($448.4 million), places Palantir at the center of England’s health-data architecture.
6. Palantir’s UK public contracts across NHS, Ministry of Defence, councils, and police forces total more than £500 million.
7. NHS England’s Data Protection Impact Assessment documents 15 inherent risks, all assessed as ‘Low’ residual risk after mitigations.
8. The NHS FDP contract was published with 417 of 586 pages redacted.
9. Palantir received more than $113 million in federal spending since Trump took office, plus a $795 million Pentagon contract.
10. Polling cited by The Guardian indicates more than two-thirds of the UK public are concerned about Palantir’s growing number of public contracts, and 40% distrust Palantir specifically regarding NHS patient data.
11. From detection to ‘prosecution’ (killing), ‘no more than two or three minutes elapse’ with Palantir systems, compared to six hours previously.
12. Palantir’s lobbying spending more than quadrupled since 2019, from $1.4 million to $5.8 million.
Palantir’s roots seem practically indistinguishable from the traditional Korean/Japanese dispatch programmers (often referred to as SI, or System Integration).
They dispatch engineers under the title of FDSEs, but in Korea and Japan, this kind of on-site deployment is often considered the lowest tier of programming.
In my own experience consulting with factory owners—advising them on hardware choices like Lumens versus Mitsubishi—I see the physical reality. The idea of absorbing a client's database into an ontology and hooking it up to an LLM sounds great in theory. However, considering the extreme fragmentation of equipment standards and data representations across different sites, I seriously question if this is a sustainable business model.
Sure, initially it’s just dispatch programming. But how can they possibly absorb all these disparate, chaotic field environments into a single platform asset? Even within a single factory, different assembly lines use entirely different equipment, often from completely different manufacturers.
The idea of interpreting every piece of equipment's specific protocol, reverse-engineering the DB schemas, standardizing the terminology, and modeling the entire approval flow seems practically impossible. Is this actually achievable? Take PLCs, for example: even if they share a standard communication protocol, the ladder logic itself is completely incompatible across different brands
Thinking about it in reverse, Palantir might have absolutely no intention of solving this fragmentation problem themselves. Their survival strategy might be to dictate the core tech stack of the end-point B2C clients, creating a structure that essentially incentivizes specific B2B vendors to fall in line. Ultimately, what makes Palantir so dangerous is the high likelihood that they will simply shift the massive cost of standardization onto those B2B subcontractors
Unfortunately, it's hard to escape the feeling that friends in high places, some lobbying and some er... reciprocal back scratching might have been instrumental.
See also senior staff at NHS England (or Digitial? can't remember) handing massive NHS compute contracts to AWS, and then leaving the civil service to become... an AWS employee.
Take a look at this job posting for example: https://www.jobs.nhs.uk/candidate/jobadvert/C9175-26-0093 .
The role is more aligned with IT/Data as obvious by the fact that the main skill requirement is SQL.
It is horrible to work for them and in fact in consulting as soon as you hear that the project is for the NHS people run and hide not to be assigned.
Same reason the US political system is falling apart - buyable businessmen eh I mean politicians in power. „Lobbying“
Because the problem they're trying to solve is this:
https://www.statista.com/statistics/1388245/uk-sick-leave-fi...
In the UK, long term sick leave with a mental health diagnosis is a way to be unemployed (ie. not working) but not have the disadvantages of that. There's the money difference: 530 pounds per month for sick leave, 338 or 425 pounds per month for unemployment. On long-term sick leave you get all advantages job seekers get (ie. "Universal Credit"), PLUS others (support for not being able to work, ESA, and support for extra living expenses due to long-term sickness, PIP). So if you don't want to work, long-term sick leave has many advantages, plus it's a big cost to employers. You don't have to look for work in long-term sick leave. In fact, nothing is expected at all (other than medical evidence).
2.8 million people are long-term ill, not participating in the workforce, not being economically productive at all, at least half due to being diagnosed with mental illness. The issue with this is that this is happening with full support of government employees, and even the courts cooperate to a lesser extent.
What the government is trying to do, in other words, is trying to kick people off, uh, let's say "mental disability", force them to work. And they need to do this without relying on government workers, because they often side with the people on sick leave.
Hence, Palantir. Being hated is a feature here, not a bug.
1. they aim to deliver product company margins with a consulting-heavy model.
2. your software purchase funds a cadre of "free" FDEs and deployment strategists who customize your install, build a bunch of data pipes/transforms, and talk to people to figure out what all the data means.
This could be a good deal if your tech org is not very competent at integrating your data, or if you have a sudden, short-term need. In the longer term, it's probably cheaper and more effective to develop a competent tech team, modernize the source data systems, and roll your own integration -- but that also requires leaders with long-term vision who are resistant to external hype and pressure.
https://www.thenational.scot/news/26055524.palantir-hired-30...
Build: you need expertise in contracts, knowing what you need and also software development.
It's obviously easier to buy than build, especially for civil service roles where they can't attract the best developers due to political/ideological constraints.
Spending 10x more on IBM or Palantir can't get them fired, but trying to build something in-house their organization don't have competence for can get them fired.
And this is even if you don't take lobbying or corruption into account.
Almost all governments have a legally defined public procurement framework. If this is overridden, it's pretty much always by elected politicians, not by regular government employees.
Also it's not like 4 years ago either UK or EU governments would expect they will soon want to get rid of all US companies in their public sector.
I'll give you an example. At a previous employer, We used Google Analytics. We paid for Google Analytics. I feel positive that as a mid size company, We shouldn't have paid for Google Analytics. The free product with 50 events in GA4 should be plenty for us. But why do we use Google Analytics in the first place? Because everyone uses Google Analytics.
I agree that sometimes Salesforce might be a good idea. However, it should be a part of an overall strategy, not just because everyone does it. This kind of deliberate tooling strategy is difficult though because the way Google Analytics or Salesforce works from what I understand is make marketing folks feel they are specialized in Google Analytics or Salesforce so they feel like they have to keep using it or their skill will become useless.
It is like resume driven development but for the whole business.
„Everyone does that“ is definitely part of decision-making process almost everywhere, but I personally have not seen companies where it’s just a cargo cult rather than a reasonable strategic choice. The obvious benefits are that it’s easier to find implementation partners, the costs are predictable and your users may already know the system, so you won’t have unnecessary friction in your ops.
1. Palantir isn't selling consulting as much as Palantir is selling the confidence you get from buying a name brand. It's the same as paying for McKinsey to provide justification to do what you already want to do.
2. Palantir actually has some good core tech. An in house team can probably do a better job just because the incentives are better aligned, but they'll be starting from behind and have to catch up.
3. LLMs aren't at a level to replace a team of FDEs. Maybe in a couple of years. The role requires too much understanding of the human systems, and too much initiative to keep the ball rolling/acknowledge and deal with real problems.
This is the kind of thing GDS and other Civil Service departments build all the time, its a completely standard kind of challenge that a small team of Devs (+ supporting staff) from a departments DDAT department does day in and day out.
The output will be open source by default and use existing standards.
That America's brightest tech minds can't solve this problem is embarrassing. (Never mind the baggage of giving a foreign, potentially adversarial nation access to something as sensitive as residency and visa information.)
I can't believe that in our timeline Europe has to think like this, but here we are.
Note that Palantir is an American company that failed to solve this problem well, and introduces an adversarial risk to the UK.
And ran out of rows
https://www.bbc.co.uk/news/technology-54423988
For context, the Homes for Ukraine refugee scheme cost 2-3 billion as of 2023. I can't seem to find an updated cost. This cost (from the article) was Palantir working for free for the first 6 months (could they have beat that, time wise?), then awarded 4.5m and 5.5m for two more 12 month terms, and now they're transitioning to something home-grown instead.
> The MHCLG [ Ministry of Housing, Communities and Local Government] said it initially needed a system which could be ready within days but, in seeking a "steadier service", later created an updated platform to meet the programme's longer-term needs and bring down costs.
I basically agree with the MHCLG's reasoning here. It's always worth at least experimenting to see if you can roll your own.
You were talking about a team of 5 cranking this out in about 2-3 months with some longer term part time involvement, with an annual cost of less than 1m and those people mostly all dellivering several product lines ( so actual cost is half or a quater ).
Governments build these kinds of systems ("collect data from a bunch of internal systems and show some public forms and have some internal processes for handling form submissions") all the time. When I worked for a local municipality, we built something like this every other month.
The difference is always having one or two devs who care. Every successful software project I've ever seen has had a few devs who care way more than is healthy
It didn't work very well, so GDS rebuilt it in-house.
Tjheir "ELITE" guide says that during "special operations" normal safeguards may need to be turned off.
Palantir's Maven Smart System ha grown into a Pentagon program of record with 20,000+ active users. "Human in the loop" may become "human rubber stamp" when the number and speed of AI recommendations exceed real human review capacity.
A Palantir-backed program reportedly operated secretly from city council members, defense attorneys, and the public.
Vendor lock-in issue: once a system becomes embedded in agency workflows, switching vendors becomes politically and operationally hard and they are trying their best to achieve this. The Army's $10 billion enterprise agreement consolidating many contracts into one Palantir platform is the cleanest example of institutional dependence.
--- tldr;
The accountability chain is broken: when harm happens, the agency blames the tool, the vendor blames the customer, the operator blames policy, and the model blames the data.
---
Also, I won't share the full report link since whenever I share something like that here, I get banned/flagged for a day.
1. ICE awarded Palantir a reported $30 million contract for ImmigrationOS, described as a platform to support immigration lifecycle operations, including enforcement prioritization and self-deportation tracking.
2. Palantir’s Maven Smart System was designated a Pentagon ‘program of record’ in March 2026, with 20,000+ active military users and a contract ceiling that grew from $480 million to $1.3 billion.
3. The US Army’s $10 billion enterprise agreement consolidates 75 separate contracts into one Palantir platform.
4. The Maven Smart System has 20,000+ military users across 35+ military tools.
5. The UK NHS Federated Data Platform, valued at £330 million ($448.4 million), places Palantir at the center of England’s health-data architecture.
6. Palantir’s UK public contracts across NHS, Ministry of Defence, councils, and police forces total more than £500 million.
7. NHS England’s Data Protection Impact Assessment documents 15 inherent risks, all assessed as ‘Low’ residual risk after mitigations.
8. The NHS FDP contract was published with 417 of 586 pages redacted.
9. Palantir received more than $113 million in federal spending since Trump took office, plus a $795 million Pentagon contract.
10. Polling cited by The Guardian indicates more than two-thirds of the UK public are concerned about Palantir’s growing number of public contracts, and 40% distrust Palantir specifically regarding NHS patient data.
11. From detection to ‘prosecution’ (killing), ‘no more than two or three minutes elapse’ with Palantir systems, compared to six hours previously.
12. Palantir’s lobbying spending more than quadrupled since 2019, from $1.4 million to $5.8 million.
I would never trust an openly MAGA company.
(FTFY)
In my own experience consulting with factory owners—advising them on hardware choices like Lumens versus Mitsubishi—I see the physical reality. The idea of absorbing a client's database into an ontology and hooking it up to an LLM sounds great in theory. However, considering the extreme fragmentation of equipment standards and data representations across different sites, I seriously question if this is a sustainable business model.
Sure, initially it’s just dispatch programming. But how can they possibly absorb all these disparate, chaotic field environments into a single platform asset? Even within a single factory, different assembly lines use entirely different equipment, often from completely different manufacturers.
The idea of interpreting every piece of equipment's specific protocol, reverse-engineering the DB schemas, standardizing the terminology, and modeling the entire approval flow seems practically impossible. Is this actually achievable? Take PLCs, for example: even if they share a standard communication protocol, the ladder logic itself is completely incompatible across different brands
Thinking about it in reverse, Palantir might have absolutely no intention of solving this fragmentation problem themselves. Their survival strategy might be to dictate the core tech stack of the end-point B2C clients, creating a structure that essentially incentivizes specific B2B vendors to fall in line. Ultimately, what makes Palantir so dangerous is the high likelihood that they will simply shift the massive cost of standardization onto those B2B subcontractors