Category: Accountability

Public procurement, like constitutional law, should be boring.

It should be about the mundane everyday activity of public bodies purchasing things so that they can fulfil their public functions.

But it is currently exciting, and my pieces at the Financial Times and here on the Ministry of Defence contracts with Palantir were disconcertingly popular.

But as a preliminary remark: nothing which follows is disparaging about the greater number of those engaged in public procurement and contract management for public authorities whose professionalism keep things from being a lot worse. The public are lucky to have you.

The points below are generally about the faults of a system – and about how it benefits cynical contractors acting in their commercial self-interest and about how non-procurement officials and their political masters lack realism.

The points below are not about those that somehow stop greater abuses from happening.

*

The first point is about money.

Loads of money.

And as with the fog at the start of Bleak House, money everywhere.

Billions – not just millions.

If you can get a lucrative public services contract, you and your pension plan and your family are not just made for life, but also for generations.

Public procurement is an absolute geyser of cash.

And – this is the important thing – central and local government are good reliable payers.

There are hardly any defaults, and there are hardly any early terminations.

Many contracts just rollover, just like the public authorities that let them.

Public authorities suing on a contract is almost unheard of, more is the pity.

Sometimes a public authority will assert its contractual rights – and this will grab attention like any rare event – but mostly public authorities will keep funnelling the money to contractors.

And because central and local government are good reliable payers then the canny contractor can use the revenue stream effectively as security for other aspects of the business.

The returns on savings or an investment fund are nothing as the percentages a contractor will make on a public contract, especially if they then sub-contract the actual provision of goods, services or works at a discount to sub-contractors – who, in turn, are sometimes made up of the very same people who provided the same things for the public authority in-house before being out-sourced.

Some may complain (and no doubt will complain below) that the above is a horrible caricature: but in my experience there is enough truth in the depiction set out above for it to be offered as a concern on this blog.

And it gets worse.

*

The second point is the general lack of transparency.

Public procurement for no good reason whatsoever is shrouded in secrecy.

Routinely the facts about public procurement – that is when huge amounts of public money are spent on things (supposedly) so public functions can be discharged – are not disclosed to the, well, public.

And this is an attitude often at the highest levels of public authorities.

The magic phrase “commercial confidentiality” is constantly invoked, often by those who want you to nod-along with their mock-earnestness.

But “commercial confidentiality” in public procurement is utter balderdash and complete flapdoodle – at least after the contract has been let.

And this secrecy cloaks so may inefficiencies and abuses – on both sides of the transaction.

But those involved know that any attempts to force public disclosure of commercial information about these contracts can be avoided, at least in any timely way.

In principle – because of the amounts of public money involved and the need for public functions to be discharged, as well as because of pretty basic things like transparency and accountability – there should not be as a general rule “commercial confidentially” in public procurement.

The contractors get the benefit of huge amounts of cash, paid on a regular and reliable basis. They really should be happy with that.

And so the price of such contracts for the contractor should include full public transparency – unless there is a reason other than commercial confidentiality for non-disclosure.

*

The third point is that the law and practice of public procurement will often favour a small group of large providers – often with deliberately forgettable corporate names – who can afford the risk and the expense of participating in elaborate procurement exercises without guaranteed return.

Some procurement exercises with their multitude of stages and questionnaires and voluminous tender documents cost a small fortune for a bidder.

Public procurement should be about non-discrimination and avoiding bias, but – counterintuitively – the complex rules to give effect to such laudable aims have the practical effect of excluding almost all providers.

Like how profit and sustainability rules in football have the practical effect of favouring already established clubs – disclosure, Aston Villa fan here – the rules of public procurement have the effect of favouring a small group of established providers.

*

The fourth point is about the closeness of some (but not all) contractors and public authorities – with the famous “revolving doors”.

Those who let contracts really should not then work for the contractors to whom those contracts were let.

Even if the integrity of such individuals is beyond reproach, unbiased public procurement – like justice – not only needs to be done, it needs to be seen to be done.

*

The fifth point is a personal bugbear about contracts which involve informational technology and intellectual property (IT/IP).

There is a tendency by non-procurement officials in public authorities – and even their political masters – towards bespoke IT/IP development.

And this suits the contractor just fine – especially if the contractor retains the property rights and gets to charge for testing and ongoing management.

And so sometimes you end up with public authorities beholden to that contractor for the IT/IP development long after the term of the initial contract.

The public body is captured.

Of course, public authorities should use, where possible, commercially-off-the-shelf (COTS) products or open source software.

And if there is a need for proprietary bespoke software then there has to be robust exit management plans and licensing arrangements so that a public authority does not become dependent on one provider.

But it seems some non-procurement officials and their political masters like gleaming new things, with wish-lists of white-boarded specifications.

Aspire-ware, vapour-ware.

Public authorities should stick to COTS products or open source software where possible, and if there really – really, really – has to be an exception, then considerable thought needs to go into not only the terms of the contract, but also into the practical contract management and re-letting of the contract, so as to avoid capture by a supplier.

*

And now to the sixth point, the saddest point of all.

The only thing worse than having (often inflexible and elaborate) public procurement rules is having no public procurement rules at all.

The recent experiences of Covid and the abuses of PPE show what happens when the rules are suspended and an anything goes approach is adopted.

The sheer amounts of cash at stake mean that corruption is pretty-much inevitable.

Ideally one would have disinterested public authorities picking-and-choosing the right supplier without the fuss of public procurement rules; but instead of picking-and-choosing you will get pick-pocketing, and at a vast scale.

As some Victorian statesman once said, not all problems have solutions.

*

And in conclusion: we are fortunate that because of the professionalism of many procurement officers and contract managers within public authorities that things are not a lot worse.

But non-procurement officials and their political masters really need to get a grip on what is going on.

For contractors, acting cynically in their own commercial self-interest, know exactly what is going on.

**

*

Over at the Financial Times at the weekend I had an “op-ed” column on how Palantir commercially captured the United Kingdom’s Ministry of Defence. That column was popular and widely shared.

The post below now unpacks the detail of that column – to show the “working-out” behind what was summarised at the Financial Times .

*

As a preliminary point, I do not have any strong opinions or detailed knowledge about Palantir as a company or about individuals connected with that company.

This post, like the Financial Times column, comes at the subject from a different direction.

In essence: what would a cold, close reading of the relevant public domain documents tell us?

Would those documents, in and of themselves, give rise to concerns about the contracts awarded by the Ministry of Defence to Palantir?

In adopting this approach, I am drawing on over twenty years of experience advising on and watching public procurement – and I happen to be a former central government lawyer dealing with public procurement matters (especially IT procurements).

What follows is my opinion based on the information available.

The concerns I express are about how the Ministry of Defence dealt with these procurements. In respect of Palantir, their land-and-expand approach is nothing other than what any rational if cynical economic agent would do when faced with the lucrative opportunities presented to it by Ministry of Defence. They appear to have done what they could get away with and perfectly lawfully.

In essence: Palantir commercially colonised the Ministry of Defence because the Ministry of Defence let them.

*
Let us start with the words of that eminent public procurement professional J. Arthur Prufrock:

“[…] And time yet for a hundred indecisions,
And for a hundred visions and revisions […]”

*

A public procurement exercise is a sequence of decisions (and indecisions), of (in a way) visions, and (often) of revisions.

Accordingly we have to re-create and trace what decisions (and indecisions) were made, and when, and by who.

*

The primary objective of any public procurement exercise is the letting (for that be the verb) of a contract – for goods and/or services, and even for grand works projects.

A public procurement exercise also should have regard to what happens after the contract is let – contract management and such like – and to the likely need for the contract to be re-let after a period of time.

What distinguishes public (sector) procurement from private (sector) procurement is that there are certain principles and procedures that a public body should follow when acting as a contracting authority.

In broad terms, the principles are those of transparency and equal treatment, and of the need for competitive tendering when possible. There should also not be any bias towards incumbents.

These broad principles are in turn enshrined (ho ho) in detailed procedures that are set out in legal codes. The current legal regime is primarily provided by the Procurement Act 2023 which replaced other laws, such as the Public Contracts Regulations 2015.

(The law and practice of public procurement rightly has many critics – and there is much criticism of the way it favours large bidders who can afford to go through the expensive and sometimes elaborate procurement processes, and also of its general inflexibility – but as the Covid experience showed us, the only thing worse than having rules of public procurement is not having rules of public procurement.)

*

Now let us look at these two Ministry of Defence contracts.

The Ministry of Defence let a contract to Palantir in 2022 and then again in 2025.

The formal notice for the 2022 contract award is here and the formal “transparency” notice for the 2025 contract award is here.

We will look at what the 2022 contract award notice (and published contract) tells us, and then we shall consider what should have happened next, and then we shall look at 2025 award transparency notice tells us. Regard will also be had to this this parliamentary debate of 10 February 2026 and to various press releases.

*

The 2022 Ministry of Defence contract award to Palantir

The notice tells us the following basic information about the award: it was for a three-year contract for “software package and information systems” to start in 2022, and the value of that contract was £75,215,711.11.

The date of the notice tells us that it was before the current Procurement Act was in force – and it it seems that the Public Contact Regulations 2015 were followed (though there were also special defence procurement regulations).

The notice also tells us this:

This passage tells us the following:

• the contract was not advertised (“…without prior publication”);
• it was a direct award;
• such direct awards usually are because of urgency or that only one supplier is capable of delivering the requirement.

Now we need to look at the (then) applicable law (assuming that is the 2015 regulations).

The starting point is that there should have been a competitive tender unless an exception applied.

The relevant threshold at that time appears to have been £138,760 (well below the £75,215,711.11 value of this contract.)

So what exception?

Under regulation 32, “[i]n the specific cases and circumstances laid down in this regulation, contracting authorities may award public contracts by a negotiated procedure without prior publication”.

Those cases are circumstances are:

• where no tenders, no suitable tenders, no requests to participate or no suitable requests to participate have been submitted in response to an open procedure or a restricted procedure – and we can assume this was not the case;
• where the works, supplies or services can be supplied only by a particular economic operator for any of the following reasons:—

(i) the aim of the procurement is the creation or acquisition of a unique work of art or artistic performance,

(ii) competition is absent for technical reasons,

(iii) the protection of exclusive rights, including intellectual property rights,

but only, in the case of paragraphs (ii) and (iii), where no reasonable alternative or substitute exists and the absence of competition is not the result of an artificial narrowing down of the parameters of the procurement; and

• insofar as is strictly necessary where, for reasons of extreme urgency brought about by events unforeseeable by the contracting authority, the time limits for the open or rest – and we can assume this also was not the case.

So we can assume it is the middle bullet-point condition.

We can also assume that the software package was not “a unique work of art or artistic performance”.

So it must be [ii] or [iii] of the middle bullet-point condition.

Given the procurement was the proprietary software then we can assume it was [iii] – though legally it would be the same if it had been [ii].

Therefore, if it was [ii] or [iii] the law then provides it there can only be such a direct award “where no reasonable alternative or substitute exists and the absence of competition is not the result of an artificial narrowing down of the parameters of the procurement”.

*

Now turning back to the 2022 procurement exercise, and following Prufrock, what decisions and indecisions had to have been taken by the Ministry of Defence at this point for there to have been a contract ward without publication?

First, there would (or should) have been an output-based specification prepared within the Ministry of Defence. Normally this would be set out in neutral terms in a discrete specification department. Setting out what is required in terms of outputs would avoid supplier or product bias. There would (or should) have also been a business case.

Second, there would (or should) have been a decision by the Ministry of Defence that only one supplier was capable of meeting that (neutral) requirement – that “no reasonable alternative or substitute exists”. This decision (or should) have involved market sourcing and analysis. It would (or should) not have been assumed that there was no reasonable alternative or substitute exists without, well, reasoning – and that reasoning based on data. The Ministry of Defence would have to be persuaded that there was no alternative.

Third, the Ministry of Defence would have been mindful that “the absence of competition is not the result of an artificial narrowing down of the parameters of the procurement”.

What should not happen is that from the beginning the objective was always just to contract with one supplier for their proprietary product, regardless of market research. This is not least that then there would not be any informed decision that “no reasonable alternative or substitute exists”.

*

As there was no advertisement published we will never know whether there was a reasonable alternative or substitute available for meeting any output-based specification.

*

At the time there was a press release from Palatir. The content of this press release would either have been agreed between the Ministry of Defence and Palantir or at least would have been known to the Minsitry of Defence.

The press release quoted an official from the Ministry of Defence:

“Palantir’s Enterprise Agreement with the MOD will accelerate the UK Armed Forces’ journey to become a truly integrated force. We’re proud to expand our long-standing relationship with the MOD through our partnership with Defence Digital, and look forward to delivering world-leading software capabilities to enable greater operational outcomes.”

That press release described the software package as follows (emphasis added):

“Worth £75 million over three years, the partnership will support the MOD’s digital transformation as it becomes a world-leading agile force of the future. Spearheaded by Defence Digital and powered by Palantir, the digital transformation will see the MOD treat data as a strategic asset, harnessing its power to deliver superior military advantage and greater efficiency across the enterprise, from headquarters to the front lines.

“This partnership aims to accelerate Defence Digital’s ambitious vision of where the UK needs to be by 2030, providing secure access to its data across all operational domains, Top Level Budgets and UK Armed Forces bases.

“Working in close collaboration with Defence Digital, Palantir software will enable the MOD to exploit data at scale and speed to make faster, better decisions across Defence. Building on more than a decade of partnership, the agreement will enable any part of UK Defence to gain access to Palantir software across multiple classifications, wherever and however it can help – turning the MOD’s digital vision into reality at pace.

“Palantir builds leading digital platforms for data-driven operations and decision-making. By helping develop a data-backed representation of Defence assets, personnel and workstreams, Palantir software brings all the data that matters into a single view through a modeling concept known as an ‘ontology.’

“As disparate data sources are integrated into Palantir software, it digests the information, cleaning and harmonising the data and mapping it to an ontology. The ontology allows users to see data not as rows and columns but as objects, properties, and the relationships between them.

“With this sophisticated, intuitive data model in place, MOD personnel can perform advance scenario planning, testing hypotheses and modelling how they would play out in real-time. Whether for front-line operations, budget decisions or maintenance planning with industry partners, Palantir’s software gives the MOD the ability to understand outcomes before making decisions.”

What this press release tells is that even in 2022 the procurement was expressly seen by both the Ministry of Defence and Palantir as the beginning of a long-term project – “journey to become a truly integrated force” – that would take until at least 2030 – “this partnership aims to accelerate Defence Digital’s ambitious vision of where the UK needs to be by 2030” – in respect of something which would become absolutely critical for the Ministry of Defence.

*

Yet this “journey” to “where the UK needs to be by 2030” was to be on a three-year contract – starting in 2022 and ending in 2025.

So that placed the Ministry of Defence in a difficult situation.

Normally a contracting authority will make provision in the agreement for a contractor to be able to “exit” after the term of the agreement so that an alternative provider can be appointed. In this way a contracting authority does not get captured by a supplier to which the authority becomes beholden.

This is especially the case where there is bespoke development and/or proprietary software (or other intellectual property) on which a public authority can become dependent.

Otherwise the contracting authority becomes trapped into a contractual relationship, unable to appoint another provider.

And a trapped contracting authority then is at the mercy at whatever the contractor wants to charge for the next contract.

A contractor will have landed and expanded.

*

The curious incident of the Exit Management Plan

We have a version of the the contract signed in 2022. This is published by the government with its notice for the 2022 contract award.

And you will see in the table of contents that Schedule 17 has headed “Exit Management Plan”.

An Exit Management Plan ensures, among other things, that a contract can be effectively re-competed at the end of the term.

(The schedules, including Schedule 17, have not been published – though this is not unusual with published government contracts.)

There is no substantive clause in the main contract incorporating schedule 17 as a whole (Alan Hansen wince at poor drafting – it is probably in Schedule 1, which is also not published) but Schedule 17 is referred to clause 32A, which provides:

The Authority’s rights of use to Project Specific Intellectual Property and Contractor owned COTS software shall be licensed in accordance with Schedules 10, 15, 17 and 22 respectively.

(“COTS software” means Commercially Off The Shelf software – like Windows etc.)

“Project Specific Intellectual Property” is not defined in the main contract – again it is probably defined in the unpublished Schedule 1 – but we can assume it means the intellectual property of Palantir which the Ministry of Defence is paying for in respect of services under the contract.

If so, Palantir was under an obligation to license the software to the Ministry of Defence in the event of an exit from providing services to the Ministry of Defence.

As this was a three-year contract for a critical service, it would have been essential for the Ministry of Defence and Palantir to agree a practical, realistic and robust exit management plan under Schedule 17, including the rights of the Ministry of Defence to use the Palantir software in accordance with clause 32A during that exit.

We do not know whether there was an exit management plan under Schedule 17, as the schedules to this contract have not been published.

If there was an exit management plan under Schedule 17, we do not know how much serious and meaningful effort was put into formulating the plan for exit after three years, in view of the express “journey” to “where the UK needs to be by 2030”.

But.

What we do know, by inference, is that whatever exit management plan there was (if any) was not practical, realistic and robust enough for there to be a competition for the follow-on contract from 2025.

This was a severe public procurement failure, even if there was a case for not advertising the first contract.

*

The 2025 Ministry of Defence contract award to Palantir

The reason we know, by inference, that whatever exit management plan there was (if any) was not practical, realistic and robust enough for there to be a competition for the follow-on contract from 2025 is because of the 2025 contract award transparency notice (published in January 2026).

(The 2025 award was under the new Procurement Act 2023.)

This transparency notice stated:

“United Kingdom Ministry of Defence (MoD), intends to award a follow on enterprise agreement to Palantir Technologies UK Limited (Palantir) for continued licencing and support to data analytics capabilities supporting critical strategic, tactical and live operational decision making across classifications across defence and interoperable with NATO and other allied nations Palantir systems.”

The transparency notice then said that there had been a direct award.

There had been no competition (again) and no advertisement (again).

What was the justification for the direct award?

Direct award justification

Single supplier – technical reasons

It is considered that the contract can be awarded directly in accordance Section 41 of the Procurement Act 2023 together with paragraph 6 (absence of competition for technical reasons) and paragraph 7 (existing services where a change of supplier would result in disproportionate technical difficulties) of Schedule 5.

This is because MoD’s data analytics capabilities use Palantir data analytics architecture that only Palantir is able licence, and which only Palantir has the design familiarity and technical expertise to fully support.

Changing supplier for this requirement would involve

rebuild of the underlying data analytics architecture needing support;

reaccreditation of the new solutions at the required security levels; and

retraining of MoD personnel;

at significant cost (including to the current of level capability and interoperability with NATO and allied partners), diversion of resource, and disruption to in-train military operations and planning.

(Links added, and quotation broke into smaller (sub-)paragraphs and emphasis added.)

And what was the value of this new three-year contract, compared with the £75,215,711.11 value of the initial contract?

The value of the new contract was £240,600,000.

This is over three times as much – and note the transparency notice says that the “underlying data analytics architecture” is already in place.

*

For completeness, the new Act’s direct award justifications relevant here are:

The following conditions are met in relation to the public contract—

(a) due to an absence of competition for technical reasons, only a particular supplier can supply the goods, services or works required, and

(b) there are no reasonable alternatives to those goods, services or works.

And:

The public contract concerns the supply of goods, services or works by the existing supplier which are intended as an extension to, or partial replacement of, existing goods, services or works in circumstances where—

(a) a change in supplier would result in the contracting authority receiving goods, services or works that are different from, or incompatible with, the existing goods, services or works, and

(b) the difference or incompatibility would result in disproportionate technical difficulties in operation or maintenance.

*

Somebody at the Ministry of Defence will have been persuaded and then decided in 2025 that:

• due to an absence of competition for technical reasons, only a particular supplier can supply the services required, and there are no reasonable alternatives to those services; and
• the contract concerns the supply of services by the existing supplier which are intended as an extension to, or partial replacement of, existing services in circumstances where a change in supplier would result in the contracting authority receiving services that are different from, or incompatible with, the existing services, and the difference or incompatibility would result in disproportionate technical difficulties in operation or maintenance.

These issues, of course, were foreseeable in 2022 and this is why there needed to be a practical, realistic and robust exit management plan.

Had there been a practical, realistic and robust exit management plan then an authority would not need to use these direct award justifications.

The Ministry of Defence had become dependent on the software service it purchased in 2022, notwithstanding the supposed three-year contract and an exit management plan.

*

There is no explanation in the contract notices for why the value of the contract had shot up so much.

In parliament last month a minister said new commitments had been added:

“As part of the development of the new enterprise agreement, the MOD negotiated a strategic partnership with Palantir last September. The SPA reaffirms the strong relationship developed between UK defence and Palantir over the past decade, and includes new commitments that this Government secured from Palantir, including £1.5 billion investment into the UK, a new UK defence tech SME mentoring scheme to help companies grow and access the US market, and a commitment that London is to be the company’s European defence headquarters.”

None of these additional capabilities also seem to have been advertised.

An earlier government press release set out these claims, word for word:

“The UK will be at the leading edge of defence innovation as the government signs a new partnership with Palantir to unlock billions in investment and deliver on the Government’s Plan for Change.

[…]

New strategic partnership with Palantir to unlock up to £1.5bn investment into the UK to deliver new jobs, growth and national security.

Palantir announces plans to make the UK its European HQ for defence and create up to 350 new high-skilled jobs, cementing the UK as a state-of-the-art defence technology hub.

Palantir and UK military to develop AI-powered capabilities already tested in Ukraine to speed up decision making, military planning and targeting.

“It will see Palantir invest up to £1.5bn to help make the UK a defence innovation leader and create up to 350 new jobs, making defence an engine for growth.

“The new partnership, signed today (18 September) by Defence Secretary John Healey, will help the UK military develop the latest digital tools and harness AI technology to accelerate decision making, improve targeting and keep the British people safe from evolving threats. Palantir has also announced plans for London to become the base for Palantir’s European defence business, establishing Britain as a hub for defence technology innovation across Europe.

“The arrangement will also support the growth of British Defence Tech companies across the supply chain, with Palantir helping to mentor and develop UK companies. This will include helping British defence start-ups and SMEs to expand into US markets, including an offering on a pro bono basis.”

*

Here presumably the “£1.5 billion investment” would not have been a relevant consideration for awarding the contract, as contracts should not be rewarded for the promise of monies flowing elsewhere.

*

On the face of it, there is now nothing stopping Palantir having the contract awarded again and again, every three year cycle, perhaps tripling the amount each time.

If the government was not able to have an exit management plan that would have allowed an advertised competition in 2025 then it is unlikely to have one in 2028 and 2031 and so on, as we go on this “journey” to “where the UK needs to be by 2030”.

Palantir have landed-and-expanded.

The Ministry of Defence is now Palantir’s new commercial colony.

*

As I concluded in my Financial Times column:

”This is a public procurement failure. Even if the 2022 contract without any advertisement was justified, the MoD should have been preparing for a competition for the next contract. Instead, Palantir’s tools were embedded in the department and the company was awarded a second contract worth over three times as much as the first.

“Indeed, had Palantir suggested 10 times as much, the MoD would have been put in a difficult position. It also looks as if this is not the only Whitehall department facing a so-called land-and-expand tactic by the company.

“We cannot know from the public documents whether any other supplier could have provided services against the same specification because no specification has been published. In three years’ time, and in every three-year cycle, it is possible that the same will happen again. Unless evidence to the contrary is provided to the public, it appears as if the government department responsible for defence has commercially surrendered to a single service provider.”

**