Metropolitan Police publishes £300 M tech pipeline after Palantir deal blocked

![Featured image]

The procurement showdown

On 20 May 2026 Deputy Mayor for Policing and Crime Kaya Comer‑Schwartz refused to sign off on a £25.3 M contract with Palantir UK for the Metropolitan Police Service (MPS). The decision was grounded in two procedural breaches: the Met only engaged a single supplier and failed to present its procurement strategy to the Deputy Mayor as required under the Mayor’s Office for Policing and Crime (MOPAC) guidelines.

MOPAC’s spokesperson stressed that, given the tight fiscal environment, any multi‑million‑pound contract must survive “full and proper scrutiny” and demonstrate clear value for money. The Met’s rebuttal highlighted the need for rapid innovation to stay ahead of hostile states and organised crime, and pointed to existing Palantir deployments across the MoD, NHS and other police forces.

What the £300 M pipeline looks like

Within days of the block, the Met published 30 tender notices covering £304 M (VAT inclusive) of technology spend slated for the next decade. The notices give us a rare, data‑driven glimpse into the force’s future architecture.

Category	Start	Duration	Estimated spend (incl. VAT)	Primary goal
Managed radio service	Aug 2028	6 y 9 m	£67.2 M	Replace legacy air‑wave network with LTE‑based push‑to‑talk and encrypted broadband for officers on the ground
Digital forensics (Labnet) re‑procurement	Aug 2027	5 y 2 m	£56.4 M	Scale forensic imaging, hash‑verification and chain‑of‑custody tooling across 30+ precincts
Connectivity framework	Q4 2026	4 y 2 m	£36 M	Consolidate broadband, 5G backhaul and edge‑computing nodes for CCTV, body‑cams and command‑center dashboards
Video analytics (AI)	FY 2027/28	1 y	£0.96 M	Deploy object‑detection models to flag weapons and unattended bags in real‑time feeds
Video search (AI)	FY 2027/28	1 y	£0.48 M	Enable natural‑language queries across the city’s 10 k+ camera archive
Situational‑awareness apps	FY 2027/28	1 y	£0.54 M	Mobile apps for officers to overlay live incident data, crowd density and threat levels

Beyond the tech‑focused contracts, the Met listed £1.6 bn of non‑tech spend, including a £672 M building‑works framework and a £204 M Physical Forensics Project (2029‑2038). A quirky outlier is a £3.13 M two‑year contract for feed, forage and hay for the force’s mounted unit.

Benchmarks and power‑draw considerations

While the procurement notices are high‑level, they hint at the underlying hardware and energy profile the Met will have to manage:

• Managed radio service – The move to LTE‑Advanced will likely involve Sector‑wide 5G base stations with a typical power consumption of ~1 kW per sector. Over a six‑year horizon, that translates to roughly 52 MWh per site, a non‑trivial addition to the force’s carbon footprint.
• Digital forensics – Labnet’s next‑gen platform is expected to run NVMe‑based storage arrays (8 TB per node) delivering >5 GB/s ingest rates. Power draw per node sits around 250 W, meaning a 30‑node deployment will consume ~2.2 MWh annually.
• AI video analytics – Real‑time inference on edge GPUs (e.g., NVIDIA Jetson AGX) averages 15 W per camera. Scaling to 10 k cameras would add ~150 kW, or ~1.3 GWh per year – a cost that must be factored into total‑ownership calculations.

These figures underscore why MOPAC is demanding rigorous value‑for‑money analysis. Energy costs alone could erode any upfront savings if the architecture isn’t optimised for low‑power operation.

Vendor lock‑in risk

MOPAC’s broader concern is the danger of vendor lock‑in. A single‑supplier contract, as seen with Palantir, can create a situation where migration costs skyrocket once the contract expires. The new pipeline mitigates this risk by:

1. Fragmenting spend across multiple categories – Radio, forensics, connectivity and AI are sourced via separate frameworks, encouraging competition.
2. Re‑using open standards – The connectivity framework is expected to be built on OpenRAN specifications, allowing the Met to swap radio vendors without wholesale replacement.
3. Mandating data portability – Tender documents for the digital forensics service explicitly require export of raw image data in standardised forensic formats (e.g., AFF, E01), reducing dependence on proprietary APIs.

If the Met proceeds with a mixed‑vendor strategy, it will be easier to benchmark individual components against industry baselines (e.g., 5G latency <10 ms**, **forensic imaging throughput >5 GB/s, AI inference latency <100 ms per frame) and replace under‑performing suppliers without a massive overhaul.

Build recommendations for a future‑proof Met

For homelab‑style enthusiasts looking to emulate a slice of the Met’s architecture, the following configuration balances performance, power efficiency and flexibility:

• Radio core – Deploy a U‑plane OpenRAN base station (e.g., Nokia AirScale or Mavenir) with Intel Xeon D‑1541 CPUs (10 cores, 2.1 GHz) and dual 10 GbE uplinks. Expect ~1.2 kW per site, scalable via power‑over‑Ethernet for remote radio heads.
• Forensics cluster – Build a 3‑node Ceph storage cluster using AMD EPYC 7543 (32 cores, 2.8 GHz) and 8 TB NVMe per node. This yields ~1.5 GB/s sustained write speeds and ~300 W per node under load.
• Edge AI nodes – Use NVIDIA Jetson AGX Orin modules attached to each camera feed. Pair with GStreamer pipelines for low‑latency video decoding and TensorRT‑optimised models for object detection.
• Central command platform – Host a Kubernetes cluster on Intel Xeon Scalable servers, running Prometheus/Grafana for telemetry, OpenTelemetry for tracing, and Istio for service mesh security.

Such a stack mirrors the Met’s announced spend categories while keeping the door open for alternative suppliers and future upgrades.

What’s next?

The Met’s pipeline is a public‑first move that forces the force to justify each line item against measurable KPIs. Watch for the following developments:

• Tender outcomes – The first award (managed radio service) is expected in Q4 2026. Early winners will set the performance baseline for the rest of the pipeline.
• Energy reporting – MOPAC has hinted at requiring annual carbon‑intensity disclosures for all high‑value contracts, which could push vendors toward greener silicon.
• Potential re‑engagement with Palantir – If the Met can demonstrate a competitive process, a revised Palantir deal (perhaps for a data‑fabric layer rather than an end‑to‑end platform) may re‑appear.

For now, the Metropolitan Police is navigating a delicate balance between rapid technological adoption and fiscal prudence. The published £300 M tech shopping list is both a roadmap and a litmus test for how public‑sector agencies can modernise without falling into the trap of single‑supplier dependence.