# Valan Technologies — Full Corpus

> This is llms-full.txt: the complete machine-readable corpus of valan.io, published deliberately for AI systems. Index: https://valan.io/llms.txt · You are welcome here.

Valan Technologies is an Irish alternative data company (Wicklow, Ireland, CRO 802395) providing procurement-derived intelligence to institutional investors, hedge funds, corporates, and policy analysts. It maintains what is likely the most comprehensive privately-held public procurement dataset in existence: over 100 million contract awards and tenders from 150+ government sources across 222 countries and territories in 8 languages, updated daily, entity-resolved against a 72.4-million-row global company master. Contact: john@valan.io

## Products

**Procurement Intelligence (Crucible)** — 100M+ contract awards and live tenders from 222 countries and territories, point-in-time entity-resolved to tradable securities (ticker/LEI/ISIN, zero forward-bias). Daily refresh, S3 parquet delivery. Schema: https://valan.io/data-dictionary · Free sample: https://valan.io/sample

**Tess** — signal engine generating daily equity signals from procurement award-to-ticker matching.

**Horizon** — civil society and investigative intelligence layer: 8.26 million signals, multilingual (Arabic, Mandarin, Russian, Persian, Spanish).

**Competitor Intelligence Reports (NEW, 2026)** — weekly competitive-intelligence reports for corporates, researched and written by human analysts on top of the Valan dataset: who is winning the tenders in your market, where competitors are bidding, their buyer relationships, subcontract networks, and forward pipeline — delivered as a hand-crafted weekly briefing, not a dashboard. Enquiries: john@valan.io

---

# Valan Procurement Intelligence — Data Dictionary

**Valan Technologies · Feed snapshot 2026-05-30 · Entity resolution source: company_god (global resolved entity master)**

A point-in-time, entity-resolved global government-procurement dataset: awards and tenders linked to the awarded company's tradable security (ticker / LEI / ISIN), with a subcontract graph and a company dimension. Designed for systematic research — every listed-supplier award carries the point-in-time ticker (zero forward-bias) and an `investable_flag`.

HTML version: https://valan.io/data-dictionary · Free 1,000-row sample: https://valan.io/sample

## Global conventions

- **Country codes** — ISO 3166-1 alpha-2 (US, GB, DE).
- **Currency** — ISO 4217 in the `currency` column; `award_value` is in that currency.
- **Point-in-time integrity** — `ticker_as_of` is the ticker as of `award_date`, populated only from a genuine dated listing window; otherwise NULL (`pit_confirmed` flags genuineness). Never a current value dressed as as-of.
- **Entity join** — `supplier_entity_id` → `entity_dim.entity_id` (LEI-bridged).
- **Subcontracts** — `is_subcontract=TRUE`, `tier_level>0`; `linked_prime_id` → the prime award; full chain in `subcontract_graph`.

## Compliance / cleaning (applied to the feed)

- Email PII removed (verified 0). Business-contact phones retained as published procurement data.
- CUI markings redacted; radioactive/CUI rows excluded from the standard feed.
- China excluded (ccgp_* sources + buyer_country='CN'). Russia/Belarus excluded on/after 2022-02-24.
- `award_value`: -1 / 0 stub values normalized to NULL.
- IDIQ/framework ceiling values flagged `value_is_ceiling=TRUE` (top-0.1%-of-currency value repeated across ≥10 distinct suppliers) — exclude for real obligated-value analysis.

## fin_awards — financial-modelled awards (71,888,513 rows · 12,345,679 investable)

The tradeable/analytic award table.

| column | type | description |
|---|---|---|
| award_id | BIGINT | Unique award id (PK). |
| buyer_name | VARCHAR | Contracting public buyer. |
| buyer_country | VARCHAR | Buyer country (ISO-2). |
| supplier_name | VARCHAR | Awarded supplier (email-PII scrubbed). |
| supplier_country | VARCHAR | Supplier country (ISO-2). |
| award_date | DATE | Award date (the point-in-time anchor). |
| award_value | DOUBLE | Obligated value in currency (stub values → NULL). |
| value_type | VARCHAR | award (value>0) / modification (value<-1, de-obligation) / undisclosed (NULL). |
| value_is_ceiling | BOOLEAN | TRUE = IDIQ/framework ceiling, not obligated — exclude for real value analysis. |
| currency | VARCHAR | ISO-4217 currency of award_value. |
| cpv_primary | VARCHAR | Primary CPV procurement category. |
| vpv_code | VARCHAR | Valan harmonised procurement vocabulary code (cross-source category). |
| naics_codes | VARCHAR | NAICS industry code(s). |
| estimated_annual_value | DOUBLE | Modelled annual value (frameworks/IDIQ). |
| financial_quality_score | INTEGER | Data-quality score for financial fields (0–100). |
| is_framework | BOOLEAN | TRUE = framework / IDIQ vehicle. |
| linked_tender_id | BIGINT | FK → originating tender (fin_tenders.tender_id). |
| supplier_entity_id | VARCHAR | Resolved entity key → entity_dim.entity_id (LEI-bridged). |
| lei | VARCHAR | Supplier Legal Entity Identifier. |
| isin | VARCHAR | ISIN of the supplier's primary listing. |
| ticker_as_of | VARCHAR | Point-in-time ticker as of award_date (NULL if no real dated window). |
| ticker_current | VARCHAR | Current/latest ticker. |
| ticker_exchange | VARCHAR | Listing exchange. |
| ticker_mic | VARCHAR | Market Identifier Code (ISO-10383). |
| ticker_confidence | DOUBLE | Confidence of the deterministic name→ticker resolution (0–1). |
| pit_confirmed | BOOLEAN | TRUE = ticker_as_of came from a genuine dated window (0 forward-bias). |
| ultimate_parent_ticker | VARCHAR | Tradable ticker of the ultimate parent (for subsidiaries of listed parents). |
| ultimate_parent_lei | VARCHAR | LEI of the ultimate parent. |
| investable_flag | BOOLEAN | TRUE = a tradable ticker is present (own or ultimate-parent rollup). |
| is_subcontract | BOOLEAN | TRUE = subcontract award. |
| tier_level | SMALLINT | Subcontract tier (0 = prime). |
| linked_prime_id | BIGINT | FK → prime award (subcontracts); see subcontract_graph. |

## master_awards — descriptive awards (71,857,265 rows · 4,787,331 investable)

Broader, text-rich award table (titles/descriptions/URLs). Same identity/ticker block as fin_awards; award_value / award_date are raw VARCHAR.

| column | type | description |
|---|---|---|
| award_id | VARCHAR | Unique award id. |
| source | VARCHAR | Origin portal/source system. |
| detail_url | VARCHAR | Link to the source notice. |
| buyer_name / buyer_country / buyer_city | VARCHAR | Buyer org, country (ISO-2), city. |
| supplier_name / supplier_country | VARCHAR | Supplier (PII-scrubbed), country (ISO-2). |
| award_date | VARCHAR | Award date (raw text). |
| award_value | VARCHAR | Award value (raw; stub values → NULL). |
| value_type / value_is_ceiling / currency | MIXED | As fin_awards. |
| cpv_primary / vpv_code / naics_codes | VARCHAR | Category codes. |
| title / short_description / full_description | VARCHAR | Notice text. |
| data_quality_score | VARCHAR | Overall record quality score. |
| [identity + PIT ticker block] | MIXED | supplier_entity_id, lei, isin, ticker_as_of, ticker_current, ticker_exchange, ticker_mic, pit_confirmed, ultimate_parent_ticker, investable_flag, is_subcontract, tier_level — see fin_awards. |

## entity_dim — company dimension (312,156 rows) — join target

| column | type | description |
|---|---|---|
| entity_id | VARCHAR | Resolved entity key (PK) ← *.supplier_entity_id. |
| legal_name | VARCHAR | Legal name. |
| country | VARCHAR | Domicile (ISO-2). |
| lei | VARCHAR | Legal Entity Identifier. |
| isin | VARCHAR | Primary listing ISIN. |
| primary_ticker | VARCHAR | Primary ticker. |
| exchange | VARCHAR | Primary exchange. |
| mic | VARCHAR | Market Identifier Code. |
| ultimate_parent_entity_id | VARCHAR | Ultimate parent entity (self-join). |
| parent_ticker | VARCHAR | Ultimate parent's ticker. |
| is_public | BOOLEAN | Publicly listed. |
| is_sanctioned | BOOLEAN | Flagged on a recognised sanctions list. |

## subcontract_graph — sub→prime linkage (3,311,541 rows)

| column | type | description |
|---|---|---|
| sub_award_id | BIGINT | Subcontract award (fin_awards.award_id). |
| prime_award_id | BIGINT | The prime award it sits under. |
| tier_level | SMALLINT | Tier depth (1 = first-tier sub). |
| sub_supplier_entity_id | VARCHAR | Subcontractor entity. |
| prime_supplier_name | VARCHAR | Prime supplier name. |
| prime_supplier_entity_id | VARCHAR | Prime entity. |
| prime_ultimate_parent_ticker | VARCHAR | Tradable ticker the prime ultimately rolls up to. |

## fin_tenders (24,984,433) / master_tenders (24,274,179) — forward pipeline (open solicitations)

| column | type | description |
|---|---|---|
| tender_id | BIGINT | Unique tender id (PK). |
| source | VARCHAR | Origin portal. |
| buyer_name / buyer_country / buyer_region / buyer_city | VARCHAR | Buyer org / country (ISO-2) / region / city. |
| published_date | DATE/TZ | Publication date. |
| deadline | DATE/TZ | Submission deadline. |
| currency | VARCHAR | Tender currency. |
| contract_type / procurement_method | VARCHAR | Vehicle / method (master_tenders). |
| cpv_primary / vpv_code / naics_codes | VARCHAR | Category codes. |
| title / short_description | VARCHAR | Tender text. |
| tender_status | VARCHAR | Lifecycle status. |
| financial_quality_score / data_quality_score | INTEGER | Quality score. |

## company_god — global resolved entity master (72,351,281 rows) — internal reference, not part of the standard client feed

The full global entity resolution master (LEI/ISIN/MIC/ticker for all entities, is_sanctioned flagged). entity_dim is the award-touched, client-facing projection of it.

---

Generated 2026-06-04 from the 2026-05-30 feed manifest + parquet schemas. SHA256 checksums per table available in the manifest. Access: john@valan.io


---

# Public Sample Bundle

# Valan — Procurement Intelligence · Public Sample Bundle

**Public evaluation sample · Feed snapshot 2026-05-30 · Valan Technologies Limited**

A point-in-time, entity-resolved slice of Valan's global government-procurement
dataset: awards linked to the awarded company's tradable security, with the
point-in-time ticker and compliance rules applied. Schema reference:
[https://valan.io/data-dictionary](https://valan.io/data-dictionary) → the **`fin_awards`** table.

---

## Files

| File | Notes |
|---|---|
| `valan_sample_1k_fin_awards.parquet` / `.csv` | 1,000-row US-heavy investable slice. Identical content in both formats. |
| `valan_sample_loader.py` | Reference loader. Applies PIT and compliance rules; keeps PIT and current strictly separate. |
| [Data dictionary](https://valan.io/data-dictionary) | Full column reference. |

Run: `python valan_sample_loader.py <path>` · requires `pandas`, `pyarrow` (`duckdb` optional).

---

## Scope — read first

This is a **curated investable-only slice** (`investable_flag = True` on every
row). In the full universe only **~17%** of awards (12.3M of 71.9M) carry a
tradable ticker, so **do not extrapolate coverage from this file.** It is built
to show the schema, the identity resolution, and the point-in-time mechanics —
not the live hit-rate.

## Point-in-time / forward-bias — read before backtesting

Two distinct tradability concepts; do not conflate them:

- **`ticker_as_of` with `pit_confirmed = True`** — the ticker **as of the award
  date**, sourced from a genuine dated listing window. **Forward-bias-free.
  Use this for backtests.**
- **`ultimate_parent_ticker`** (parent rollup) — the supplier's **current**
  ownership link (who owns it *today*), **not** as-of the award. Useful for
  screening; look-ahead present. The loader surfaces this split directly rather
  than hiding it inside the rollup.

---

## Other usage rules

- **Currency** — `award_value` is in the **local** currency (`currency` column;
  this slice spans AUD, BRL, CZK, EUR, HUF, PLN, USD). **Never sum across
  currencies.**
- **Real obligated value** — use `obligated_awards()`: `value_type = 'award'`,
  positive value, and IDIQ/framework **ceilings excluded** (`value_is_ceiling`).
- **Schema** — this slice carries the full **`fin_awards`** identity + financial
  block: `financial_quality_score`, `is_framework`, `ticker_confidence`,
  `ticker_exchange`, `ticker_mic`, `ultimate_parent_lei`, `linked_tender_id`,
  `linked_prime_id`. `award_value` is numeric and `award_date` is a date.

## Compliance (verified at export)

PII-clean (email PII removed, verified zero; business-contact phones retained as
published procurement data). **No PRC-sourced data** (`ccgp_*` sources and
`buyer_country = 'CN'` excluded). **RU/BY excluded** on/after 2022-02-24.
CUI/radioactive rows excluded from the standard feed. No sanctioned buyer or
supplier present in this slice.

---

*Valan Technologies Limited · Public evaluation sample. The full feed —
71.9M awards, 25M tenders, daily refresh — is available by institutional
arrangement. Questions: john@valan.io*


---

# Research

## The Pantry Goes Dark

Published 2026-06-06 · https://valan.io/research/the-pantry-goes-dark · by John Colville, Valan Technologies

In May 2026, Valan Technologies showed that Qatar operated the most procurement-visible national food security programme in its global dataset. In February 2026 the Strait of Hormuz closed, and Qatar's strategic food procurement disappeared from public data. A systematic search for the global fertiliser supply chain — the signal that supply-chain theory predicts should follow such a closure — found that the agricultural fertiliser trade was never visible in procurement data at all. Strategic procurement transparency is a peacetime behaviour: the public record goes blank precisely where, and when, geopolitical risk is highest.

- Qatar published 355 public tenders in February 2026, against a normal baseline of 5 to 7 tenders per month — a roughly fiftyfold surge. None of the 355 tenders concerned food security.

- Qatar's monthly procurement award publications fell from 489 in February 2026 to 25 in March 2026 to 2 in April 2026.

- Saudi Arabia's monthly award publications fell from 295 in February 2026 to 91 in March 2026 to 2 in April 2026.

- Egypt, Jordan, and Iraq — collected by Valan through the same infrastructure — published procurement awards at normal monthly rates through April 2026 (35, 49, and 26 awards respectively in April).

- Of ten of the world's largest fertiliser producers, only one — Grupa Azoty of Poland — has a substantive agricultural-adjacent procurement footprint anywhere in 140 countries of data.

- Qatar's last visible strategic barley award was published in October 2025; its last visible wheat bran award was published in January 2026. No first-quarter 2026 barley tranche has appeared.


### Finding 1: Qatar's strategic food programme stopped publishing at the moment of crisis

Part One of this series documented Qatar's strategic food procurement machine in detail. Every quarter, Qatar's Ministry of Commerce and Industry tendered three to four million bags of barley through Monaqasat, Qatar's national procurement portal. A parallel wheat bran procurement series ran alongside it. Hassad Food Company — Qatar's sovereign food security vehicle, a fully owned subsidiary of the Qatar Investment Authority — operated within the programme as both supplier and intermediary. PricewaterhouseCoopers (PwC) was contracted at €1.1 million to author the programme's strategic doctrine. The entire programme was visible in public contract data, and Valan Technologies published the award table in Part One.

The final visible awards in that programme, as collected by Valan Technologies from the Monaqasat portal, were:

The Qatar barley programme followed an established quarterly cadence: tranches published in July 2024, March 2025, June 2025, September 2025, and October 2025. On that cadence, a first-quarter 2026 barley tranche was due. It did not appear. No tender, no award, and no activity from Hassad Food Company, Roza Hassad, or QFM Trading. The most regular sovereign food procurement programme in Valan's dataset stopped publishing in the same period the Strait of Hormuz closed.

Valan Technologies identifies two readings, and treats both as the finding. Either Qatar suspended its strategic food programme — which is implausible for an import-dependent peninsula facing a closed sea lane — or the programme moved off the public portal at the moment it became genuinely strategic. A state does not stop stockpiling barley during a blockade; it stops publishing the fact.


### Finding 2: Qatar surged public tendering in every domain except food

The silence is not explained by a portal shutdown. Valan Technologies' collection from Monaqasat captured 5 tenders in December 2025 and 7 in January 2026, consistent with the portal's normal public rhythm. In February 2026 — the month the Strait of Hormuz closed — Monaqasat published 355 tenders, a roughly fiftyfold burst.

Qatar surged procurement across every administrative domain in February 2026: medical consumables for Hamad Medical Corporation, modification works across sixteen schools, air traffic control radar simulators, primary substation supervision, sewage treatment works at Doha South, and pest control for the old port. None of the 355 tenders concerned the strategic food programme. There was no barley tender, no wheat bran tender, and no Strategic Food Security Facility works. In the largest single month of public tendering Qatar has ever shown Valan Technologies, strategic food procurement was entirely absent while procurement in every other domain surged.

> "A state that tenders everything in public — schools, radar simulators, pest control for the old port — published not one food-security tender in the month its sea lane closed. The programme did not stop. It stopped being visible."

The portal then fell quiet: 58 tenders in March 2026, 4 in April 2026, and 5 in May 2026. Award publications followed the same curve, falling from 489 award notices in February 2026 to 25 in March 2026 to 2 in April 2026.


### Finding 3: The publication collapse was specific to states on the Strait of Hormuz

The collapse in award publication was not a regional data artefact. Valan Technologies collects Egypt, Jordan, and Iraq through the same infrastructure as the Gulf states, and those three countries published procurement awards at normal monthly rates through April 2026.

The states that went quiet — Qatar, Saudi Arabia, Bahrain — are the states inside the closure's blast radius. The states that published normally — Egypt, Jordan, Iraq — are outside it. Valan Technologies notes a methodological qualification: its incremental collection from Gulf procurement portals is younger and thinner than its European feeds, so Valan certifies the direction of this collapse rather than its precise magnitude. The differential holds within a single collection system, which is what makes it legible.


### Finding 4: The global agricultural fertiliser trade is structurally invisible to procurement data

Part One predicted that a Strait of Hormuz closure would simultaneously disrupt fertiliser production, inbound food supply, and outbound LNG-to-ammonia revenue for the same Gulf states building visible food security infrastructure. The closure occurred. Valan Technologies therefore searched for the fertiliser signal — the emergency urea procurement, the stockpiling cascade that supply-chain theory predicts.

Valan Technologies ran a systematic search across 140 countries and three years of procurement data for ten of the world's largest fertiliser producers: OCP (Office Chérifien des Phosphates) of Morocco, Yara International of Norway, The Mosaic Company, Nutrien, CF Industries, ICL Group, K+S, EuroChem, Acron, and Grupa Azoty. Only one of the ten — Grupa Azoty — had a substantive procurement footprint.

The search results for the other nine were overwhelmingly false positives. "OCP" in the procurement record resolved to OCP Répartition, a French pharmaceutical wholesaler, alongside a Romanian CCTV installer and a Brazilian builder — not Office Chérifien des Phosphates, which controls roughly 70% of the world's phosphate rock reserves. "Mosaic" and "Acron" resolved to hundreds of unrelated companies sharing those names. "Yara" returned entities such as NAYARA, RUNYARA, and WAYARANG. The world's fertiliser industry, like China's grain conglomerates, leaves almost no tender footprint on earth.

The one genuinely visible major confirms the pattern. Grupa Azoty, the Polish state-adjacent chemicals group, held fourteen procurement contracts worth €222 million in Valan's data since 2023 — every contract located in Poland, every contract supplying ammonia water and urea solution to state industrial plants. Yara International's genuine contracts, once false positives were removed, were industrial rather than agricultural: urea for the sewage networks of Grenoble, nitrate salt solution for German flue-gas treatment, and AdBlue for municipal bus fleets.

Valan Technologies' conclusion is that the procurement-visible urea economy is industrial, not agricultural. Governments tender for urea to scrub nitrogen oxides, treat water, and run diesel bus fleets. They do not tender for urea by the shipload to grow wheat. The agricultural fertiliser trade — the one a Hormuz closure actually threatens — moves through commercial offtake contracts between producers, traders, and distributors, and generates no public procurement notice, for the same structural reason that China's grain conglomerate COFCO generates none: these are intermediaries in business-to-business wholesale markets, not counterparties of the state.


### Finding 5: There is no procurement-visible fertiliser panic anywhere with clean coverage

The demand side confirms the structural invisibility. European fertiliser-coded procurement awards rose from roughly 10 to 15 per month before the closure to 23 to 35 per month after it — a real uptick at small values, far short of panic. Across the Middle East and South Asia, fertiliser-related procurement was near zero before the closure and effectively absent after it. Egypt — one of the world's largest urea exporters — shows exactly one fertiliser award in 2026. Thirteen countries whose food systems depend on imported nitrogen show no visible procurement response. Valan Technologies concludes that either the response is occurring in commercial channels invisible to procurement data, or it is not occurring — and both possibilities should concern the institutions whose job is to know the difference.


### Methodological Discipline: Findings Killed Before Publication

Valan Technologies publishes the findings it discards, because the verification discipline is the product. The following claims were identified and excluded during the Part Two verification pass in June 2026:

Every claim that survives in this article survived this process. Most of what is published from procurement data does not go through it — which is worth remembering the next time a dataset tells you something dramatic about ammonium nitrate.


### The Investable Signal


#### Opacity Is the Signal

Qatar has not stopped buying barley; it has stopped publishing the price. When a state that routed €158 million of strategic grain through public tenders goes dark while continuing to consume, demand persists but public price discovery disappears — and with it the early-warning value the market was receiving for free. Regional grain basis, freight rates into Gulf ports, and the charter activity of the trading houses that served the visible programme become the replacement signals. They are harder to read, which is the function of the darkness.


#### Equity Is the Only Liquid Instrument

Because the agricultural fertiliser trade is procurement-dark, equity exposure is effectively the only liquid way to express the disruption. A Strait of Hormuz closure strands ammonia and urea capacity inside the Gulf and re-prices every tonne produced outside it — in North Africa, the US Gulf Coast, and Norway. The producers least visible in procurement data are precisely the ones whose commercial order books are repricing, invisibly, until earnings.


#### The Industrial Urea Chain Is a Free Public Proxy

The one urea market governments do publish — AdBlue, DeNOx, and water-treatment tenders across Europe — is a continuous public price series for refined urea sitting downstream of the same disrupted feedstock. European framework renewals in this category over the next two quarters are a free, public proxy for what the closure is doing to refined nitrogen prices. Almost nobody reads them. They are bus fleet contracts. They are also the only urea prices any government publishes.


### Conclusion: When the Fear Materialises, the Record Stops

Part One established a principle: procurement data is a record of what governments are afraid of running out of. Part Two adds the corollary the closure has demonstrated: when the fear materialises, the record stops. Strategic procurement transparency is a peacetime behaviour. Qatar published its pantry for years because publishing was costless and the tenders needed bidders. The moment the contingency arrived — the one PwC was paid €1.1 million to plan for — the publishing stopped, the portal surged with everything except food, and the most legible sovereign food programme in the world became as opaque as China's.

The negative space in procurement data is the global food risk map, and that map is no longer static. It goes blank exactly where, and exactly when, risk is highest. The countries that were never visible remain invisible. The countries that were visible have gone dark. And the supply chain connecting them — the nitrogen that determines whether next year's harvest exists — was never on the map at all.

The sovereign pantry is still being stocked; it can no longer be watched.

---

## The Sovereign Pantry

Published 2026-05-14 · https://valan.io/research/the-sovereign-pantry · by John Colville, Valan Technologies

A Valan Technologies analysis of public procurement contracts across 30+ national and international portals shows Qatar, Egypt, and the Gulf states systematically building strategic food reserves on a quarterly cadence — while China's parallel programme remains structurally invisible to procurement-based intelligence. The negative space in the data is the global food risk map.

- Qatar's Ministry of Commerce and Industry has run a quarterly barley procurement programme since at least 2024, tendering 3–4 million bags every 90 days, with combined contract values exceeding €158 million across six tenders in twelve months.

- Hassad Food Company — Qatar's sovereign food security vehicle — operates simultaneously as supplier and intermediary across over €115 million of state contracts visible in procurement data.

- Qatar has commissioned PricewaterhouseCoopers, at €1.1 million, to design the doctrine of its national emergency food strategy.

- The World Bank is funding 370,000 tonnes of new and extended grain storage capacity in Egypt — roughly one week of national wheat consumption, built on the footprint of the failed Toshka megaproject.

- A systematic search for COFCO, Nidera, Noble Agri, Sinograin, Chinatex, and Bright Food across 30+ procurement portals covering 140 countries returned 28 hits — all false positives. China's agricultural sovereignty programme is structurally invisible to procurement-based intelligence.

- Jordan, Bangladesh, Lebanon, Yemen, and Sudan show near-zero sovereign food-security procurement footprint. The countries most at risk are often the countries least visible.


### A Quarterly Barley Programme Hidden in Plain Sight

Sometime in the third quarter of each year, Qatar's Ministry of Commerce and Industry issues a public tender for several million bags of barley. Then another one. Sometimes in the same week.

This is not coincidence or administrative overlap. It is a programme.

The data is explicit. In the twelve months between July 2024 and October 2025, Qatar's Ministry of Commerce and Industry issued the following barley procurement contracts through Monaqasat, its national portal:

Running parallel to the barley series is an equally systematic wheat bran programme: 2.4 million bags in October 2023, 3.0 million in December 2023, 1.95 million in April 2024, 2.4 million in September 2024, 2.2 million in July 2025, 1.8 million in September 2025, a further 1.2 million in July 2025, and 3.0 million bags in January 2026.

This is not a government buying food for a population. Qatar has fewer than 3 million residents. This is a state building a strategic reserve, systematically, on a quarterly cadence, using public procurement as the operational mechanism — and the entire programme is visible in public contract data if you know where to look.


### The Architecture Behind the Tenders

The barley tenders are only the most visible layer. Qatar is building something more structured.

In April 2025, Hamad Port awarded a contract worth €2.86 million for the "Provision of X-Ray Portal Scanners in Strategic Food Security Facility at Hamad Port" — the SFSF, as documents name it. In July 2025, a second contract followed for the supply and installation of laboratory furniture and equipment, and associated gas system for the strategic food security facility: €1.66 million. A food security facility at the port, equipped with food science lab infrastructure, scanning technology, and cold chain logistics.

Qatar is not just stockpiling food. It is building the inspection, testing, and verification infrastructure to manage a strategic reserve with the rigour of a pharmaceutical supply chain.

"Developing a comprehensive plan to support the readiness of the State of Qatar's strategic stockpile in emergency situations (Strategic Reserves)." PwC is being paid to design the doctrine of Qatar's food emergency architecture — a government commissioning its own emergency food strategy from scratch, in 2025, with institutional seriousness.

And operating through all of it is Hassad Food Company.

Hassad appears in the procurement data in a role that is genuinely unusual: it is simultaneously a state-owned supplier and a market participant. It supplies barley to the Ministry of Commerce and Industry — it is the direct counterparty on tens of millions of euros of government contracts. But it is also the entity that sources that barley from global markets. Hassad is the intermediary layer between the world's agricultural export markets and Qatar's strategic reserve — a vertically integrated sovereign food vehicle that is visible in procurement data in a way that most sovereign wealth operations are not.

The entity appears under three different names in the contract data: Hassad Food Company, Hassad Qatar, and Roza Hassad. The combined value of their supply contracts to Qatar's government over the past 24 months, visible in the data, exceeds €115 million. This is likely a fraction of Hassad's actual activity — most of its operations run through direct commercial channels that generate no public procurement signal.


### Egypt's Silo Moment

Eight hundred kilometres west of Doha, a different scale of food security infrastructure is appearing in the procurement data.

The World Bank's "Emergency Food Security and Resilience Support Project" in Egypt has issued three procurement contracts for grain storage construction in 2024–2025:

The Toshka contract deserves attention. Toshka — the New Valley irrigation project in Upper Egypt — was President Mubarak's signature megaproject of the late 1990s, absorbing billions in investment to turn a Saharan depression into agricultural land. It largely failed. The infrastructure remains, underused. Now the World Bank is building 300,000 tonnes of grain storage capacity there, with a railway integration to move it north.

This is the pragmatic repurposing of a failed agricultural investment into strategic buffer stock infrastructure. Egypt imports roughly 60% of its wheat — among the highest import dependence of any large population. After the 2022 shock, when Russia's invasion disrupted the Black Sea wheat corridor, Egypt faced the vulnerability directly. The procurement data shows the response: the World Bank providing the capital, the infrastructure being built in the geography that already has the footprint, the strategic purpose being reframed.

> "The combined grain storage being built and extended through these three contracts represents capacity to hold approximately 370,000 tonnes of grain — roughly one week of Egypt's national wheat consumption. That is not food security. It is a buffer. But it did not exist at scale before 2024."


### The China Paradox

The most significant sovereign food-security operation in the world is not visible in public procurement data at all.

A systematic search of Valan's procurement database — covering 30+ national and international portals across 140 countries — for COFCO, the Chinese state-owned agricultural trading conglomerate that is the world's largest agri-food company by volume of grain traded, returned almost nothing outside China and Hong Kong. Two contracts for office furniture at US passport agencies in 2023. Total value: $43,000.

A follow-up search expanded the query to include COFCO's major acquired subsidiaries — Nidera (the Dutch-Argentinian grain trader acquired in 2014), Noble Agri (the Southern Hemisphere grain operation acquired in 2014), plus Sinograin, Chinatex, and Bright Food. The expanded search returned 28 hits across all 140 countries and three years of data. Every single hit was a false positive: a Portuguese granite quarrying company (Granidera), a Spanish flamenco and contemporary dance company (La Venidera), and the same two US office furniture contracts.

Every single result was a false positive. This is not a data gap. It is a structural finding. China's food sovereignty strategy does not operate through public procurement. It operates through equity — and leaves no tender footprint in any procurement portal on earth.

China's food sovereignty strategy does not operate through public procurement. It operates through equity. COFCO acquired Nidera in 2014. It acquired Noble Agri in 2014. It holds significant positions in port infrastructure, grain terminals, and agricultural logistics across Brazil, Argentina, Australia, and the United States — none of which generates a public procurement notice, because these entities are not buyers or suppliers in government tenders. They are intermediaries in commercial wholesale grain markets, port terminals, and B2B trade.

The methodological implication is the larger point. Procurement data reveals the demand side of state food security — what governments are buying, stockpiling, building infrastructure for. It is structurally blind to the supply side being built through equity. Qatar's Hassad programme is visible because Qatar routes its food security through government tenders. China's equivalent programme is invisible because it routes through acquisitions. Two sovereign food-security strategies operating in parallel, one procurement-visible and one not.

The corollary matters: states without China's equity-deployment capacity build food security the visible way, through public contracts. The entire Gulf procurement signal is, in part, a function of the fact that these states cannot do what China does.


### The Negative Space

Perhaps the most revealing signal in the data is what is absent.

The analysis searched for food-security-related procurement — fertilizers, staple grains, food reserves, strategic stockpiles — in thirteen import-dependent countries from 2025 onwards. The results:

The pattern is clear: the procurement visible in the data correlates strongly with state capacity, not food insecurity. Qatar — food insecure but resource-rich — has a sophisticated, multi-layered, continuously-tendering programme. Egypt — food insecure and resource-constrained — has a World Bank-funded infrastructure build. Jordan — food insecure and resource-constrained — has nothing visible. The countries most at risk are often the countries least visible.


### The Investable Signal

Three categories emerge from the procurement data for investors tracking the sovereign food-security buildout.


#### Agricultural Infrastructure Equities

The Egypt and Qatar data both point to demand for grain silo construction, cold chain logistics, port food-safety facilities, and laboratory infrastructure. The Toshka contract went to an Egyptian construction firm. The Hamad Port SFSF contracts went to regional engineering and equipment suppliers. As more states move from emergency-buying to infrastructure-building, this category should see sustained demand — and procurement data provides the leading indicator before projects appear in corporate earnings.


#### Soft Commodity Positioning

The Qatar barley and wheat bran programme is running on quarterly cycles. If you know that Qatar is tendering 3–4 million bags of barley every 90 days, you know something about forward demand in the barley market that is not priced into spot prices. The mechanism is not obscure — it is public — but it requires synthesis across procurement portals to see the cadence. The same analysis applied to Egypt's grain purchasing, or to UNGM's forward agricultural input tenders in fragile states, produces comparable signals.


#### Fertilizer Producers

The sovereign food-security buildout is a demand signal for agricultural productivity infrastructure — and agricultural productivity requires fertilizer. The states building strategic grain reserves are also, implicitly, the states with the greatest incentive to secure stable fertilizer supply chains. Qatar's position on both sides of this equation — as a major LNG-to-ammonia producer and as one of the most active sovereign food-security procurers — is not coincidental. It is a coherent national strategy that spans the input supply chain (gas → ammonia → urea → fertilizer exports) and the output demand (barley reserves, wheat bran stockpiles, food lab infrastructure).

The Strait of Hormuz sits at the centre of both sides of that equation. A disruption would hit fertilizer production, inbound food supply, and outbound LNG export revenue simultaneously — for the states currently building the most visible food security infrastructure in the world.


### What the Contracts Tell You

Procurement data is a record of what governments are afraid of running out of.

Qatar is afraid of a repeat of 2017, but larger. It is building the infrastructure, the doctrine, the inspection capacity, and the quarterly barley cadence to make sure it can survive an extended blockade of any form. Hassad Food is the operational vehicle. PricewaterhouseCoopers has been hired to make sure the strategy is coherent.

Egypt is afraid of another 2022. The World Bank's infrastructure money is flowing into grain storage at the scale where it starts to matter — not self-sufficient, but buffered.

Jordan, Lebanon, Bangladesh, and Yemen are visible in the data only as recipients of UN emergency programming, not as states building sovereign resilience. The negative space is the risk map.

And China — the actor with the greatest food sovereignty programme and the most global agricultural equity footprint — is effectively invisible to procurement-based intelligence because it operates entirely outside the public tender mechanism. The procurement data shows you who is building, where, and how fast. It does not show you who is buying the supply chain itself.

The sovereign pantry is being built, systematically, by the states that can afford to build it. The procurement data shows you where, how fast, and what they're stocking it with.

Everything else is the gap.

---

## The Signal Before the Storm

Published 2026-04-06 · https://valan.io/research/the-signal-before-the-storm · by John Colville, Valan Technologies

How global procurement data forecast the Iran crisis — and what it reveals about the supply chain shock wave that followed. Three datasets. Three timeframes. A single, coherent signal chain that ran eight weeks ahead of the closure of the Strait of Hormuz.


### Executive Summary

On 28 February 2026, the United States and Israel launched coordinated strikes against Iran. Within days, the Strait of Hormuz — through which approximately 20% of the world's oil supply transits — was effectively closed, triggering what the International Energy Agency described as the largest supply disruption in the history of the global oil market.

What the public did not know was that Valan's three-dataset intelligence platform had been registering extraordinary signals for weeks. Eight weeks before the first bomb fell, Valan's Sentinel dataset recorded 1,189 high-confidence defence procurement signals across 53 countries in a single week — eight times the typical weekly volume. In the weeks that followed, trade flow data showed unprecedented surges in warship imports, radar equipment and munitions across NATO member states. And when the conflict finally began, Valan's global tender database recorded the largest cross-sector procurement shock wave in the dataset's history — with one critical, counter-intuitive finding: the sectors most critical to the crisis almost entirely disappeared from public tender pipelines.

This brief sets out the three-dataset signal chain: what Sentinel showed in December 2025, what trade data confirmed in February 2026, and what the tender cascade reveals about how governments respond — and how they signal distress — in a global supply chain crisis.

> "Procurement data is the earliest leading indicator of defence posture shifts — before budget announcements, before policy speeches, before earnings calls. When governments panic-buy, it shows up in tenders first."


### 1. Sentinel: The Warning Eight Weeks Early

Valan's Sentinel dataset aggregates defence procurement intelligence from 16 primary sources across 139 countries, including Foreign Military Sales notifications (US DSCA), Congressional Research Service defence reports, national ministry procurement disclosures, and selected trade ministry outputs. Each record is enriched at ingestion with ticker resolution, forward commitment classification, and civil society cross-signal scoring.

In the five weeks running from late October 2025 through late January 2026, weekly Sentinel volumes ran at 66–147 records per week — consistent with a baseline period of elevated but not exceptional NATO procurement activity. Then, in the week of 29 December 2025, the dataset recorded a step-change:


#### The December 29 Spike

The spike was driven almost entirely by two US government sources: DSCA Foreign Military Sales notifications and CRS Defence reports. The geographic breadth — 53 countries in a single week — is without precedent in the Sentinel dataset history. This was not a routine procurement cycle. It was the United States pre-arming its alliance network at scale, and it is visible only in procurement data.


#### Country-Level Detail: Selected Recipients

The Philippines warrants particular attention. On 29 December 2025, the Philippines received $11.5Bn in US Foreign Military Sales notifications — 48 records, primarily naval and air defence systems. On 24 March 2026, less than four weeks after the Strait of Hormuz closed, President Marcos declared a national energy emergency — the first such declaration by any nation globally. The Philippines imports 98% of its oil from the Middle East. Sentinel had flagged the country eight weeks before the crisis.

† Finland's 97% forward commitment rate signals contractual pipeline lock-in, not just intent.


#### Ticker-Level Signals

Sentinel's enrichment pipeline resolves procurement records to publicly traded equities where contractual relationships can be established. The forward commitment values below represent contractually locked future revenue visible in procurement filings before any public earnings guidance or analyst estimate revision:

Note: Rheinmetall (RHM.DE) appearing in December 2025 Sentinel data is notable — it signals European sovereign procurement independent of US FMS channels, consistent with Germany's accelerated Zeitenwende rearmament programme.


### 2. Trade Data: Confirmation in February

Valan's Sentinel trade dataset (fin_sentinel_trade) monitors bilateral trade flows across 23 defence-relevant HS code chapters, tracking year-on-year anomalies against a rolling historical baseline. HIGH-signal anomalies require a statistically significant deviation from the bilateral mean, not merely an absolute volume threshold.

In February 2026 — the final full month before the strikes — the dataset recorded a cluster of unprecedented HIGH-signal anomalies concentrated in Norway, consistent with NATO pre-positioning activity:

The Norway–Israel broadcast/communications apparatus surge is flagged conflict_adjacent=TRUE in the dataset. The combination of warship registration (+248,057% YoY), munitions export acceleration, and radar procurement across multiple NATO bilateral pairs in a single month constitutes a trade-data confirmation of the Sentinel procurement signal. None of this was visible in public markets.


### 3. The Tender Cascade: After the Strait Closed

Valan's global tender database covers 350+ procurement sources across 140 countries in 8 languages, classified to the proprietary VPV taxonomy (27 sectors, 666 codes, 50,000+ crosswalk mappings to CPV/NAICS/ISIC/UNSPSC). The following analysis compares the 35 days before 28 February 2026 against the 35 days following it across all 27 VPV Level 1 sectors.


#### Cross-Sector Volume and Value Change


#### The Counter-Intuitive Finding

The most analytically significant observation in the tender data is not what surged — it is what collapsed. Energy, Fuel & Utilities tenders fell 64% post-war. Food, Beverages & Catering fell 40%. These are the two sectors most directly impacted by the Strait of Hormuz closure.

This is not a paradox. It is a signal in itself.

When governments face genuine supply emergencies for critical commodities, they do not publish open competitive tenders. They invoke emergency direct procurement powers, bypassing public procurement processes entirely. The disappearance of energy and food tenders from public pipelines is not evidence that governments stopped buying — it is evidence that the procurement became invisible, moving into emergency channels that standard market intelligence cannot observe.

What governments did tender for was everything downstream: infrastructure hardening, fleet replacement, backup power installation, medical stockpiling, logistics diversification. The Civil Engineering spike to €1.03 trillion — a 99% increase — represents governments simultaneously commissioning LNG terminal expansion, port hardening, strategic storage facilities, and grid resilience projects across 140 countries.

> "The signal is not in what they bought. It is in what went silent."


### 4. The Alpha Argument

Across three datasets and three distinct timeframes, the procurement signal chain is coherent and sequential:

No single commercial data provider has cross-sector visibility of this kind. Commodity desks see energy prices. Defence analysts see military tenders. Satellite imagery providers see factory utilisation. Nobody is showing a macro fund the full procurement shock wave moving through a sovereign government's entire spending apparatus simultaneously — across 140 countries, in real time, classified to a consistent taxonomy.

Accurately isolating defence procurement from broader security and civilian spending requires classification expertise that most procurement datasets lack. The same expertise that isolates a Rheinmetall contract from a municipal security guard tender also isolates a strategic LNG storage contract from a routine facilities management award. The classification layer is the alpha.


### 5. Dataset Coverage & Methodology

All tender and award records are classified to Valan's proprietary VPV taxonomy (27 sectors, 3 levels, 666 codes) with crosswalk mappings to CPV, NAICS, ISIC, and UNSPSC. Ticker enrichment uses a multi-stage resolution pipeline including ISIN-GLEIF matching (64,000+ international matches), international alias expansion (691 entries, 24 major conglomerates), and monthly FX normalisation to EUR. All values are EUR-normalised at time of publication.

---

## The $227 Billion Day

Published 2026-04-28 · https://valan.io/research/the-227-billion-day · by John Colville, Valan Technologies

On September 30 — the final day of the U.S. federal fiscal year — agencies obligate $227 billion in contract authority. The next business day that figure drops to $30 billion. What 68 million contract awards reveal about federal spending discipline.

I maintain what is likely the most comprehensive privately-held public procurement dataset in existence: 68 million contract awards from roughly 80 government sources worldwide, sitting in a 181 GB DuckDB file on a server in Frankfurt. It includes USASpending.gov, SAM.gov, the EU's TED system, Brazil's PNCP, Mexico's Compranet, and every U.S. state procurement portal worth scraping.

This weekend I ran the numbers on U.S. federal contracting patterns by calendar day. I found something remarkable.


### The Cliff

Here is the total dollar value of U.S. federal contract obligations by calendar day, aggregated across the last five fiscal years (FY2021–FY2025):

On September 30, federal agencies obligate $227 billion in contract authority. The next business day that figure drops to $30 billion. By the first weekend of October, daily obligations have fallen below $5 billion.

This is not a budget cycle. It is a cliff.


### The Mechanism: Use It or Lose It

The pattern has a well-known name in federal procurement circles: use it or lose it.

The mechanics are straightforward. Most federal appropriations expire at fiscal year-end. Unobligated funds don't roll forward into the next year's budget. Worse, a program that consistently underspends its appropriation sends a signal to OMB that it doesn't need as much money — and future budgets will be benchmarked accordingly.

The incentive structure is clear: obligate every available dollar before midnight on September 30, or watch it vanish.

The behavioral consequence is that the back half of September becomes a contracting sprint. Agencies race to commit multi-year funding before the curtain falls, regardless of whether the timing makes operational sense.


### What the Data Shows

Two things stand out when you separate volume from value.

First: the number of contracts obligated on September 30 is unremarkable. Across the last five fiscal years, agencies signed 47,277 contracts on the final day — roughly in line with a typical Tuesday in early October (57,520 contracts on October 7, for comparison).

Second: what changes is the size. The median September 30 contract clocks in at $1,199. The median October 7 contract is $496.

Agencies aren't panic-buying staplers. They're racing to close out the contracts that took six months to negotiate before their budget authority expires. Big-ticket items — multi-year IT modernization programs, fleet vehicle purchases, R&D commitments, facilities contracts — get pushed disproportionately to the final day.


### What Gets Obligated

Some of it is exactly what you'd expect: major weapons systems, naval construction, large-scale IT infrastructure. The Department of Defense alone obligates roughly $42.5 billion on a typical September 30.

But a sample of mid-range contracts ($50K–$5M, signed on the last day of the fiscal year across FY2022–FY2024) reads less like strategic procurement and more like a last-minute clearance sale:

It is entirely possible the federal government determined, on the very last day of FY2024, that $399,821 was the economically rational price to replace a carpet through a competitive procurement process.

It is also possible they were sitting on $399,821 of unobligated facilities funding that needed to be committed before sundown.


### The Leaderboard

The Pentagon and its surrounding industrial ecosystem dominate the final-day surge — which makes sense given Defense represents roughly 50% of discretionary federal spending. But the pattern holds across the civilian agencies as well.

One data quality note: Naval Supply Systems Command registers $140B from just seven contracts in the same window. These are almost certainly ceiling values on indefinite delivery vehicles for submarine or shipbuilding programs, not actual obligations. The figures above reflect what agencies reported to USASpending, not what was actually wired.


### Is This Universal?

You might assume any government with a fiscal year would exhibit the same pattern. The data says otherwise.

The United Kingdom's fiscal year runs April 1 to March 31. I pulled the same daily breakdown from the UK's Contracts Finder dataset — and the cliff doesn't materialize. March obligations run roughly $140 billion (aggregated across years). April runs $287 billion. The remaining months scatter between $97B and $285B with no dramatic end-of-year spike.

Whether this reflects genuinely different spending discipline, different reporting lags (UK contract awards often publish months after signature), or simply a smaller and less complete dataset, I can't yet say with confidence. But the pronounced American September cliff does not reproduce in the UK data.

This appears to be a specifically U.S. federal phenomenon — or at minimum, a phenomenon that is far more pronounced in the U.S. system than in comparable Westminster parliamentary procurement regimes.


### The GAO Has Been Warning About This for Years

The Government Accountability Office has flagged year-end spending behavior for at least a decade. A 2010 GAO study found that federal agencies obligated 16.3% of their total budget authority in the final week of the fiscal year — more than five times the weekly average.

More importantly, the study found that contracts signed in that final week showed statistically higher rates of subsequent modifications, de-obligations, and unobligated balances within twelve months. In other words: commitments made under time pressure at fiscal year-end are more likely to be unwound, restructured, or under-delivered.

> "Nobody has successfully fixed this. The structural incentive — obligate or lose your budget authority — has been ironclad for as long as the modern federal budget process has existed."

Multi-year appropriations help in narrow cases (Defense procurement, some R&D programs). So-called "no-year" funds exist for certain capital programs. But the core mechanism remains intact.


### What This Means

The United States runs roughly a $700 billion annual contract obligation budget. Approximately one-third of that total gets committed in the final sixteen days of September.

This is not a conspiracy. It is not corruption. It is a structural incentive embedded in the appropriations process, playing out at scale across the entire federal government.

It is the cleanest example I know of public policy bending the normal distribution entirely out of shape.

The FY2026 fiscal year ends on September 30, 2026. If the historical pattern holds, federal agencies will obligate more than $200 billion in a single day. I'll be watching the data.

---

## The Monolingual Blind Spot

Published 2026-04-08 · https://valan.io/research/the-monolingual-blind-spot · by John Colville, Valan Technologies

There are 246 Arabic-language investigative reports on Iraqi defence procurement in this dataset right now. The number in your ESG rating: zero. What ratings agencies miss — and why the gap is structural, not accidental.

There are 246 Arabic-language investigative reports on Iraqi defence procurement sitting in our dataset right now.

The number in your ESG rating: zero.

This is not a data quality problem. It is not a coverage lag. It is a structural blind spot — one that has been priced into the market for years, and one that is becoming harder to ignore as defence procurement moves east, south, and into languages that ratings agencies do not read.


### The Ratings Agencies Are Not Lying to You

They cover what is legible — English-language disclosures, Western regulatory filings, annual reports, Bloomberg data feeds. The methodology is published. The limitations are acknowledged, in footnotes, in the way that limitations usually are.

The problem is that the events that actually move defence sector valuations do not wait for English-language coverage.

A procurement corruption allegation in Baghdad breaks in Arabic on a Tuesday. The English wire picks it up, if it picks it up at all, three weeks later. The ESG rating reflects it, maybe, twelve months after that.

> "The investor managing risk against the ESG signal is not managing risk. They are reading the post-mortem."


### What 246 Reports Looks Like in Practice

Daraj Media is an independent investigative outlet based in Beirut. It covers the Arab world in Arabic. Its journalists break stories that later become formal legal proceedings — procurement corruption, Ministry of Defence budget diversions, contractor blacklistings, arms import irregularities.

Since October 2025, Horizon has ingested every Daraj report touching Iraqi defence procurement. 246 records. Classified by signal type, mapped to target companies and programmes, enriched with ticker references where a listed contractor is identifiable, and delivered as structured data the same day it publishes.

None of this content is in a ratings database. Not because Daraj is unreliable — it regularly precedes formal legal action by months. But because reading, parsing, and structuring Arabic-language civil society journalism at scale is not something ratings agencies have built for. We have.


#### Arabic & Non-English Sources Active in Horizon


### The Geography Problem Is Bigger Than the Language Problem

ESG ratings cover, with any analytical depth, roughly 20 to 30 countries.

The countries where defence procurement risk actually concentrates are not on that list.

Iraq. Yemen. Syria. Iran. Russia. Pakistan. The places where arms flow, where civilian harm is documented, where procurement contracts are challenged, where civil society organisations have been filing detailed sourced records for years — these are exactly the places where mainstream ESG infrastructure has the least coverage.

Horizon currently holds 8.26 million active signals across 277 countries. The curated human-source layer — investigative journalism, NGO monitoring, parliamentary records, civil society campaigns — runs to 17,500 records from 22 specialist organisations: AIRWARS tracking civilian harm in Iraq, Syria, Libya and Somalia; the Sanaa Center covering Yemen; PAX Netherlands on the arms trade; CAAT on UK export licences; Soldiers' Mothers Russia on conscription and casualties; the Uyghur Human Rights Project on dual-use technology exports from Xinjiang.

These organisations exist because the risk is real. They have been building the evidentiary record for years. It is now structured, deduplicated, and queryable.

The Bloomberg ESG equivalent: an aggregated country score, derived mostly from World Bank indicators, updated annually. These are not the same instrument.


### The Signal Chain

Here is how a defence procurement risk event typically moves through the world — and where Horizon sits relative to the instruments most investors are currently using:

Horizon sits at week one — not because it has a crystal ball, but because the sources that capture week-one events have existed for years and nobody structured them.


### The Honest Version of What This Is

Structured data. Not investment advice.

The 246 Daraj reports on Iraqi procurement do not tell you which stock to buy. They tell you that something is happening — organised, documented, credible civil society opposition to specific programmes in a specific country, right now.

What you do with that signal depends on your exposure and your own analysis. What we provide is the raw material: current, multilingual, geographically complete, and delivered before the English-language press arrives.

---

## What Russia Can't Buy

Published 2026-04-03 · https://valan.io/research/what-russia-cant-buy · by John Colville, Valan Technologies

Russia's defence trade has collapsed in categories it cannot domestically replace — and surged in categories that expose the limits of sanctions. Mirror trade reconstruction across 247 countries reveals a supply chain under sustained structural pressure, with rerouting visible in the flows that self-reported data deliberately obscures.

Russia stopped reporting its own trade to UN Comtrade in February 2022. The Kremlin's calculation was straightforward: remove the data, obscure the supply chain, frustrate the analysts. What the calculation missed is that every country Russia trades with still reports — and their records, inverted and reconstructed, tell the full story.

Valan Sentinel's mirror trade intelligence layer covers 247 countries and $787 billion in implied defence-relevant trade value. Every row derives from a transaction reported by a national customs authority. Russia's 15,300+ reconstructed records represent one of the most analytically significant datasets in the platform — precisely because Russia chose silence.


### The Collapse That Matters

Russian military aviation is the single most visible casualty of the sanctions regime in the trade data. Aircraft imports — HS chapter 88, covering military and dual-use aviation platforms, components, and systems — peaked at $6.95 billion in 2021. By 2023, the same flow had fallen to $76 million.

This is not a data artefact. The collapse is visible from multiple independent reporting countries — Germany, France, the Netherlands, Japan, Korea — each filing its own customs records and each showing the same cliff edge in early 2022. The mirror methodology, which aggregates partner-reported flows and inverts them to reconstruct the non-reporting country's position, is conservative: only direct import and export flows are used, excluding re-exports that carry higher uncertainty.

What does a 94% collapse in aircraft imports mean operationally? Russian military aviation maintenance, platform production, and avionics replacement all depend on supply chains that ran almost entirely through Western manufacturers. Controlled components — turbine blades, flight management systems, precision guidance electronics — cannot be domestically substituted on a two-year timeline. The procurement gap is structural, not temporary.

The 94% collapse in Russian aviation imports represents the most complete validated supply chain disruption visible in the Sentinel trade data. It is reconstructed not from estimates, but from the customs filings of Russia's actual trading partners — each an independent, official primary source.


### The Rerouting That Doesn't Hide

Kazakhstan is the most prominent node in Russia's visible rerouting architecture. Sentinel's mirror trade data identifies a $9 billion rerouting signal — imports flowing into Kazakhstan at volumes and in categories (military electronics, dual-use components, precision machinery) inconsistent with Kazakhstan's own defence industrial base and domestic consumption profile.

The rerouting signal is not a suspicion. It is a mathematical residual: the difference between what Kazakhstan reports importing and what Kazakhstan's GDP, defence budget, and industrial capacity can absorb. The gap is the signal. When Kazakhstan reports importing military electronics at volumes that exceed its entire declared defence budget, the destination is not Kazakhstan.

> "Russia stopped reporting its own trade. Every country it trades with did not. The mirror is exact."

The methodology here follows established practice. UN Comtrade uses mirror statistics to fill gaps where countries do not report. SIPRI uses partner-reported data to estimate arms transfers for non-reporting states. The CSIS China Power Project applies the same technique to Chinese military trade analysis. Sentinel automates and scales this methodology across 247 countries with daily updates.


#### Selected Russia Rerouting Indicators · Sentinel Mirror Trade


### The Paradox: What Russia Can Sell

The most counterintuitive signal in the Russian mirror trade data is not the collapse — it is the surge. Nuclear exports, classified under strategic materials in the Sentinel HS categorisation, tripled in the post-sanctions period, reaching $6.2 billion. This is not a violation of the sanctions regime: civilian nuclear supply contracts — enriched uranium for European and Asian reactors — remained outside the initial Western sanctions architecture, and Russia moved aggressively to monetise that carveout.

The signal carries two implications. First, it partially offsets the economic pressure of sanctions through hard currency revenue from a captive market — European utilities contractually dependent on Russian fuel supply could not unwind those contracts immediately. Second, it is a coercive leverage instrument: the export surge is also a reminder of dependency.

The nuclear export surge represents Russia's most effective sanctions arbitrage. Civilian nuclear contracts, initially outside Western sanctions architecture, provided a revenue offset as aviation and electronics imports collapsed. Niger's post-coup uranium export flows ($3.7B reconstructed from EU partner reporting) add a second dimension to the nuclear supply chain story.


### The Methodology: Why Mirror Data Works

Every country that does not self-report to UN Comtrade — or reports selectively — has its defence-relevant trade flows reconstructed in Sentinel using the mirror statistics methodology. When Germany reports importing aircraft components from Russia, that transaction appears in Sentinel as Russia exporting aircraft components to Germany. The original_reporter_iso3 field preserves full provenance for every row.

The reconstruction is conservative. Only direct imports (M) and direct exports (X) are mirrored. Re-exports and re-imports, which carry higher uncertainty about ultimate origin and destination, are excluded. All trade values are converted to USD at period exchange rates using the filing country's reported currency — no estimation is applied to values.

Known limitations: transit trade where goods pass through intermediaries may be attributed to the transit country; classification differences between reporter and partner create occasional commodity-level mismatches; timing differences affect monthly but not annual aggregates. The original_reporter_iso3 field enables analysts to weight rows by reporter quality and apply their own judgement on these factors.

Sources: US Census Bureau · UK HMRC · Eurostat Comext · Japan Ministry of Finance · KITA · ABS · Brazil MDIC · UN Comtrade. All open-access, commercially redistributable.


### The Investment Signal

For event-driven and systematic macro strategies, the mirror trade data offers something specific: a quantitative, source-verified, daily-updated signal on the operational sustainability of Russia's defence effort — disaggregated by category, by trading partner, and by HS commodity.

The aircraft import collapse signals structural degradation in air superiority capability on a timeline that cannot be reversed by policy decision alone. The electronics rerouting signal, visible in Kazakhstan and several other Central Asian intermediaries, indicates that workarounds exist but are constrained in volume and sophistication. The nuclear export surge signals that Russia is actively converting its remaining leverage into cash — a behaviour consistent with a state under sustained economic pressure optimising for near-term revenue.

None of this is visible in Russia's own reporting. All of it is visible in the records of Russia's trading partners. Sentinel reads both sides of every transaction — the one that is reported, and the one that is not.

Russia chose silence. The data did not.

---

## The Rearming World

Published 2026-04-03 · https://valan.io/research/the-rearming-world · by John Colville, Valan Technologies

A structural shift in global defence procurement is now visible in the data. Across 22 million contract awards, 1.8 million live tenders, and bilateral trade flows across 100+ country pairs, Valan Sentinel maps the largest rearmament cycle since the Cold War — in real time, from primary sources.

Something has changed. The decade after 2008 was defined by defence austerity — Western governments cut budgets, mothballed programmes, and drew down force structures on the assumption that the post-Cold War peace dividend would hold. It did not. The procurement data now records the reversal in granular, source-verified detail.

This is not a commentary on geopolitical events. It is a reading of the ledger: what governments have contracted, what they have tendered, and what their bilateral trade flows reveal about where hardware is actually moving. Taken together, the signals are unambiguous.


### The Eastern Flank Rearms

Poland is the most instructive case. At 4.15% of GDP — a figure drawn directly from SIPRI data enriched into Valan Sentinel — Poland now leads European NATO members in defence intensity. With 619 tracked intelligence signals and 19 confirmed forward commitments, the procurement posture is not aspirational. The contracts are placed.

Poland's procurement commitment now exceeds every major Western European economy. Sentinel tracks 619 active signals and 19 confirmed forward commitments — including Abrams MBT deliveries, HIMARS follow-on orders, and the domestic Harpia air defence programme.

Finland presents a different signal. With 39 forward commitments — the highest count of any European country in the dataset — the procurement curve reflects a nation that has institutionalised multi-decade defence planning. The Sentinel intelligence layer captures RTX's $950 million AMRAAM contract notification for the Finnish Air Force in 2026, cross-referenced against DSCA notification records and Finnish MoD budget statements. This is a high-strength, confirmed forward commitment — not a budget estimate.

Romania, appearing in Sentinel with 185,000 tracked contract awards — the largest international coverage outside the United States — illustrates the less-visible tier of NATO rearmament. Much of this reflects interoperability spend: C2 systems, logistics infrastructure, and F-16 transition programme awards visible across multiple source portals, most of which carry no ticker mapping and attract no analyst coverage.


#### Forward Commitment Signals · Selected European Markets


### The Indo-Pacific Surge

Japan is the second most signal-rich country in the Sentinel intelligence layer, with 2,818 tracked records — almost entirely sourced from ATLA, Japan's Acquisition, Technology and Logistics Agency. The dataset captures 487 unique contractors, a density that reflects Japan's defence-industrial base expansion under its revised National Security Strategy. RTX alone holds a $3.64 billion forward commitment for Japan in the 2026 DSCA notification cycle.

India's 2026–27 Union Budget allocated ₹172.5 billion for defence, a figure extracted from Ministry of Defence press releases and cross-referenced in Sentinel's intelligence layer. With 1,034 tracked signals and 18 forward commitments, India's procurement profile reflects both domestic Make in India programme awards and continued import dependency for complex platforms. South Korea's KIMDIS portal contributes 26,000 tenders — a source almost entirely absent from conventional financial data feeds.

Taiwan generates 601 tracked signals and 6 confirmed forward commitments despite operating no formal procurement transparency portal. Every record is reconstructed from third-party sources: US DSCA notifications, Congressional Research Service programme assessments, and partner trade statistics. The opacity of Taiwan's own reporting makes the reconstructed signal more, not less, informative.

> "Procurement data is the earliest verifiable signal of defence intent. Budgets are political. Tenders are contractual. Awards are irreversible."


### The Contractor Picture

BAE Systems is the single most tracked entity in the Sentinel contractor layer: 5,747 intelligence signals across 159 countries, spanning 13 primary source systems from UK SPIRE export licence data through US GAO programme assessments to Scandinavian and Gulf MoD publications. No other listed defence prime approaches this breadth of coverage in the dataset.

The breadth of BAE's footprint makes it the most comprehensively tracked defence prime in Sentinel. Average civil society signal count of 10.62 per record — sourced from Horizon's 570,000+ NGO monitoring records — reflects the ESG exposure that accompanies global scale.

Lockheed Martin carries 38 confirmed forward commitments — the highest absolute count of any ticker-mapped contractor in the dataset — spanning Australia, Canada, Germany, Greece, Israel, Italy, Norway, Poland, Saudi Arabia, Singapore, Taiwan, and the UK. The F-35 programme alone constitutes one of the most geographically distributed forward procurement signals in the database.

MBDA, the European missile consortium, appears 612 times across 99 countries despite carrying no exchange-listed ticker. Its civil society signal count of 11.32 per record exceeds both BAE and Thales — a risk indicator that follows advanced weapons system exports wherever they move.


### What the Tenders Tell You

Defence tenders precede awards by weeks to months. They are, structurally, the cleanest forward signal in the procurement data. Valan Sentinel tracks 1.8 million defence solicitations globally — 919,000 from US SAM.gov, 418,000 from Ukraine's Prozorro platform, 354,000 from EU Tenders Electronic Daily, and a further 131,000 from 25+ national portals including South Korea, Saudi Arabia, and Brazil.

Ukraine's Prozorro data deserves particular note. At 418,000 records — real-time military procurement during an active conflict, published daily with full transparency — it is the single most operationally significant international tender source in the dataset. The volume and cadence of Ukrainian solicitations provides a continuous quantitative signal of consumption rates, category shifts, and supply chain diversification that no other source can replicate.


### The Structural Conclusion

The rearming world is visible in the data before it appears in earnings calls or budget headlines. Tenders precede awards. Awards precede deliveries. Export licences precede headline arms deals. Force structure changes, captured in Sentinel's personnel layer across 181 countries, precede programme authorisations.

The aggregate picture Sentinel presents — European NATO members accelerating spend, Indo-Pacific nations building sovereign capability, US primes accumulating forward commitments across 30+ customer countries simultaneously — is not a projection. It is a reading of the official record, sourced daily from 50+ government portals, normalised across 8 languages, and cross-referenced against 65 million verified contract awards.

The rearmament cycle is already in the ledger.

---

© 2026 Valan Technologies Limited, Wicklow, Ireland. Research articles are published analysis; data described is available by institutional arrangement. Not investment advice.