Paris Produced 21—or More Than 22—Saudi-French Agreements. The Public Ledger Still Does Not Add Up to One Deal Total.

Saudi Arabia and France left their August summit with a political success and an accounting problem.

The Saudi and French public record does not agree on how many agreements were announced. Saudi reporting put the number at 21. The English joint statement published by the Élysée says “more than 22 agreements and memoranda of understanding”. It also says only that “a number” of instruments were signed during the visit. [S1] [S2]

Neither government has published a numbered schedule that reconciles either count. That makes the first task more basic than calculating a grand total: establish the population of instruments.

The available disclosures support a public ledger of at least 18 identifiable agreements, frameworks, transactions and administrative announcements. They do not support one defensible “deal value”. The package combines a €6 billion development ambition, export-credit capacity of up to $5 billion, a potential Aramco procurement-and-technology value above $3.7 billion, an up-to-$3 billion financing framework, awarded commercial contracts, a planned joint venture, an investment registration and several non-financial MoUs. Adding those figures would combine different units of economic meaning and risk double-counting.

The correct verdict is not that nothing happened. Several instruments are commercially material. It is that the Paris package is a portfolio, not a transaction—and its public disclosure is not yet sufficient to produce either a complete count or a committed-capital total.

Last verified: 1 September 2026.

The public ledger

The table below records instruments identifiable in official releases and high-quality contemporaneous reporting. “Status” describes the legal or delivery stage disclosed publicly; it is not a legal opinion on documents that have not been published. Values retain the issuer’s own qualifiers.

Items 1–4 and the general sectors appear in the joint statement. Items 6–16 were described by Arab News in its account of the signing programme; the Aramco and finance items also have issuer disclosures. [S1] [S3] [S4] [S5]

The ledger deliberately does not force the BNP licence into the agreement count. Its announcement occurred during the visit, but an investment registration issued by a regulator is a different instrument from a bilateral agreement or commercial contract. It may be one of the official tally’s “announcements”; the underlying schedule is unavailable.

Why there is no responsible headline aggregate

Four tests prevent the published figures from being added.

They measure different things. Qiddiya’s roughly €6 billion is an ambition over an entire development lifecycle. The Bpifrance $5 billion is a proposed financing ceiling. Aramco’s $3.7 billion is a potential combined value across two procurement agreements and one exploratory technology MoU. Alstom’s €500 million is reported as an awarded supply contract. SAUR’s $150 million relates to a contract for which a future joint venture intends to bid. One is project cost, another credit capacity, another procurement potential, another contract award and another opportunity pipeline.

They may overlap. The shopping line is intended to finance French-company contracts in Saudi infrastructure, urban development, transport and healthcare. Reporting identifies Riyadh Metro rolling stock and AlUla hospitality among contemplated projects. If a French contract is both announced separately and later financed through the shopping line, counting the contract and the facility as independent investment counts the same economic activity twice. [S3] [S4]

The currencies and periods differ. Euros and dollars cannot be combined without an exchange date. More importantly, conversion does not make a multi-year maximum comparable with a signed purchase order.

Commitment is not disclosed. “Up to”, “around”, “potential” and “ambition” are not decorative words. They define the exposure. The conservative committed-value subtotal cannot be computed because even the awarded-contract disclosures do not provide executed documents, payment schedules, cancellation terms or the value allocated among Aramco’s three instruments.

The most accurate aggregate as of the verification date is therefore: not publicly determinable.

The count discrepancy is a disclosure failure, not a semantic curiosity

The Élysée statement contains two formulations within the same document. Paragraph 11 says “a number” of agreements and MoUs were signed in defence, healthcare, AI, emerging technologies and entertainment. Paragraph 12 says the investment roundtable announced “more than 22” agreements and MoUs. [S1]

Saudi Gazette, summarising the Saudi-French package, says the two countries signed 21 agreements and MoUs. [S2] These could refer to different populations: state-level signings, roundtable announcements, commercial contracts or instruments signed before but presented during the visit. “More than 22” could also include additional private-sector arrangements not in the Saudi count.

Those are possible explanations, not documented reconciliations. Without a schedule that gives each instrument a name, parties, date and type, readers cannot determine which explanation is correct.

The missing schedule matters because count inflation is easy when one project produces several documents. A financing arrangement may have a government joint statement, a lender framework and individual buyer credits. An industrial package may contain two contracts and one MoU. Conversely, a single press release may bundle multiple agreements. Counting announcements, signatures and underlying projects as though they were interchangeable produces a politically useful number but a poor economic dataset.

France and Saudi Arabia can resolve this cheaply: publish the Strategic Partnership Council annex or an agreed signing list with document titles and parties.

Bindingness runs on a spectrum

The public language supports five broad categories.

At the firmest end are reported contract awards and purchase agreements: the €500 million Alstom train order, CMA CGM–Red Sea Gateway Terminal arrangement, and Aramco’s two procurement instruments. These are more advanced than MoUs, but “contract” does not disclose whether value is minimum, estimated, call-off or conditional.

Next are institutional financing frameworks. Bpifrance’s shopping line can become real buyer credit when eligible French contracts, a borrower, a lender, pricing and guarantees are agreed. Until individual transactions close, the ceiling is capacity, not cash.

Then come investment and joint-venture arrangements. Alstom’s assembly plan and SAUR–Nesma vehicle could build Saudi industrial capability. The disclosed evidence does not establish plant capex, ownership, award, local payroll or delivery date.

MoUs establish a field and a process. Mistral and HUMAIN have a strategically coherent proposition, but their published language says they will explore dedicated compute; it does not state a reserved volume, price or commissioning date. Qiddiya’s French instrument is explicitly exploratory, while the €6 billion remains a lifecycle ambition. [S6] [S7]

Finally, a licence is administrative permission. BNP Paribas can now establish an eligible regional-headquarters entity, subject to the Saudi regime. The licence does not prove that regional decision rights or staff have moved.

This spectrum is not a ranking of merit. MoUs can mature into large projects; contracts can be delayed or amended. It is a ranking of what the present evidence supports.

The capital map points in both directions

The package is strategically more interesting than a simple French-export story.

Saudi capital is considering a conspicuous move into metropolitan France through Qiddiya. If developed, it would export the Kingdom’s entertainment-development model and place Saudi capital directly in a politically visible European consumer asset. That is unlike domestic Vision 2030 procurement.

French exporters and financiers, meanwhile, are positioned for Saudi rail, port, water, aviation, industrial and healthcare demand. The Bpifrance facilities are designed to improve financing availability for contracts executed by French companies. The economic benefit to Saudi Arabia will depend on procurement terms, local content, technology transfer and whether financing accelerates viable projects rather than preserving marginal ones.

The AI relationship is reciprocal but asymmetric in a different way. HUMAIN can supply capital and Saudi compute; Mistral supplies models and European strategic positioning. The value will depend on whether the partnership produces reserved capacity, Saudi deployment, local technical work and intellectual property—not the signature itself.

AlUla sits outside a narrow capital-flow frame. Extending cooperation to 2035 reinforces a state-to-state platform covering archaeology, heritage, training, tourism and sustainability. [S3] Its significance is institutional continuity rather than a disclosed transaction value.

Some announcements are continuations, not new starts

The summit brought genuine new signatures, but several items have a prior history.

The AlUla agreement extends and amends an established bilateral relationship rather than creating it. The Saudi Electricity–Bpifrance arrangement follows a cooperation MoU signed in November 2025. Vallourec already had a long-term relationship supplying oil-country tubular goods to Aramco, including Saudi industrial capacity; the latest purchase agreement should be evaluated as part of that commercial continuum, not as France entering the market for the first time. [S8]

Le Monde reported that the Jeddah Terminal 4 investment had been announced previously in autumn 2025. [S9] A summit can legitimately advance, formalise or showcase an existing programme. The problem begins only when a re-announcement is counted as wholly new capital without its chronology.

A clean ledger therefore needs two dates: the original commercial commitment and the state-visit announcement. It should also show whether the new instrument changes value, scope or legal status.

The strongest case for the Paris portfolio

The package fits identifiable Vision 2030 constraints. Rail and port contracts address transport capacity. Water-treatment collaboration targets project infrastructure. Export credit can lower the financing friction facing Saudi buyers. AI and industrial technology links address capability, not merely goods. Assembly and localisation language points towards domestic production.

France also offers Saudi Arabia an alternative set of technology, finance, tourism and cultural partners at a time when supply-chain resilience has strategic value. The relationship spans oil services and metros on one side, AI and entertainment on the other. That diversity reduces dependence on any single sector.

The most credible items have named counterparties, a defined object and an instrument: Alstom trainsets; Aramco drilling equipment and OCTG; Saudia aircraft finance; a Bpifrance export-credit architecture. These can be monitored through purchase orders, financial close, deliveries and operating assets.

The summit also created an institutional container—the first French-Saudi Strategic Partnership Council—in which future obstacles can be escalated. [S1] That mechanism may prove more durable than the headline count.

The countercase: diplomatic breadth can conceal weak conversion

The portfolio’s breadth is also its weakness. The public list blends mature procurement with early exploration, making it easy for political communications to transfer the credibility of a signed contract to an unfunded concept.

The largest visible figures are among the least committal. Qiddiya’s €6 billion is an ambition; Bpifrance’s $5 billion is an “up to” line whose operational modalities remain to be finalised; Aramco’s $3.7 billion is potential combined value without an allocation; and the $3 billion Saudi Electricity framework is another ceiling. A narrative built from those four numbers would be dominated by maxima and optionality.

Delivery risk also differs by jurisdiction. The French destination requires land, planning, environmental, financing and intellectual-property resolution. Saudi projects require procurement, borrowing and localisation decisions. Cross-border cooperation does not remove those gates.

Nor is export credit free capital. It supports lenders against non-payment and helps French exporters compete; the underlying Saudi borrower still owes financed amounts under transaction terms. The economic test is whether the acquired asset produces adequate public or commercial value.

Finally, the absence of a consolidated annex weakens accountability. If governments cannot agree publicly on the denominator within a week of the summit, later claims about implementation percentages will be difficult to audit.

What Vision 2030 gains—and what remains unproven

Vision 2030 gains access to French equipment, finance, industrial capabilities and cultural-development experience. It also gains an outbound-investment test: whether a Saudi giga-project company can develop a viable destination in a demanding European planning and consumer market.

What remains unproven is conversion. No public document shows how much cash is legally committed, how much will be Saudi- or French-funded, how much procurement will be localised, or how many of the MoUs contain milestones and termination rights.

The policy value should be assessed through five outcomes: assets delivered, financing drawn, Saudi jobs and capabilities created, commercially sustainable activity in both countries, and transparent public exposure. Announcement count is not an outcome.

The 90-day evidence test

By late November 2026, the public record should answer six questions.

1. Population: will the two governments publish one reconciled list of the 21 or 22-plus instruments?
2. Qiddiya: is there a development agreement, site-control instrument, rights package, planning pathway or funded budget?
3. Export credit: have operational modalities been executed, and has any named buyer credit reached mandate or close?
4. Procurement: will Aramco disclose counterparties, allocated values, term and purchase-order mechanics?
5. Localisation: will Alstom, the Aramco suppliers and the water joint venture disclose Saudi manufacturing, employment and training commitments?
6. AI: will Mistral and HUMAIN convert exploration into priced, reserved and energised compute capacity?

A transaction should move in the ledger only when evidence changes its state: MoU to definitive agreement; facility to signed financing; award to financial close; plan to construction; construction to operation. That is the difference between a summit communiqué and a Vision 2030 delivery record.

Related Vision 2030 Context

• Qiddiya’s French theme-park proposal: the commitment test
• Aramco’s $3.7 billion French package, disaggregated
• The Bpifrance $5 billion shopping line explained
• BNP Paribas’s Saudi regional-headquarters licence
• AlUla’s Saudi–French partnership now runs to 2035

Sources

1. [S1] Élysée, “Joint statement by France and Saudi Arabia”, 25 August 2026. https://www.elysee.fr/en/emmanuel-macron/2026/08/25/joint-statement-by-france-and-saudi-arabia-on-the-occasion-of-the-visit-of-his-royal-highness-prince-mohammed-bin-salman-bin-abdulaziz-al-saud-crown-prince-and-prime-minister-of-the-kingdom-of-saudi-arabia-to-the-french-republic
2. [S2] Saudi Gazette, “Saudi Arabia, France deepen strategic partnership with 21 agreements and MoUs”, 25 August 2026. https://saudigazette.com.sa/article/664026/saudi-arabia/saudi-arabia-france-deepen-strategic-partnership-with-21-agreements-and-mous
3. [S3] Arab News, “Saudi Arabia, France sign agreements spanning defense, AI and investment”, 24 August 2026. https://www.arabnews.com/node/2655730/amp
4. [S4] Saudi Ministry of Finance and National Debt Management Center, joint statement with Bpifrance Assurance Export, 24 August 2026. https://mof.gov.sa/en/MediaCenter/news/Pages/News_24082026.aspx
5. [S5] Aramco, “Aramco enhances its global partnership ecosystem through collaboration with French companies”, 24 August 2026. https://www.aramco.com/en/news-media/news/2026/aramco-enhances-its-global-partnership-ecosystem-through-collaboration-with-french-companies
6. [S6] Qiddiya Investment Company via Saudi Press Agency, MoU with the French government to explore development in Cergy-Pontoise [Arabic], 24 August 2026. https://www.spa.gov.sa/ar/w2660962
7. [S7] Mistral AI, “Mistral x HUMAIN”, 24 August 2026. https://mistral.ai/news/mistral-x-humain/
8. [S8] Vallourec, “Vallourec signs a ten-year agreement with Saudi Aramco”, 15 November 2023. https://www.vallourec.com/app/uploads/2023/11/20231115-Press-release-Vallourec-Aramco.pdf
9. [S9] Le Monde, “‘Colossal’ theme park project near Paris unveiled as Saudi Arabia’s MBS visits France”, 25 August 2026. https://www.lemonde.fr/en/international/article/2026/08/25/colossal-theme-park-project-near-paris-unveiled-as-saudi-arabia-s-mbs-visits-france_6756817_4.html