LeptisCode
    Back to Writing

    Data Residency Is the New Sales Pitch: Snowflake's $300M Saudi Bet and the Lesson for North African Governments

    Mohanid Elragel

    Saudi Arabia made in-country hosting a condition of doing business, and Snowflake built local regions. North African governments can use the same lever.

    Data Residency Is the New Sales Pitch: Snowflake's $300M Saudi Bet and the Lesson for North African Governments

    Global cloud companies now sell to governments in this region on one promise above all: your data stays in your country. Saudi Arabia shows how much leverage that gives a state. North Africa can use the same lever.

    What Snowflake announced

    On 27 October, TradeArabia reported that Snowflake formally opened its Riyadh headquarters for the Middle East, Turkey and Africa. Alongside it, the company described a "projected $300 million investment" in Saudi Arabia through 2030, working with the Ministry of Communications and Information Technology.

    The package has three parts:

    • Two more cloud regions in the Kingdom, on top of its existing Google Cloud region in Dammam, so customers can pick a provider and still meet data residency rules.
    • A skills programme with the ministry to train up to 25,000 people in data and AI over five years.
    • A larger local team and partner network run from Riyadh.

    Read the $300 million figure carefully. According to Middle East AI News, it comes from an IDC analysis of Snowflake's projected economic impact and customer value in Saudi Arabia. It is not a sum Snowflake has committed to spend. The plan itself was first announced at LEAP in September; October brought the formal opening.

    The interesting part is not the number. It is why a US data company opened local regions at all: Saudi buyers made data residency a condition of doing business.

    The lesson for North Africa

    Libya has already set itself a large target. Its National AI Strategy 2026 to 2030, launched on 1 June, aims to automate 50% of government transactions by 2030. It also targets digital identity for 70% of the population and digitising 70% of paper records.

    Automating half of all government transactions means citizen records, civil registry, tax, health and visas all flow through software. Where that data lives becomes a national question, not just a technical one. The strategy seems aware of this. According to iAfrica, it plans a Libya Sovereign Cloud for 2026 to 2027, a unified digital ID by 2027 and a national data exchange platform by 2028. The ethics charter adopted alongside it also names digital sovereignty.

    The Saudi example offers two lessons:

    • Residency is a bargaining chip. Once a government makes in-country hosting a rule, global vendors build local capacity to compete. Libya, Tunisia and Egypt can write the same rule into tenders.
    • The rule has to come before the systems. If ministries build first and set residency rules later, they will be migrating live citizen data out of foreign clouds for years.

    The strategy itself lists the obstacles: no unified AI body, fragmented legislation, limited quality data and a skills gap. Good system design cannot fix all of that, but it can stop new systems from making it worse.

    How to design government systems that keep data in-country

    Data residency is not a server location you tick on a form. Data leaks out through backups, logs, email providers and AI calls long after the main database is safely local. These are the rules we would set for any public-sector system:

    1. Classify data first. Decide what must stay in the country (identity, civil registry, health, tax) and what can live anywhere (public pages, open data). Build the rules around the first group.
    2. Build so you can move. Use containers, open databases such as PostgreSQL and open standards. A system tied to a cloud service that only exists abroad cannot move to a sovereign cloud when one launches.
    3. Keep the keys at home. Encrypt data and hold the keys in-country, under state control. A backup stored abroad is then unreadable without them.
    4. Hunt the hidden exits. Analytics scripts, SMS and email gateways, error trackers, CDNs and chat widgets often send citizen data to foreign servers. Audit every third-party call before launch.
    5. Use AI without exporting data. Run open-weight models on local servers, or send only redacted fields to outside services. Never pass full citizen records to a foreign API.
    6. Treat logs and backups as data. They hold the same personal details as the main system and need the same residency rules.
    7. Put it in the contract. Every tender should require in-country hosting, audit rights, and an exit plan that returns all data in an open format when the contract ends.

    None of this needs a $300 million budget. It needs the rules written down before the first line of code.

    How LeptisCode can help

    LeptisCode is a software consultancy based in Tripoli and Dublin. We design and build bilingual Arabic and English systems with data residency in mind from day one, and we can review existing systems for the hidden exits listed above.

    If your ministry or agency is planning work under the national AI strategy, get in touch.

    Sources

    • Snowflake announces $300m investment in Saudi Arabia, TradeArabia, 27 October 2026
    • Snowflake commits $300m to Saudi Arabia, Middle East AI News
    • Libya Adopts the Ethical Charter and Artificial Intelligence Strategy, Libyan Government Information Authority, 2 June 2026
    • Libya Adopts AI Ethics Charter and Launches National AI Strategy 2026 to 2030, iAfrica