From GPUs to payments: Saudi Arabia's technology stack is coming together

A stack, not a stunt There is a temptation, in a month with this much news, to cover cloud regions and fintech funding rounds as separate beats entirely — infrastructure reporters covering one, fintech reporters covering the other. That separation misses the more useful story: neither layer is fully legible without the other, and September 2026 happened to produce enough news at every layer simultaneously to make the connections between them visible in a way they rarely are. Individually, a cloud region launch and a fintech funding round look like unrelated stories. Put enough of them side by side, as September 2026 did, and a layered technology stack becomes visible: compute and infrastructure at the bottom, financial and enterprise applications running on top, with each layer depending on the one beneath it. Layer one: compute and cloud Every layer described below depends on this one existing first, which is why it is worth starting the stack from the bottom rather than the more attention-grabbing layers above it. The base layer is physical: chips, data centres and cloud regions. AWS confirmed its Saudi cloud region for December 2026 , backed by a planned $5.3 billion investment, while Microsoft's Azure region is set for November. Beneath the cloud layer sits the physical data centre build-out: MIS detailed a $1.2 billion expansion to 192 MW of capacity, and NHC Innovation outlined an $800 million project, Khuzam Digital Valley, with a potential 65 MW and a build-out running to 2033. This is the layer that takes years and billions of dollars to build, and it is also the layer every other layer ultimately depends on. Layer two: AI models and platforms Nothing above this layer works without it, and nothing below it is worth much on its own. This layer is where raw compute capacity becomes a usable product, and it is also where most of September's largest partnership announcements sit, precisely because it is the layer both international technology companies and HUMAIN have the clearest commercial incentive to build out together. On top of that physical infrastructure sits the model and platform layer. AWS and HUMAIN are expanding a dedicated AI zone to as much as 50 MW, built around AWS's Bedrock, the ALLAM model and HUMAIN Fabric. Adobe's more than $4 billion partnership with HUMAIN brings generative AI tools, including an Arabic-culture-specific image model, to as many as 27 million eligible users. HUMAIN's strategic investments in Mozn and Arabic.AI extend that platform layer further into financial-crime detection and Arabic-language AI specifically. None of this layer works without the compute layer beneath it — models and AI zones need the physical data centre and cloud capacity to actually run on. Layer three: fintech and financial infrastructure This is the layer where the stack stops being theoretical and starts generating measurable transaction volume and revenue. Fintech is the clearest example of a sector where enough of the layers below already exist that a company can build a large, revenue-generating business today rather than waiting years for infrastructure still under construction — which is a large part of why this layer produced the month's biggest individual dollar figures outside the infrastructure announcements themselves. The third layer is where AI and cloud infrastructure meet real financial transactions. Abwab.ai's AI infrastructure for SME lending underwriting, which already sits behind more than SAR 10 billion in processed loans, needs compute to run its models. Tarabut's open banking infrastructure, processing more than 5 billion API calls, needs reliable infrastructure to connect banks and fintechs. Tabby's $233 million Series F and barq's $329.5 million Series A both fund companies whose products depend on the layers beneath them. This layer is where the abstract capacity built in layers one and two starts generating actual transaction volume and revenue. Where the mobility and construction layers fit Fintech is the clearest illustration of this stack simply because it is the furthest along, but it is not the only sector occupying this position within it, and it is worth naming the others explicitly rather than letting fintech stand in for the whole layer on its own. The stack is not limited to fintech and consumer payments. Sirdab's logistics and warehousing platform, FlyAkeed's enterprise travel product, and Fitting's construction procurement software all sit at the same "layer three" position as Tabby and Tarabut — businesses whose products depend on reliable cloud infrastructure to run, generating transaction data as they scale that could eventually support AI tools layered on top, the way Abwab.ai's underwriting product already does inside fintech specifically. None of these companies has disclosed an AI product yet, but their position in the stack is structurally identical to where fintech sat a few years before AI-assisted underwriting became a visible, funded category.