Everything AI announced at LEAP 2026

An AI-heavy LEAP, catalogued LEAP 2026 produced a dense cluster of AI-related announcements, spanning new cloud regions, expanded compute capacity, model partnerships and data-centre commitments, according to my.gov.sa's official summary of the event . Individually, TechScoop has covered several of these announcements in depth; this piece pulls them together into a single, definitive rundown of what was actually announced, organised by what each piece of news actually changes rather than by who announced it. HUMAIN at the centre of almost everything HUMAIN, Saudi Arabia's state-backed AI company, appears as a named partner in several of this year's LEAP announcements, including the AWS AI-zone expansion and Adobe's expanded partnership and image-model work. In each case HUMAIN is the domestic partner to a global technology provider, with the resulting infrastructure and services based in Saudi Arabia — a pattern TechScoop looks at more closely in a separate piece on how HUMAIN is becoming the centre of Saudi Arabia's AI strategy . New cloud regions from AWS and Microsoft Two of the most concrete infrastructure announcements at LEAP were competing hyperscaler cloud regions. AWS confirmed it will launch its Saudi cloud region in December 2026, backed by a planned $5.3 billion investment, as TechScoop covered in its dedicated piece on the AWS region timeline . Microsoft, meanwhile, set its own Saudi Azure region for a November launch, arriving even earlier than AWS's, per TechScoop's coverage of Azure's Saudi region timing . Both regions matter beyond bragging rights over launch dates: they give enterprises and government bodies the ability to run workloads, and in particular AI workloads, with data kept inside Saudi borders — an important consideration for regulated industries subject to data-residency requirements. AWS and HUMAIN expand a dedicated AI zone Beyond the general-purpose cloud region, AWS and HUMAIN separately announced an expansion of a dedicated Saudi AI zone to as much as 50 megawatts of capacity, built around Amazon's Bedrock AI platform, the Arabic-focused ALLAM model, and HUMAIN's own compute fabric, with a target tied to 2028, according to TechScoop's coverage of the AI zone expansion . That announcement is distinct from the general cloud region news precisely because it is purpose-built for AI workloads rather than general enterprise computing, and the megawatt figure attached to it is a direct proxy for how much AI training and inference capacity the zone can ultimately support. Adobe's multi-billion-dollar bet on localised generative AI Adobe used LEAP to commit more than $4 billion to an expanded Saudi partnership built around Firefly, its generative AI tools, and Express, its design app, integrated with HUMAIN and aimed at a base of 27 million eligible users, as TechScoop detailed in its piece on Adobe's expanded Saudi commitment . The scale of that figure places Adobe among the largest disclosed corporate commitments at the event, and it is notable for being explicitly built around locally contextualised generative AI rather than a generic global product rollout. An AI image model designed around Saudi and Arab culture A more specific piece of that Adobe partnership is a dedicated AI image model, built through Adobe's Firefly Foundry working with HUMAIN, designed around Arabic-language prompting and local cultural context, as TechScoop covered in its piece on the culturally contextualised model . That puts localisation at the centre of the project rather than treating Arabic support as an add-on. AMD and Cisco's presence in the AI conversation AMD and Cisco were also named among the technology companies participating in LEAP's AI-focused announcements, according to my.gov.sa's summary of the event, though the fuller scope of their individual commitments was less granular in public materials than the AWS, Microsoft and Adobe deals. Their presence nonetheless underscores how broad LEAP's AI-vendor participation was this year, spanning chipmakers and networking infrastructure providers alongside the hyperscalers and software companies carrying the largest headline figures. Data-centre capacity to support all of it None of the above works without physical data-centre capacity to house it, and LEAP's announcements extended into that layer too: MIS outlined a $1.2 billion expansion plan targeting 192 megawatts of capacity, covered in TechScoop's piece on MIS's data-centre expansion , while NHC Innovation detailed an $800 million Khuzam Digital Valley project with a path to roughly 65 megawatts of potential capacity by 2033, as TechScoop reported in its coverage of the Khuzam project . Both are part of the same underlying story: AI ambitions at this scale require physical infrastructure — land, power and cooling — built out in parallel with the cloud and model announcements getting most of the headlines. TechScoop connects that infrastructure layer directly to the power and cooling constrai