Abu Dhabi-based AI fintech startup Mal pulled in $230 million in the first half of 2026, a single round that accounted for 37% of all Middle East and North Africa fintech funding, even as the region logged only 57 transactions. The numbers are neat. The power behind them is less so. A tiny set of big investors, a few favored markets, and a lot of talk about innovation are doing most of the work while the broader field thins out.
The Money Flows Upward
Dubai-based data firm MAGNiTT said MENA fintech funding fell 9% year-on-year in the first half of 2026. Its founder and chief executive, Philip Bahoshy, put the imbalance in plain language: "FinTech funding looks considerably stronger through the headline number than through the breadth of activity underneath it." He added: "A smaller number of large rounds, concentrated with one investor and one market, are carrying the [first] half’s resilience. That distinction matters more than the top-line figure."
That concentration is the story. Mal helped fintech companies in the region raise $387 million in the first half of 2026, while the UAE and Saudi Arabia captured 85% of regional fintech funding. Egypt ranked as the third-largest fintech market. In the UAE fintech scene, Fasset followed Mal with a $51 million transaction. In Saudi Arabia, Madfu and Stitch led the rounds with $26 million and $25 million, respectively. Together, those three transactions were worth $102 million, less than half the amount secured by Mal.
Payments solutions drew 28% of all deals, making them the leading fintech sub-industry. The report said deals worth less than $100 million accounted for 62.7% of total funding, compared with almost 80% in the same six months of the previous year. Bahoshy said: "The second half will show whether other investors and other markets step in behind this activity, or whether FinTech’s real health continues to be closer to what the deal count already shows."
Construction, Surveillance, and the IDF Pipeline
In Israel, Buildots announced that it completed a $130 million funding round. The real estate AI and computer vision company said the round brings its total capital raised since its founding in 2018 to about $297 million. The new funds are expected to support continued development of its technology and system, as well as employee recruitment, marketing and global expansion. The latest investment round was led by O.G. Venture Partners, with participation from Lightspeed Venture Partners, Intel Capital, Mohari Ventures, Human Capital, Qumra Capital, Poalim Equity, Viola Growth and Avigdor Willenz.
Buildots said it has developed a system that bridges the physical world and the world of AI, creating a unified shared status report for all parties involved in a real estate project, including the developer, main contractor and subcontractors. It said the system can reduce project duration by an average of about 15% through early identification of delays and coordination problems. The company said its technology is based on AI and computer vision models derived from eight years of collecting real-time data from construction sites. Its workflow includes feeding in work plans, which the system compares against on-site photography taken every few days by a project manager or a Buildots representative. The system performs comparative analysis across schedules, 3D models of construction sites and field documentation, and can identify and classify hundreds of types of construction work from video alone.
Buildots said its clients include Intel, Azrieli Group, Ashtrom, Peretz Boneh HaNegev, Tidhar and other construction companies and real estate developers worldwide, such as STO Building Group, JE Dunn, Mortenson, Bouygues and HOCHTIEF. The international real estate AI market is valued at about $405 billion as of mid-2026 and is expected to rise to up to $1.4 trillion by 2030, according to Research and Markets, with a compound annual growth rate of roughly 34% to 36%.
Buildots was founded in 2018 by Roy Danon, who serves as CEO, Aviv Leibovici, CPO, and Yakir Sudry, CTO. All three are graduates of the IDF's Talpiot R&D leadership training program. The company currently employs about 400 people, 260 of whom are in Israel and the rest in the US, Canada, the UK, France, Germany and Denmark. It is working to recruit an estimated 100 additional employees, including several dozen in Israel. Roy Danon said: "The challenges we are solving in data center construction also exist in office building, residential tower, school, and hospital projects. Many times it is discovered too late that the project is deviating from the plan. In data centers, because of the size, complexity, and investment, every such deviation generates huge losses."
Ziv Kop, managing partner at O.G. Venture Partners, said: "We have been investing for over two decades in companies that create new categories, and the successful companies are those that build a technological infrastructure through which the entire industry begins to work. Buildots' unique models and its ability to collect data from the field bring the AI revolution to the traditional construction world and the tech construction world, which today includes data centers, chip manufacturing plants, and defense-tech facilities. Buildots is doing for the construction industry what Mobileye did back in the day for the automotive world."
Saudi Arabia's AI Statecraft
In Saudi Arabia, the state-owned artificial intelligence startup Humain unveiled a new Arabic-language model developed with Chinese AI firm MiniMax. Announced during the LEAP technology conference in Riyadh on Sept. 3, the development signals that the kingdom continues deepening AI ties with China while simultaneously pursuing major US technology partnerships. Humain, a PIF-owned AI company launched last year, said it had commissioned MiniMax to deliver Humain-m3, a frontier Arabic-language model based on the Chinese company's AI systems.
The model is initially available as a research preview through Humain Node, the company's platform for giving developers, researchers and enterprises access to AI models and inference capabilities. Humain said Humain-m3 achieved the highest average score among the frontier models tested across seven public Arabic-language benchmarks, and it plans to publicly release the model after the research and evaluation phase. The language of access is open. The ownership is not.
Across the region, the same pattern keeps repeating: concentrated capital, state-linked firms, military-trained founders, and technology built to organize labor, construction, finance, and language from above. The pitch is efficiency. The structure is control.