Mubadala Investment Company grew its asset base 17 percent to $385 billion in 2025 — a figure Khaldoon Khalifa Al Mubarak, the fund's Managing Director and Group CEO, described as proof that "bolstering AI capabilities in Abu Dhabi" was not a strategic footnote but the central driver of the fund's growth narrative. The number matters to talent professionals for one reason: it is the capital base underwriting six distinct AI-focused operating entities — G42, MGX, AIQ, Bayanat, Khazna Data Centers, and M42 — that, taken individually, look like competing employers but, examined from the sovereign-fund layer, behave as a single integrated talent ecosystem. For a senior AI engineer or applied scientist in H1 2026, the operative career decision is not which Mubadala-backed entity to join. It is whether to enter the constellation at all — and if so, at which node, at which career stage.
What the Constellation Actually Looks Like
The six entities under the Mubadala AI portfolio umbrella are not interchangeable, and conflating them is the most consistent error made by Western candidates who arrive in Abu Dhabi with a CV tuned for a single hiring context. They are architecturally distinct.
G42 — in which Mubadala holds a direct stake — is the parent structure housing Core42 (sovereign cloud and AI infrastructure, 1,100-plus employees across 68 nationalities as of Q1 2026, per G42 group communications and Core42 company profile), Presight (applied AI for government and enterprise), and the Inception Institute of Artificial Intelligence, which produced Jais, the Arabic large language model. Core42 is the entity building the physical AI infrastructure for Stargate UAE: a $30 billion, one-gigawatt compute cluster in Masdar City, in consortium with OpenAI, Oracle, NVIDIA, and SoftBank, with the first 200-megawatt phase expected live in 2026. The engineering profiles Core42 is staffing now are data-center-scale AI infrastructure leads and sovereign cloud architects — roles that did not exist at scale in Abu Dhabi two years ago and that draw directly from former hyperscaler operations at Amazon Web Services, Google Cloud, and Microsoft Azure.
MGX — a technology investment company co-founded by Mubadala and G42, launched in 2024 under CEO Ahmed Yahia Al Idrissi, who leads from a three-decade combined operational and investment background — is not an operating AI company. It is a capital deployment vehicle targeting $100 billion in assets, investing across frontier AI models (OpenAI, Anthropic, xAI), AI infrastructure, and semiconductors. MGX's talent footprint is a 40-to-70-person team of AI investment specialists, deal principals, and sector experts who operate at the intersection of frontier-model evaluation and sovereign capital allocation. The roles MGX posts are not ML engineering roles; they are due-diligence-capable technical analysts who can evaluate a transformer architecture and a term sheet in the same week. At Davos 2026, Khaldoon confirmed Mubadala's position as Vice Chairman of MGX — a governance signal that the fund treats the vehicle as a strategic, not purely financial, instrument.
AIQ — the AI joint venture between ADNOC (60 percent) and G42 (40 percent), operating at the intersection of energy-sector domain data and AI engineering — is the most technically specific node in the network. AIQ builds production AI for drilling performance, reservoir modelling, and subsurface simulation on ADNOC's hydrocarbon data, which is the most extensive petroleum dataset in the Arabian Peninsula. Its hiring is narrow: machine learning engineers who can translate energy-sector operational data into production models, with a strong preference for profiles carrying prior exposure to PDO, Saudi Aramco Digital, or QatarEnergy Digital's technical pipelines. Compensation for senior ML roles at AIQ tracks the $110,000-to-$165,000 USD-equivalent band, tax-free, per ENTRA's Middle East AI Salary Index for H1 2026.
Bayanat — listed on the Abu Dhabi Securities Exchange (ADX: BAYANAT), majority-owned by G42 — is the geospatial AI and earth-observation data platform. Its technical requirements are specialist and non-overlapping with the other portfolio entities: remote sensing engineers, computer vision specialists with satellite imagery domain expertise, and geospatial data scientists. Bayanat's ADX listing provides public-company equity visibility that the unlisted portfolio entities cannot match, and its hiring has been running quietly through Q1 and Q2 2026 in roles that rarely surface in general AI job boards.
Khazna Data Centers — the G42-anchored data center infrastructure provider contracted to build and operate Stargate UAE's physical layer — is the construction and operations engineering node. Its 2026 hiring is the most volume-intensive in the ecosystem: data-center build management, power systems engineers, network infrastructure leads, and sovereign cloud operations roles. Khazna's compensation structure for senior engineering roles is tracking AED 45,000-80,000 per month base, tax-free, per ENTRA Middle East AI Salary Index H1 2026, with the Stargate UAE contract creating a sustained three-to-five-year demand horizon that the company has not faced at this scale before.
M42 — Mubadala's healthcare and life sciences AI vehicle — is the node least covered in AI talent journalism and the most differentiated for applied AI engineers interested in clinical and genomics domain work. M42 operates SEHA hospitals, Cleveland Clinic Abu Dhabi, and the Aster DM network under a unified AI transformation mandate, with ML engineers building diagnostic imaging models, clinical workflow optimisation tools, and genomic sequencing pipelines. The profile M42 is recruiting in 2026 is narrow: healthcare AI engineers who combine model deployment experience with regulatory compliance literacy — profiles that have become scarce in the US market as large hospital systems compete for the same talent class.
Why This Reads as One Market, Not Six Employers
The mechanics that make this a talent ecosystem rather than a collection of discrete employers are institutional, not organic.
Three structures are running simultaneously. First, the ADGM Academy partnership with Abu Dhabi's Department of Government Enablement (DGE) Mawaheb Talent Hub — formalized in February 2026 — is targeting 620 cross-sector job placements in 2026 across Abu Dhabi's sovereign-tech stack, with the Software Development Accelerator Programme, Technology Infrastructure Programme, and Strategic Business Intelligence Programme running as dedicated tracks feeding directly into the Mubadala portfolio entities and their government partners. The DGE Mawaheb Hub trained over 10,000 individuals in 2025 and facilitated more than 6,000 placements (per official DGE and ADGM announcement, February 2026) — with G42 a confirmed delivery partner on the remote-work and AI training tracks. The hub's 17 MoUs signed in 2025-2026, covering Al Ain and Al Dhafra regional expansion alongside central Abu Dhabi, created 3,000 new job opportunities that ENTRA's tracking indicates are absorbing into the Mubadala portfolio entities at a higher rate than any other employer category in Abu Dhabi.
Second, the MBZUAI placement pipeline, which ENTRA covered in detail in May 2026, routes approximately 56 percent of its 2026 graduating cohort into G42-umbrella or Mubadala-backed entities (per ENTRA H1 2026 graduate placement tracking) — not because of formal placement agreements, but because of the compensation and visa architecture those entities offer. The UAE Golden Visa mechanism is central here: MBZUAI employers who sponsor a fresh graduate's UAE Golden Visa within 30 days of hire are filing for a 10-year renewable UAE residency instrument at zero cost to the individual. The standard package for an MBZUAI master's graduate entering a Core42 or M42 role in 2026 runs AED 420,000 to 520,000 base per year ($114,000-$142,000, tax-free), plus AED 60,000-90,000 housing allowance, UAE Golden Visa covered, and relocation assistance. Against that package, the argument for a competing US offer requires a very specific lifestyle or project preference that compensation alone cannot resolve.
Third, MGX's own internal AI tooling — its Alpha Intelligence investment platform and its AI Talent Scout, launched as part of the firm's 2025 AI-native operations upgrade (per MGX official communications and Davos 2026 presentations) — has systematised the pipeline between portfolio company technical requirements and candidate sourcing in a way that makes the ecosystem behave more like a single large employer's internal mobility programme than six separate companies posting on LinkedIn. A principal ML engineer who cycles through a Core42 infrastructure role before moving to an AIQ energy-AI position is not changing employers in any meaningful career-narrative sense. They are deepening sovereign-tech domain expertise across the same capital stack.
Why This Matters for Senior ICs Considering Abu Dhabi
G42's own AI Talent Report, released in April 2025 in collaboration with Semafor and drawing on responses from 750 AI specialists, identified a consistent gap between what engineers value and what they currently experience: compensation ranked as the top career driver (68 percent called it important), but satisfaction with current compensation sat at only 43 percent — a 25-point gap that is structurally larger in the US and UK markets than it is in Abu Dhabi's tax-free environment. The same report flagged job security (70 percent important, 48 percent satisfied) and access to advanced AI projects (explicitly listed as a top offer driver alongside salary and benefits) as the second and third decision factors.
For a senior IC at a US frontier lab or a principal at a European hyperscaler operating in that 25-point satisfaction gap on compensation — and who is also running projects that are either behind a corporate roadmap decision or constrained by a slow-moving governance structure — the Mubadala portfolio ecosystem resolves at least two of those three concerns simultaneously. The compensation structure is verifiable: senior specialists in the Mubadala network are clearing AED 750,000-plus annually in base alone (approximately $204,000 at current rates, tax-free), with principal-level and director-level packages running to AED 110,000-250,000 per month before performance bonuses. The project access question is resolved by the infrastructure itself: a senior AI engineer at Core42 in 2026 is staffing a one-gigawatt compute cluster. The scale of the engineering problem is not abstract.
The third gap — job security — is where the sovereign-capital structure matters most. Mubadala's $385 billion AUM is not a venture fund that re-prices in a down round. The capital behind Core42, AIQ, Bayanat, and Khazna is Abu Dhabi sovereign, and the UAE National AI Strategy 2031 is the mandate. Neither of those change on a VC-cycle timeline.
What a UAE Golden Visa does not offer — and what candidates must model honestly — is the permanent-residency path available in Bahrain or the US Green Card equivalent. The 10-year renewable UAE Golden Visa is a strong instrument, but renewal depends on maintaining the qualifying salary floor, which requires staying employed within the ecosystem. For candidates who want a permanent anchor, the calculus is different than for those running a five-to-seven-year financial sprint.
Three Things to Watch in H2 2026
Stargate UAE phase-one activation. The first 200-megawatt tranche of the Masdar City compute cluster is expected to come online in 2026, triggering a second-wave hiring cycle at Khazna and Core42 for operations and reliability engineering roles. Watch Khazna's ADX-adjacent corporate announcements and G42's LinkedIn hiring velocity through Q3 for the timing signal.
MGX's portfolio talent formalisation. MGX has deployed its AI Talent Scout internally; the question for H2 2026 is whether it formalises an external-facing talent mobility programme between portfolio entities — a structure that several sovereign investment funds in Singapore (Temasek) and Norway (Norges Bank Investment Management) have piloted with mixed results. If MGX publishes a cross-portfolio internal mobility framework, it will materially change the career-narrative proposition for candidates weighing a single portfolio-entity offer.
ADGM Academy expansion. The 620-placement target for 2026 is an ADGM-stated commitment, not a projection. Monitor whether the ADGM Academy's Q3 placement reporting — typically released through ADGM's official media channel and the WAM national news agency — confirms that the Software Development Accelerator and Technology Infrastructure tracks are hitting their absorption targets inside the Mubadala portfolio entities. A miss would signal supply constraints at the graduate layer; a hit would confirm the pipeline is structurally operational.
For a senior AI engineer running the Gulf calculus in June 2026, the Mubadala portfolio is not a hiring manager to convince. It is a capital structure to position yourself inside.
Global Career PlatformFind AI talent. Find your next role.
Booking is hotels. · Airbnb is apartments. · ENTRA is global careers.