Forecast upgrades are being powered by data centres, chips, and cheap-ish money. Inside companies, AI is still failing to turn into measurable profit for most executives.

New Adobe research shows Middle Eastern enterprises have largely solved for cloud and AI adoption. The harder problem is turning that infrastructure into integrated execution and measurable business impact.

As AI demand surges, technology companies are signing nuclear and energy deals that expose a deeper truth: growth now depends on infrastructure, politics, and who controls electricity.

Data centres are becoming a major new electricity load as AI scales, forcing governments and utilities to rethink grid investment, efficiency standards and the pace of renewable build-out.

Global semiconductor revenue climbed 21% to $793.4 billion in 2025, as the AI infrastructure boom pushed Nvidia past $100 billion in chip sales and turned high-bandwidth memory into the market’s new battleground, according to preliminary figures from Gartner released on January 12, 2026.

From the US to Europe and the Gulf, AI investment is surging, but questions remain about monetisation, risk and who ultimately bears the cost.

Microsoft Report Shows 64% of the UAE Uses AI, Representing the World’s Most Advanced AI Adoption System

At CES in Las Vegas, Jensen Huang didn’t just unveil “Rubin.” He reframed Nvidia as the company selling rack-scale AI factories — chips, networking, security and software bundled into one system — because the next fight is about inference cost, power, and control of the full stack.

The UAE-based AI firm and hybrid cloud provider say the new platform will let governments and large enterprises deploy “sovereign” generative AI inside their own infrastructure, keeping sensitive data on-premise.

Nvidia’s biggest investments last year were not about faster chips or higher growth. They were about preventing loss of power as AI inference overtook training and infrastructure economics began to shift.

Meta’s acquisition of Manus, an AI agent startup with more than $120m in annual revenue, signals a shift in the AI race away from models and toward execution and monetisation.

The challenges organisations faced in 2025 were not caused by lack of innovation, but by weak foundations. AI, cybersecurity, and cloud infrastructure all exposed the same underlying problem.

The acquisition will expand ServiceNow’s security workflow offerings and advance AI-native, proactive cybersecurity and vulnerability response across all connected devices.

Gartner’s latest “Companies to Beat” analysis reveals how the AI market is shifting from experimentation to structure, with platforms consolidating at the core while competition intensifies at the application layer.

G42’s AI infrastructure arm is establishing a European base in Ireland as governments and enterprises seek compliant, high-performance compute to scale production AI.

In Salesforce’s 2025 State of Service data, AI absorbs repetitive cases while reps move up the value chain - but the constraint is usually unified data and governance, not model capability.

OpenAI’s senior deal hire, US infrastructure consolidation, India’s incumbent-led acquisitions and quiet GCC capital flows all point to the same shift: tech is moving from experimentation to institutional control.

The reported $7bn deal would be one of ServiceNow’s largest acquisitions and signals continued consolidation around security as an enterprise workflow.

OpenAI, Anthropic and their peers may be the miners, and NVIDIA the shovel seller—but the quiet force shaping the entire boom is ASML, the Dutch firm whose machines make cutting-edge chips possible. Yet as computing scales, a second kind of kingmaker is emerging inside enterprises: the governance systems that decide whether AI can be trusted at all.