As industry pressures mount, the margin for error has narrowed, writes Hanno Van Niekerk, Market Leader – Sub Saharan Africa

The world is in the middle of the largest LNG infrastructure buildout in history. Dozens of mega-projects are either under construction or in advanced planning. Across the globe, capital commitments exceed $90 billion.

A massive surge of LNG capacity is slated to come online between 2026 and 2028, led by the US, Australia and Qatar.

As Shell's CEO Wael Sawan puts it, LNG is the "biggest contribution to the energy industry over the next decade."

The engineering, procurement, and construction (EPC) firms tasked with executing these projects are relentlessly under pressure. They must deliver on schedule, on budget, and to a standard that satisfies both investors and operators who will run these assets for decades.

In LNG, time really is money. A modern liquefaction facility can easily cost $30-40 billion and take three to five years to build. Once operational, every week of delay can cost tens of millions of dollars.

The overarching issue? Many of them are still managing projects with outdated tools and technology. So, the opportunity lies not in the projects themselves – but in the execution gap.

The data foundation question

EPC and operator organizations are swimming in data: sensor information, engineering data, maintenance records, and operational logs. But a lot of this data is inconsistently governed and can't be accessed in a way that produces useful operational insight at speed.

Deloitte's 2026 Oil and Gas Outlook identifies this as a defining challenge: firms face "shifting policies, rising costs, and new opportunities in LNG and digital transformation, requiring agility."

This highlights a common organizational blind spot. The right approach is a single operational view – one platform providing integrated, real-time visibility across the asset, where every employee is working from the same picture.

A unified data environment tells businesses where the data lives, who has access, and how it's been transformed. Without that, engineers and data scientists can't work efficiently – AI models are only as good as the data underneath them.

Thai chemical giant SCG Chemicals represents a brilliant case in point. Its leadership set a "zero unplanned downtime" target ­– a vision that would have seemed untenable without the right technology.

By implementing prescriptive AI on top of a robust data platform, they increased plant reliability from 98% to 100% and achieved a 9x return on investment within six months of deployment. Their teams can now predict equipment health, monitor performance in real time, and drive continuous optimization – all from a single platform.

In another example, Brazil’s AP Consultoria e Projetos shows how execution speed can be structurally improved through digitization. The multidisciplinary EPC firm adopted unified engineering platforms to tackle execution bottlenecks.

By moving engineering workflows to the cloud, AP Consultoria created a single, shared data environment where civil, mechanical, piping and instrumentation teams could work in parallel rather than sequentially.

Pipe support design was automated using AI models trained on historical engineering expertise – cutting analysis time by 90% and stress analysis review time by 60%.

Faster documentation, fewer design revisions and less rework means projects move quickly from design through to construction readiness.

Radical collaboration as competitive model

Deloitte's 2026 outlook notes that this year will likely see AI technologies "move from pilots to enterprise-wide deployment" as companies build on early successes. The firms driving that transition share a common trait ­– they aren't locked into single-vendor ecosystems.

Radical collaboration means building open, agnostic platforms that bring ecosystems together rather than locking customers in. It means integrating operational data with enterprise data – ERP, engineering systems, geospatial data, weather – to build AI models that reflect the full complexity of the real world.

This entails connecting operational technology with IT across a single governed environment. Customers want flexibility across equipment, platforms, and ecosystems – with data that stays secure, governed, and AI-ready.

For EPCs, this means engineering data built during construction doesn't get lost at handover. It also means the workforce is empowered by AI to make better decisions faster.

A field engineer on-site can ask, "How do I perform maintenance on this equipment?" and get an answer pulled from internal manuals and databases, tailored to that company's knowledge base. This capability is available now, and it changes what's possible for the people managing these assets day to day.

As the industry approaches a skilled labour cliff, AI empowers operators to maximize productivity – without depending on expertise that is increasingly difficult to find and retain.

The window is open

The IEA forecasts 7% growth in global LNG trade in 2026. The projects are moving. The EPCs that secure the next wave of LNG contracts will be the ones that can demonstrate they delivered on schedule, handed over intelligently, and gave operators the tools to reach full production fast.

Industrial intelligence and radical collaboration are the practical answer to the most expensive problem in LNG right now: closing the gap between what these projects promise and what they deliver.