Davos 2026: AI Took Centre Stage - What Executives Should Do Next
The World Economic Forum's Annual Meeting in Davos put AI at the core of every serious conversation about growth, competitiveness and national capability. Leaders from finance, industry and government debated opportunity and risk, with a clear thread running through the week: AI is no longer an initiative - it's infrastructure, policy and P&L.
The event also called out the gap between elite debate and everyday impact. As one co-chair put it, the meeting "can't remain an echo chamber." The signal for executives: move past panel talk and translate AI headlines into operating decisions.
Elite perspectives seek relevance in real outcomes
The opening sessions acknowledged Davos' distance from those affected by its ideas. The goal set on stage was simple: fewer agreeable panels, more tension that leads to action. That's a healthy constraint for boards - pressure-test assumptions, bring in dissent, and hold leaders accountable to measurable results.
Executives across sectors returned to the same pillar: AI touches work, productivity, national competitiveness and energy. The conversation has outgrown pilots. It's now about capacity, cost, and capability at scale.
AI as national infrastructure
NVIDIA's Jensen Huang put it plainly: "AI is infrastructure… every country has its electricity, you have your roads - you should have AI as part of your infrastructure." He added that with the right local expertise, training models is "not so incredibly hard" anymore.
Huang outlined a practical stack that leaders can use to set priorities:
- Energy: Reliable, affordable electricity and cooling
- Chips: Access to GPUs/accelerators and supply commitments
- Cloud: Scalable compute and data pipelines
- Models: Foundation and domain models, updated frequently
- Apps: Tools and services people actually use
Strategy takeaway: place your bets layer by layer. Secure energy and compute first, then models and applications. Skip a layer and your plan stalls.
The energy reality: cost, capacity, constraint
Schneider Electric's leadership framed the core trade-off: AI drives growth, and it also lifts demand on grids and data centres. The twist is that AI can also optimise the same systems that feed it - from load balancing to facility efficiency.
Action for COOs and CFOs:
- Set energy KPIs for AI workloads (kWh per query, per task, per dollar of output)
- Pilot AI-driven efficiency in plants and data stacks before scaling
- Lock in multi-year energy and compute agreements with clear cost curves
- Shift non-urgent training jobs to off-peak windows and greener regions
For context on the bigger picture of energy demand, see the International Energy Agency.
Leadership fluency before enterprise rollout
Accenture's approach started with its top leaders first. Train the people setting targets and budgets, then let that fluency cascade. Without it, governance drags and pilots die in committees.
What works:
- Run an executive bootcamp: core AI concepts, risks, and use-case economics
- Publish guardrails: data use, IP, privacy, and vendor standards
- Map value: three use cases per function with expected ROI and owners
- Create a simple scoreboard: costs avoided, hours saved, revenue added
If your team needs a quick way to level up, explore executive-focused resources at Complete AI Training.
Tariffs and pricing: resilience is still a CFO sport
On trade, Amazon's Andy Jassy shared a blunt tactic: pre-buy inventory to protect customer prices under tariff pressure. It helped early in 2025, but stock ran down by autumn - a reminder that buffers buy time, not certainty.
What to set up now:
- Quarterly tariff scenarios with clear triggers for pricing and sourcing
- Flexible inventory strategies: selective pre-buy, multi-supplier lanes
- Customer messaging that sets expectations without eroding trust
- A dynamic pricing playbook that protects margin and share
Your next 90 days
- Pick your layer bets: Energy, compute, cloud, models, apps - decide where to lead and where to partner
- Stand up an AI cost model: Track unit economics per use case, not just total spend
- Secure capacity: GPU commitments and energy contracts aligned to your roadmap
- Train leaders first: 20 key executives through a focused program; publish policy the same month
- Ship two needle-moving use cases: One for productivity, one for revenue - prove value fast
The bottom line
Davos closed with clear tension and clear upside. AI is being treated like roads and grids - essential, capital-intensive and foundational to growth. The executives who win will treat it the same way: make layered bets, manage energy and compute as core inputs, and build leadership fluency before scale.
For event context and ongoing updates, visit the World Economic Forum.
Your membership also unlocks:
