Building the organizational capability behind the technology
Meta expects its capital expenditure during 2026 to reach between $125 billion and $145 billion.
At the midpoint, that is approximately:
- $370 million every day
- $15 million every hour
- More than $250,000 every minute
Meta is only one part of a much larger investment cycle. When hyperscalers, semiconductor manufacturers and other infrastructure providers are included, The Josh Bersin Company estimates that the 2026 AI infrastructure investment run rate is approaching $1 trillion a year.
The scale is extraordinary. But the more important question is not how much companies are spending.
It is:
What must organizations do to turn all this technological capacity into business value?
The answer will not be found in data centres, chips or AI models alone.
Technology creates capacity. Organizational capability determines whether that capacity produces a return.
Capacity is not the same as value
A capital investment gives a company the ability to do something it could not do before, or to do more of something it already does.
A new factory creates production capacity. A distribution centre creates logistics capacity. An aircraft creates passenger capacity. An AI data centre creates computing capacity.
But none of these assets creates value automatically.
A factory will not produce a return if customers do not want what it makes. An aircraft will not create value if its routes are unprofitable. AI infrastructure will not deliver a return simply because employees have access to more computing power.
The return depends on the decisions made around that capacity: which customer problems it will solve, which processes must be redesigned, what should be automated and where human judgment remains essential. It also depends on whether the organization can measure the resulting effect on productivity, revenue, cost and risk.
These are not primarily technology questions.
They are business questions.
This is the asymmetry many organizations risk overlooking. They are investing heavily in AI infrastructure, platforms and applications, but often much less deliberately in the organizational capability required to use them well.
The infrastructure may appear on the Balance Sheet. The judgment, shared understanding and alignment needed to make it productive will not.
Both are essential to the return.
Start with the business problem, not the tool
AI discussions often begin with what the technology can do. It can summarize documents, analyse data, write code, create content and automate workflows.
These capabilities are impressive, but they do not constitute a business case.
The better starting point is the performance problem. What is the organization trying to improve? What is preventing stronger performance today? What value would solving the problem create, and what new costs or risks might the solution introduce?
Research from The Josh Bersin Company makes the principle clear: strong AI implementations begin with business problems and outcomes, not with the technology itself. Technology-led implementation is one of the fastest routes to disappointing results.
This sounds obvious, but it requires discipline.
When a powerful new technology becomes available, activity can easily be mistaken for progress. Pilots multiply. Teams acquire licences. Employees experiment with tools. Innovation targets are announced.
Some of this exploration is necessary. But unless it is eventually connected to customer value, operating performance or strategic priorities, greater adoption does not necessarily mean greater business impact.
The purpose of AI investment should not be to use more AI. It should be to improve how the business performs.
Faster execution raises the value of good judgment
AI is changing the speed at which work can be completed.
Individuals can analyse more information, generate more alternatives and execute tasks more quickly than before. This creates enormous potential, but it also introduces a less discussed risk.
When execution was slower, weak decisions often had time to surface. People could challenge assumptions, notice unintended consequences and correct course before a poor idea spread too far.
When execution accelerates, poor judgment can scale before anyone notices.
An organization may use AI to solve the wrong problem more efficiently, automate a process that should have been redesigned or optimize one function while creating costs elsewhere. It may generate more output that customers do not value or act before different parts of the organization are aligned.
Access to more computing power does not automatically lead to better business decisions.
It can also help organizations make poor decisions faster and at greater scale.
That is why business acumen becomes more important as AI advances, not less.
AI can generate options, recommendations and forecasts. People must still determine which questions matter, which trade-offs are acceptable and which decisions support the wider business.
Three capabilities that help make AI investment pay
Three organizational capabilities are especially important in moving from technological possibility to measurable performance:
- Business acumen
- Work redesign
- Cross-functional alignment
These capabilities overlap, but each addresses a different part of the investment challenge.
1. Business acumen: understanding how value is created
Most organizations are filled with capable functional experts.
Engineers understand technology. Salespeople understand customers. Finance understands financial reporting and capital allocation. Operations understands capacity, quality and delivery. HR understands people and organizational systems.
The problem is rarely a complete lack of expertise. It is that people often make decisions through a strong functional lens without a shared understanding of how those decisions interact across the enterprise.
A technically impressive AI initiative may still fail commercially. A productivity improvement in one department may increase cost or complexity somewhere else. An automated customer process may reduce operating expense while damaging customer loyalty. Faster sales activity may increase revenue while placing new pressure on delivery capacity and working capital.
Business acumen helps people see these connections.
It is more than financial literacy. It means understanding how customer value, revenue, cost, assets, cash flow and risk connect across the business. It also means recognizing that a decision that improves one number today may weaken another part of the system tomorrow.
Klas Mellander described business as a continuous transformation of resources. Money becomes materials, equipment, inventory, capabilities and ultimately customer value.
The important question is not simply how much was spent. It is what that spending becomes and whether it helps the organization generate greater value over time.
The same principle applies to AI. The technology is a resource. Its value depends on how it is used.
2. Work redesign: changing the system, not just adding a tool
Many organizations begin their AI journey by adding tools to existing jobs and processes.
A manager creates a report faster. A programmer produces code more quickly. A marketing team generates more content. A customer service representative receives suggested responses.
These gains matter, but they represent only an early stage of transformation.
The larger opportunity appears when the organization asks a more fundamental question:
If we designed this work today, with AI available from the beginning, would we organize it in the same way?
Often, the answer is no.
Existing workflows were created around constraints that may no longer apply. Information was difficult to access. Analysis required substantial manual effort. Expertise was concentrated in a small number of people. Work moved through multiple systems and handovers because there was no practical alternative.
AI changes many of these conditions.
That means organizations need to examine work from end to end rather than automating isolated tasks. They must decide what can be automated, where people and AI should work together and where human judgment adds the greatest value. They also need to reconsider approval steps, responsibilities, governance and performance measures.
This is work redesign, not tool deployment.
It cannot be delegated entirely to the technology department because it affects structures, roles, incentives and decision rights. Nor can it be driven by HR alone.
Business leadership, technology, operations, finance and HR all need to work from the same intended business outcome.
Without that shared starting point, AI is easily added to old processes without addressing whether those processes still make sense.
3. Cross-functional alignment: ensuring decisions hold together
AI increases what individuals and functions can do independently. But businesses do not succeed through independent optimization. They succeed when decisions across the organization reinforce one another.
Sales can use AI to identify more opportunities, but operations must be able to deliver. Procurement can reduce purchasing costs, but product quality and supply resilience must be protected. Finance can reduce spending, but the organization must retain the capabilities needed for future growth.
Each functional decision can appear rational on its own and still weaken the business as a whole.
This risk increases when every function uses AI to improve its own metrics faster than the organization can align around enterprise outcomes.
Cross-functional alignment requires more than communicating the strategy clearly. People can receive the same information and still interpret it differently, prioritize different outcomes and make conflicting decisions.
Alignment is built through dialogue, experience and reflection. Teams need opportunities to surface assumptions, work through disagreements and experience how decisions in one area affect performance elsewhere.
This is where shared mental models matter.
People do not need to agree on every decision. But they do need a common understanding of how the business creates value, which constraints are real and which trade-offs are acceptable.
That shared understanding allows teams to disagree productively while still reasoning from the same picture of the business.
Without it, AI increases execution speed while coordination becomes the bottleneck.
The true investment is larger than the infrastructure
The distinction between capital expenditure and operating expenditure is useful, but it can obscure the full investment picture.
Data centres, equipment and some technology infrastructure may be recorded as assets and depreciated over time. But making that infrastructure productive requires continuing expenditure on energy, maintenance, software, data, cybersecurity, integration, specialist talent and governance.
It also requires changes in how people work.
Capital investment therefore creates future operating expenditure. A complete business case cannot evaluate the physical infrastructure separately from the organization required to operate it.
This is particularly relevant to HR and L&D.
Buildings, servers and machinery can appear as assets on the Balance Sheet. Most spending on learning and organizational capability is recorded as a current operating expense.
That accounting treatment is necessary, but it can influence how organizations talk about spending.
Technology is described as an investment.
Learning is described as a cost.
Yet accounting classification does not tell us which spending will ultimately create value.
A server is not automatically a good investment because it is capitalized. A learning programme is not automatically a good investment because it develops people.
Both must produce a meaningful return.
Some technology becomes expensive, underutilized capacity. Some learning produces attendance records but little change in performance.
The real distinction is not between physical assets and people.
It is between spending that creates useful future capability and spending that does not.
Why AI training alone will not make the investment pay
Organizations clearly need to help employees use AI responsibly and effectively. People require practical guidance on security, verification, appropriate use and the integration of new tools into their work.
But knowing how to operate AI does not tell someone which business problem is worth solving, which output should be trusted or which trade-off supports the strategy. It does not reveal whether a productivity gain creates customer value or merely shifts cost elsewhere.
These are judgment questions.
Judgment develops when people make decisions under realistic conditions, experience the consequences and reflect on what happened.
An explanation can create knowledge. A framework can create useful language. But capability requires people to apply that knowledge when goals conflict, resources are constrained and outcomes unfold over time.
This is why experiential learning has an important place in AI transformation.
People need opportunities to test assumptions, make trade-offs and discover how decisions in one part of a business affect another. That is how abstract understanding becomes usable judgment.
A different role for HR and L&D
HR and L&D have an important role in making AI investment pay, but that role cannot be limited to providing courses about the technology.
Their larger contribution is helping build the human and social infrastructure around it.
This requires moving from a content-first approach to a performance-first approach.
Instead of beginning with “What training should we provide?”, the conversation should begin with the business outcome. What is the organization trying to improve? What is preventing stronger performance? Is the real barrier knowledge, skill, process, incentives, structure or alignment?
Training may be part of the solution, but it should not be assumed to be the whole solution.
L&D can also create environments where leaders and teams practise difficult decisions before applying them at scale. HR can help redesign roles, responsibilities and workflows so that AI and human judgment complement one another rather than compete.
This is a shift from delivering learning activity to developing organizational capability.
It also requires stronger business acumen within HR and L&D themselves. To influence work redesign and performance, they must understand how the organization creates value and speak credibly about business outcomes, not only learning outcomes.
This shift is not simply about making learning more engaging. It is about connecting capability development directly to performance.
Research from The Josh Bersin Company indicates that organizations with the most advanced corporate learning capabilities are six times more likely to meet or exceed their financial targets. The implication is not that learning alone produces those results. It is that organizations that treat learning as part of the performance system tend to build stronger capabilities for execution, adaptation and change.
That is a very different proposition from measuring success through participation, course completion or content consumption.
What a capable organization looks like
A capable organization does not necessarily have the most AI tools.
It uses AI with greater clarity and discipline.
People understand how an initiative supports customer value and business performance. Functions recognize their interdependencies. Leaders distinguish activity from impact. Workflows are redesigned rather than simply automated.
Teams can move faster without losing sight of cash, profitability, customers and risk. They know when to rely on AI, when to challenge it and when human judgment must remain decisive.
But perhaps the clearest test is this:
When leaders face pressure in different parts of the organization, can they still reason together and arrive at decisions that reflect a shared understanding of how the business creates value?
If they can, decisions are more likely to remain coherent across functions and levels.
If they cannot, alignment will always be fragile, regardless of how quickly the organization can execute.
That shared ability to reason, decide and act is organizational capability.
Turning technological capacity into business value
According to Gartner, hyperscalers could invest $6.3 trillion in AI infrastructure by 2030. Whether that forecast proves accurate remains to be seen.
But the underlying challenge applies to every organization investing in AI.
Technology must eventually create value.
That value will not come simply from purchasing access to more powerful systems. It will come from choosing the right business problems, redesigning work, strengthening judgment and aligning decisions across the organization.
Building the organizational capability behind the investment
Making AI investment pay requires more than technical knowledge. Leaders and teams need to understand how the business creates value, make decisions across functional boundaries and recognize how choices involving people, capacity, customers and capital affect one another.
This is where experiential learning can make an important contribution.
Celemi’s business simulations place participants inside a realistic business system. They make decisions together, experience the consequences and reflect on how their choices affected the wider organization. This helps turn abstract ideas about investment, strategy and organizational capability into shared, usable understanding.
Different simulations address different parts of the challenge.
Celemi Apples & Oranges™ helps participants understand how investments and operational decisions affect profitability, cash flow and the Balance Sheet.
Celemi Decision Base™ develops strategic thinking and cross-functional understanding by asking teams to make capital investment decisions while balancing markets, customers, capacity, people and financial performance over time.
Celemi Tango™ helps leaders explore how decisions about customers, projects and people affect financial performance and the development of intangible assets such as know-how, reputation and long-term organizational strength.
The purpose is not to provide one predetermined answer to AI transformation. It is to help people develop the shared mental models and judgment needed to make better decisions as technology, work and organizational priorities continue to change.
Technology, such as AI investments creates possibility.
Organizational capability turns it into value.
Understanding the business is only the beginning.
The real capability is knowing how to act on that understanding together.