Annie Tsai

From multi-billion-dollar CapEx surges to sweeping org changes, the recent weeks reveal a critical question at the heart of the AI economy: how are companies funding all of this infrastructure required to compete?

This week, we look at where the money is going, what’s being cut or reshaped to pay for it, and how that tension is showing up inside balance sheets, boardrooms, and team rosters.

The Infrastructure Bill is Coming Due

Meta is raising $29 billion in private credit to fund hyperscale data centers and energy-intensive infrastructure. Alphabet is upping its 2025 CapEx to $85 billion. Microsoft is spending close to $100 billion. Amazon is right there too. These are not software budgets. These are infrastructure plays at energy-sector scale. Read More → Read More →

Morgan Stanley estimates that over $3 trillion will be spent on AI infrastructure globally by 2028, including $1.6 trillion to buy GPUs from NVIDIA and others. The majority of that is being financed rather than coming from cashflow. Putting aside the question of real need as the article speculates… Read More →

Who’s Paying for All This? A Look at Human Capital

Automation helps. Model efficiency helps. But those alone don’t cover the bill. So the question has to be asked: Is a tradeoff being made between investing in AI infrastructure and retaining staff?

Microsoft: Balancing Profit With Pressure

In July, Microsoft announced another wave of layoffs, eliminating more than 15,000 roles despite strong quarterly results. CEO Satya Nadella called layoffs the:

“…enigma of success in an industry that has no franchise value.”

Nadella emphasized that while profits remain strong, Microsoft is investing heavily in AI infrastructure, including an estimated $80 billion in CapEx for 2025. He also notes that overall headcount remains “relatively unchanged” despite the cuts. Read More →

And yet, a few weeks earlier, Microsoft President Brad Smith acknowledged that record capital expenditures – largely around AI infrastructure – have placed pressure on the company to reduce operating costs. Read More →

Intel: Refusing to Write Blank Checks

At Intel, CEO Lip-Bu Tan announced that 15% of the company’s global workforce would be cut as part of a sweeping strategic reset. His message was direct:

“Every investment must make economic sense. There are no more blank checks. We will build what our customers need, when they need it, and earn their trust through consistent execution.”

While I don’t disagree with removing reporting layers in service to gaining efficiency and velocity, it’s worth noting that in this moment, the horizons for building and executing shareholder value driving activities seem very different when it comes to AI versus human capital. Perhaps this is what is driving some of the cognitive dissonance being experienced more broadly by the workforce en mass? Read More →

Is the Pressure Indeed Financial?

If these decisions indeed are in part less about existing known productivity gains and are more about managing capital ahead of massive expenditures eventually driving massive organizational reinventions, then leaders, you have your work cut out for you. It’s true, the cost of infrastructure including depreciation, GPU leases, cloud service contracts, data center builds and more is growing rapidly. That capital has to come from somewhere.

For Managers and Teams: How to Signal Strategic Value

In this environment, it’s critical that you are both productive and actively demonstrating how your work supports financial, operational, and strategic goals of your company. If you haven’t already, consider implementing:

• Link your work to cost drivers or value generators. Show how you’re optimizing costs and/or driving revenue. Got dashboards? Add it there and make it a part of the ongoing operational conversation.
• Prioritize compute-aware decisions. Engineering, ops, and product leads should be thinking in terms of model efficiency, workload scheduling, and infrastructure ROI. It is easy to scale up and down in the cloud – only spin up what you need and no more.
• Document and surface impact on an ongoing basis. Assume leadership is not monitoring your work and instead actively tell the story of how you and your team contribute to the AI economy – operationally and financially and leading AI-first mindset adoption organizationally.

The Takeaway

AI infrastructure has moved over to becoming a core business expense and companies are making visible, and often difficult, tradeoffs to pay for it. This tension is showing up in CapEx line items, org charts, and headcount decisions.

Whether you’re leading strategy or shipping daily ops, this moment calls for clarity: understand how your work ties into both the economics and execution of AI. Show how you reduce drag, create value, and/or drive alignment. In an environment where capital is increasingly earmarked for compute, teams that can make their impact both obvious and strategically essential will be the ones that endure.