Why the Companies That Bet Biggest on AI Infrastructure Strategy Are Struggling Most
In 2025, Oracle signed a $300 billion computing contract with OpenAI, positioning itself as a primary infrastructure provider for the next phase of artificial intelligence. To fund the buildout, the company cut approximately 21,000 employees, roughly 13% of its workforce.
The bet is now in serious trouble, from a direction almost nobody modeled. Oracle is building a nearly one-gigawatt data center in Wisconsin to help fulfill the OpenAI contract. When credit rating agencies reviewed the company’s financial position, Oracle’s rating was downgraded to BBB-, with the downgrade attributed directly to heavy AI spending and uncertainty about the ability to generate profit from it. Under Wisconsin regulations, that downgrade now requires Oracle to post more than $7 billion in collateral just to connect the facility to the power grid, with ongoing maintenance costs exceeding $100 million a year.
This is not a story about AI failing to work. It is a story about capital allocation outrunning organizational readiness. A sound AI infrastructure strategy is no longer optional insulation against this kind of outcome. It is the difference between a bet that compounds and one that unravels.
The Scale of an AI Infrastructure Strategy Gone Wrong
Oracle is not spending in isolation. Amazon, Microsoft, Alphabet, and Meta are expected to spend approximately $600 billion on AI infrastructure strategy execution during 2026. This is capital expenditure at a scale that moves credit markets, not just balance sheets.
Microsoft offers a second case study, distinct from Oracle’s balance sheet story but no less instructive. Microsoft’s stock is down more than 24% over the past twelve months, a steeper decline than the rest of the Magnificent Seven. Despite $190 billion in AI infrastructure capital expenditure this year, the company remains capacity constrained. It is now shopping for cloud capacity from competitors, including Amazon and Google, according to people familiar with the discussions. CFO Amy Hood confirmed publicly that Microsoft prioritizes its own AI products for scarce computing resources before allocating what remains to Azure customers, a tradeoff that contributed to one of the company’s steepest post-earnings stock declines in years.
Internally, the pressure has produced a return to stack ranking, the controversial performance review system from the Steve Ballmer era, and a culture one former Microsoft executive described as reminiscent of “the old Windows era, where you lead with a lot of fear and a billy club in your hand.”
Two of the most sophisticated, well capitalized technology companies in the world made AI bets large enough to move credit ratings and stock prices. Both are now managing consequences that were not in the original plan.
The Proof Gap
This is the section that answers the question every board is quietly asking: is any of this actually working?
- 42% of companies abandoned most of their AI initiatives before reaching production in 2025, up from 17% the year before, and 46% of projects were scrapped between proof of concept and broad adoption (S&P Global Market Intelligence, Voice of the Enterprise: AI and Machine Learning 2025, survey of 1,006 enterprises across North America and Europe).
- 95% of AI pilots delivered no measurable profit and loss impact, based on analysis of 300 public AI deployments (MIT Sloan, The GenAI Divide: State of AI in Business 2025).
- Only 25% of AI initiatives delivered the return on investment that was expected, according to a survey of 2,000 global CEOs (IBM CEO study, 2025).
The most defensible framing of this data is not that AI does not work. It is that capital expenditure is highly visible while value capture remains lagging and uneven across firms. Most organizations have not yet built the operating discipline to make AI pay off at the scale they are spending on it.
Investor Michael Burry recently put it bluntly, stating publicly that Wall Street has turned on Big Tech’s AI spending because “the market has voted,” favoring demonstrated returns over continued capital expenditure. That signals this concern has now reached public investors, not just internal skeptics inside the companies making these bets.
The Governance Vacuum
Pearl Meyer’s survey of 108 executives and board members found that only 32% could say the C-suite as a group is accountable for AI strategy, and that boards and executives frequently disagree on who actually owns the decision.
At the same time, ManpowerGroup’s 2026 Talent Shortage Survey of 39,000 employers across 41 countries found that AI skills have become the hardest skill set to recruit globally for the first time, surpassing traditional engineering and IT capabilities. ManpowerGroup’s chief executive Jonas Prising framed the shift plainly: AI is not replacing jobs so much as reshaping work, and the companies that connect productivity gains to real opportunity and career growth will be the ones best positioned to compete.
Put plainly, companies are spending record amounts on AI infrastructure while admitting they cannot find the people qualified to run it, and cannot agree internally on who is accountable for the outcome. That is not a technology gap. It is a leadership structure gap, and it sits at the center of any credible AI infrastructure strategy.
Companies are spending record amounts on AI while admitting they cannot find the people to run it, and cannot agree on who is accountable for the outcome.
What the Data Actually Rewards
Not every organization is caught in this trap. PwC’s 2026 Global AI Jobs Barometer, published June 15, 2026, offers the counter-evidence. Professionalized jobs, meaning roles redesigned around AI collaboration rather than replaced by it, are growing twice as fast as democratized jobs, with 42% higher wage growth since 2021. New tasks being added to AI-exposed roles are 2.5 times more likely to rely on empathy, judgment, and creativity than tasks in roles with less AI exposure.
Walmart is the clearest large-scale case study. Rather than treating its 1.5 million associates as a cost center to shrink, the company deployed AI tools across its workforce that cut shift-planning time for team leads from 90 minutes to 30 minutes and made product search up to 75% faster, outcomes the company has documented directly.
The pattern holds across the data available to us. The organizations seeing measurable return from AI are consistently the ones that redesigned roles around human and AI collaboration deliberately, rather than defaulting to headcount reduction as the AI strategy itself.
The Board’s Real Job
The boards and CEOs navigating this well are no longer asking whether to invest in AI. That question is settled. They are asking a harder one: who in this organization is accountable for making sure the investment produces what it promised, and does that person have the standing to say no to spending that will not pay off.
That is not a CFO question alone, and it is not a CIO or Chief AI Officer question either. It requires an executive who can hold operational leadership and AI governance at the same time, without mistaking enthusiasm for the technology for a plan to deploy it responsibly. This is what Hager Executive Search calls the Platinum Knowledge Worker, and it is precisely the profile most search processes are not built to identify, because most evaluate AI fluency as a checklist item rather than a form of judgment.
Distinguishing genuine AI judgment from performed familiarity with the technology is exactly what Hager’s Cognitive Search Methodology was built to do. For a deeper look at how to evaluate that distinction in C-suite candidates, see How to Hire an Executive Who Actually Understands AI.
This applies whether the organization is a Fortune 500 company retrofitting AI into decades-old workflows or a much smaller company built AI-native from its first hire. The accountability question does not change with company size. Only the starting conditions do, a distinction worth its own discussion.
The Board Question for the Next Meeting
Before approving the next AI infrastructure or restructuring decision, every board should be able to answer one question with real specificity: who is personally accountable if this investment does not deliver, and what authority do they have to change course before the cost becomes irreversible.
If the honest answer is “the C-suite as a group,” that is not accountability. It is diffusion of responsibility at exactly the moment clarity matters most. A sustainable AI infrastructure strategy starts by closing that gap before the next dollar is committed, not after the credit rating reflects it.
About Hager Executive Search
Hager Executive Search is a San Francisco-based retained executive search firm serving growth-stage and mission-driven organizations. Three-time Forbes-recognized among America’s Best Executive Search Firms, Hager combines management consulting rigor with retained search to diagnose organizational gaps before launching a search, including whether candidates bring genuine AI judgment or only its appearance. Learn more at hagerexecutivesearch.com.
All data and quotes in this article are linked to their original source at first mention. Full reporting on the Oracle and Microsoft case studies is available via the Jerusalem Post and Business Insider respectively.