Feature
White Paper
The Future of Leadership Development: AI-Driven Manager Enablement
Uber spent the last year rewarding engineers for burning through AI tokens as fast as possible.
In June, the People team got the bill.
That bill was 23% of the People and Places division, roughly 340 jobs across HR, recruiting and workplace functions. Uber's official explanation is org complexity: teams had grown fragmented, operating too far from the business. That may be accurate. It also describes almost every HR function that has grown quickly and hasn't had its budget scrutinized by a new executive yet.
The more useful context is that Uber's CTO disclosed that the company had burned through its entire 2026 AI coding budget in four months, and the company was scrambling to install spending caps. A few weeks later, 23% of HR was gone.
This story isn't just about Uber. It's a story about where AI budget decisions get made, who makes them and who pays when they misfire. HR is almost never involved in that first decision. And that needs to change.
What the Uber Memo Said, and What It Didn't
The official explanation for the June 3 cuts to Uber's People and Places division came from Jill Hazelbaker, who had been named president and chief corporate affairs officer on May 11, the same day longtime CPO Nikki Krishnamurthy departed. Three weeks into her expanded role, Hazelbaker sent an internal memo describing the function as "too complex and fragmented, with overlapping responsibilities, unclear ownership and teams operating too far from the businesses and partners they support."
That's org design language. It's also the kind of language a new executive reaches for when she's inherited a cost center without an established relationship with it.
Uber has consistently and explicitly stated that AI wasn't the cause of the layoffs or restructuring. Around the same time, the company told engineers that their budget would be capped at $1,500 per month per AI tool.
The two statements can coexist without a connection. The question for HR leaders isn't whether Uber is being straight with us. It's whether something created the conditions that made cutting 340 people from the People division feel reasonable.
The Tokenmaxxing Tab Came Due
In 2025, burning AI tokens as fast as possible was treated as a productivity signal. Uber certainly wasn’t alone here. Amazon, Microsoft, Meta and Walmart pushed hard to get their engineers to use AI for everything. Some ran internal leaderboards rewarding heavy token use among its engineers, a practice that became known as tokenmaxxing.
In Uber’s case, the company reported 95% of its engineers were using AI coding tools, primarily Anthropic's Claude Code. AI-generated code made up roughly 10% of the codebase. Those numbers got treated as wins.
By April, the budget was gone. COO Andrew MacDonald called it a "head-exploding moment" on a podcast, adding that the company needed to start treating token consumption as a cost, not an adoption metric. Again, Uber wasn’t alone in this.
The pattern isn't new.
Early ERP rollouts produced dashboards measuring system logins instead of process improvements. Digital transformation budgets in the 2010s tracked app deployments. Each time, the activity metric replaced the value metric until the bill made that impossible. Tokenmaxxing was the AI era's version, but, because of the cost, it collapsed faster than most.
When a profitable company running 25% year-over-year gross booking growth needs to rebalance after an AI overspend, it doesn't cut engineering. It doesn't cut product. It finds the functions where the cost of cutting is hardest to measure immediately.
Every leader knows that HR has lived in that category for a long time. That’s why it’s so hard to swallow the company line that they weren’t connected.
HR Isn't in the Room When the Budget Gets Spent
Most HR leaders think about AI's impact on their function in terms of the tools they buy: better recruiting tools, smarter analytics, automated onboarding and more. That part is true. It's also not the one that determined what happened to 340 people at Uber in June.
The conversation that actually shapes HR headcount and authority is happening in finance and engineering, often without HR present. AI budget decisions, governance questions, tool adoption rates and their cost implications: these land in a CTO's report or an engineering all-hands before they ever get to a workforce planning conversation.
That's when staying in your lane stings you.
The critique HR has made for years about tracking headcount-to-HR ratios, or measuring training completion instead of skill acquisition, applies directly to tokenmaxxing as well. HR should have standing in that argument and its implication on the workforce. Most of the time, it doesn't. And that’s a huge miss.
A 2026 Careerminds survey of 600 HR professionals who had overseen AI-driven restructuring found that 90% would approach it differently in hindsight. More than half said the cuts weren't worth it. One in three companies had already rehired between 25% and 50% of the roles they eliminated. When Klarna cut its workforce down from 5,500 to 3,400 employees with public celebration, it watched customer satisfaction collapse. Then it began rehiring.
The $10 million in savings evaporated, but that didn’t hurt as much as the institutional knowledge disappearing.
Webinar
On demand
Engaged, But Not Convinced: What's Eroding Beneath Your Strongest Engagement Numbers
Perceptyx EX Benchmark '26: leadership trust, change confidence and conviction gap truths at large enterprises.
Watch Now
Webinar
On demand
Retention Is a Communication Problem: New Data from 6,000+ Employees
This session gives leaders a clear diagnosis and practical path to treat retention as an infrastructure problem.
Watch Now
Webinar
On demand
Do Learning Programs Really Work? How to Turn Education Into Engagement In Healthcare
See how leaders are using learning programs to build trust with healthcare professionals and create measurable engagement.
Watch Now
Webinar
On demand
Fuel Your Front Line: Workforce Trends Shaping 2026
Get expert insight into the tools and tactics top frontline teams will need in 2026.
Watch Now
Webinar
On demand
Fix the Disconnect Between Projects and Profit
Your margins are hiding in disconnected data. Let's find them.
Watch Now
Webinar
On demand
AI in Customer Service: Faster Resolutions, Happier Customers
Don’t let rising demand burn out your team. See how to build a smarter, more resilient support org.
Watch Now
3 Conversations HR Should Be a Part Of
HR can’t police AI, but it does have to be far enough upstream in AI investment decisions that the recalibration doesn't always land on People functions first. Here’s where HR should be showing up at the bare minimum:
-
The workforce capacity discussion. If your engineering teams are at 95% AI tool adoption and the CTO is still announcing a blown budget, that's a workforce planning signal. Adoption rate and output quality are decoupled. Someone needs to be tracking what productivity actually means when the cost of production is running unchecked. That's an HR job, and most HR teams aren't doing it.
-
The output quality discussion. HR has spent years pushing back on blunt metrics. Apply the same skepticism to AI metrics. Token consumption is not value delivered. HR teams that can connect AI tool usage to business outcomes, rather than just adoption rates, are positioned to contribute to AI conversations before they become budget crises.
-
The governance discussion. When AI spending runs unchecked and hits a wall, the correction typically hits people. HR should see that coming, understand what capacity the organization actually needs and know what would be lost by cutting it. That requires being in the room where the AI budget gets built, not just the room where the headcount reduction gets announced.
None of this requires HR to become a technology or finance function, though. Instead, it requires HR to stop treating AI budget decisions as someone else's problem.
The Cautionary Tale Is Already Written
HR leaders watching this and thinking "that couldn't happen here" should ask themselves a more specific question: Does your CFO know what the AI budget for the People function is? Does HR leadership know what the AI budget for engineering is? If both answers are no, the structural conditions that produced the Uber outcome are present.
Nikki Krishnamurthy had been Uber's CPO since 2018. She knew the People function, had relationships across the business and carried the institutional context to defend what the division actually produced. The executive who inherited HR oversight had three weeks before her first major people decision.
The AI budget is now an HR concern whether HR has claimed that territory or not. When companies overspend on AI and need to rebalance, functions without a clear line to revenue are the easiest place to justify cuts. Uber did it. More will follow.
The only durable answer is being part of the conversation before the bill arrives, with data on what the People function produces and where it can lend insight in the biggest AI questions, not just what it costs.
Waiting for that conversation to come to you is how you end up in the memo.
Editor's Note: How else is HR being pushed to adapt to AI?
- How AI Is Rewiring People Strategy and What HR Can Do to Adjust — HR leaders see AI transforming work beyond automation — reshaping teams, culture and people strategy. The future is “human-engaged” work, not human-replaced.
- If We Want AI to Help HR, HR Has to Join the Conversation — Engineers are designing AI systems to address problems that are rooted in the very systems HR understands best.
- The Chief People Officer's New Mandate — The Chief People Officer's mandate is expanding as AI transforms our workplaces. They are becoming architects of enterprise performance.