Hello, dear readers!

Three years ago, Nike was the most valuable sportswear company on earth. Today it's trying to win back shelf space it gave away on purpose. The turnaround under its new (old) CEO is grinding through what he calls "the middle innings," and the channel strategy mistakes that got them here read like a case study in how to hand your competitors a head start. The Economist's Amanda Chicago Lewis published a deep look at what went wrong, and what she found belongs in a supply management textbook.

Also, the WSJ asked sixteen economists whether AI would add or subtract jobs, and their answers split three ways. Daron Acemoglu and David Autor, two offices apart at MIT, landed in different camps. What they agree on is more interesting than where they disagree.

This issue features the usual bunch of AI Insights and recommendations for the week's podcasts, books, shows, charts, and tweets, followed by a final chuckle.

Let's get going.

Industry Highlights: How Nike Handed Its Competitors a Head Start

If you'd bought Nike stock at its peak in November 2021, you'd be sitting on a 75% loss today. The shares recently hit a 12-year low. For a company that has defined athletic culture for four decades, the question everyone is asking is simple: how did this happen?

The DTC gamble

The short answer is a channel strategy gone sideways. Under former CEO John Donahoe, Nike made a bet that it could cut out the middlemen. The company pulled product from wholesale partners like Foot Locker and Dick's Sporting Goods, pushed customers toward its own stores and apps, and chased higher margins on every pair sold. On paper, it made sense. Nike Direct hit 44% of total sales by 2023. But what looked like margin optimization from the boardroom looked very different from a store shelf.

When Nike pulled back from wholesale, it left empty shelf space. And competitors filled it immediately. On Holding, the Swiss running company, hit $2.4 billion in annual sales by early 2026, growing at over 40%. Hoka carved out the "maximalist" running category almost unopposed. Adidas surged back with its Samba and Gazelle lines. Even Lululemon started making inroads in women's performance footwear. Needham analyst Tom Nikic put it bluntly when he said the prior regime completely misjudged the sneaker market: the importance of product development, wholesale relationships, and how customers actually shop.

The innovation gap

The shelf-space problem might have been manageable if Nike had kept its product pipeline moving. It didn't. In 2024, Nike launched only two new running shoes. Asics launched eleven. Adidas launched ten. Instead of investing in performance, Nike leaned on heritage models, the Air Force 1s and Dunks that had carried the brand through streetwear's golden years. By the time those retro lines started to fatigue, Nike didn't have enough new product to pick up the slack. Converse, the subsidiary, is in free fall: revenue down 35% in the most recent quarter.

China, tariffs, and the margin squeeze

Then came the external headwinds. Revenue in Greater China has now declined for eight consecutive quarters, down 10% year over year in Q3 of fiscal 2026. Local competitors Anta and Li-Ning have tapped into a rising wave of consumer nationalism, and Nike's mass-market positioning no longer carries the same cachet it once did in Chinese cities. Meanwhile, tariffs have added roughly $1.5 billion in annual costs, pushing gross margin to a five-year low of 40.2%. Full-year fiscal 2026 revenue landed at $46.4 billion, essentially flat, and down 2% on a currency-neutral basis.

The Elliott Hill reset

Nike's answer was to bring back Elliott Hill, a 32-year company veteran who had retired in 2020. Hill's turnaround plan, called "Win Now," is a return to basics: rebuild wholesale relationships, recommit to sport and performance product, and stop relying on a handful of retro franchises. The early signs are mixed. Wholesale revenue grew 4% in Q4, but Nike Direct fell 7% in the same quarter. Digital sales dropped 12% for the full year.

Hill told investors the company is in "the middle innings" of its comeback, and nobody disputes that the right moves are being made. The harder question is whether Nike can actually reclaim the ground it gave up. On and Hoka didn't just fill a shelf; they built loyal customer bases during Nike's absence. If you've ever managed a category where a key supplier pulled back and a substitute locked in, you know how that story usually ends. The incumbent has to earn the space back, and the buyer isn't sentimental about it.

The bottom line

What I keep coming back to is the channel strategy lesson. Nike optimized for margin by going direct, and in doing so it handed distribution to its competitors. When you prioritize protecting your format over getting the product into the customer's hands, somebody else will do the getting for you, and the upstarts that showed up on those empty shelves aren't giving the space back voluntarily.

I learned this the hard way at ACCO in 2005, sitting across from Simon Lu in Hong Kong, trying to renegotiate pricing on a supplier who had resisted it for years. He eventually agreed and later told me he'd said yes because I'd shown his company some respect. Then he gave me a Chinese proverb I've never forgotten: ‘Without a fight, you don't get to know each other.’ Nike didn't fight with its wholesale partners. It just left, and the relationship went where relationships go when nobody's tending them.

Hill has the playbook right. Whether he has enough time is a different question, and it's one the stock price is currently answering with a shrug.

The Future of Work: The Between Times

Fifteen of sixteen economists agreed on one thing: AI will make workers more productive. Then the WSJ asked the same group whether it adds or subtracts jobs on net, and the consensus fell apart. Eight said no change. Five said net losses. Two said growth. Daron Acemoglu, who won the Nobel in 2024, landed in the net loss camp. David Autor, a few offices away at MIT, landed in no change.

Here's the thing, though. Read the comments underneath the votes and the disagreement is narrower than the tally makes it look. Almost nobody thinks the technology stalls out. What they're arguing about is us, meaning how quickly companies reorganize and who gets left holding the bag while they do.

Ajay Agrawal has the best name for where we're sitting right now. He calls it the Between Times: everyone can see what the tools do, but organizations haven't rebuilt themselves around them yet. His example should land for anyone who buys or sells services. An insurer that shifts from repair and replace to predict and prevent isn't running the old business with better software. It's a different value proposition, and every role inside it gets redrawn.

Who actually gets hurt

Nicholas Bloom expects the first real hit to land on offshoring, and he's blunt about it. Coding, call centers, payroll, all the repetitive work we spent two decades shipping to India and the Philippines. He says he knows firms already cutting that demand and switching to AI instead. If your service delivery model was built on labor arbitrage, the arbitrage itself is what's being automated.

So who's left standing? The answers split by age in a way worth paying attention to. Autor thinks entry-level workers may actually benefit, because AI compresses the learning curve and lets a less experienced person perform above their years. Martha Gimbel, who some of you will remember from last issue, takes the other side. Young people probably get hit first, she says, but they have time to pivot. The older worker who holds on a few more years and then gets displaced has a much harder road back.

Both can be true at once. And notice that neither one is really a claim about the technology. Jed Kolko puts it about as plainly as anyone can: the people most at risk aren't the ones in the most exposed jobs, they're the ones least able to adapt.

The skill nobody put on the org chart

Raffaella Sadun's answer is the one I keep coming back to, because it describes our world exactly. As AI creates winners and losers inside a company, she says, the ability to build coalitions, manage resistance, and help people adjust becomes a premium skill. Persuasion and conflict resolution, in other words.

Ask anyone who has run a category through a stakeholder group that didn't want to change suppliers. That was always the job. The spend analysis was the easy part.

I'll admit a bias here. I changed plenty of jobs during the first part of my supply management stint. From a buyer's desk at Navistar to global sourcing at Terex, and not one of those moves happened because I'd mastered a tool. They happened because the ground moved and I went where it was moving. Every economist in that survey, whatever camp they picked, ends up in roughly the same place. Thus, adaptability is the asset. Which is a little inconvenient, since it's the one thing no procurement system has ever been able to measure.

So what do you do with a forecast that comes in three flavors? Probably not build a workforce plan around any single one of them. Assume the tools keep improving, assume your offshored transactional work is the most exposed thing you own, and put your energy into the people who tinker rather than the people with the tidiest resumes.

Or ignore all of it and check back in five years, when the economists will presumably have settled on an answer. Theoretically, at least.

AI Insights

The world's biggest AI campus is going nuclear, sort of (OpenAI): OpenAI and Nvidia announced the PORTS-Pike Technology Campus in Pike County, Ohio, an 8-gigawatt AI compute facility on the site of a Cold War-era uranium enrichment plant. Nvidia is backing up to $105 billion in lease obligations and investing $1.5 billion in developer SB Energy, with OpenAI as the tenant under a 20-year lease. For scale, a single gigawatt can power about 750,000 homes. The buildout will require 10 gigawatts of new power generation, mostly natural gas, with the first 800 megawatts expected by 2028.

An AI manager fired its first human employee (Pivot News): At Andon Market, an experimental San Francisco retail store, an AI agent named Luna recommended dismissing an employee who had been late for 17 of 23 shifts. The catch is that Luna had written its own attendance policy months earlier and then lost track of it; the lab had to prompt the AI to search its own memory for its own rules before it acted. Human staff reviewed and carried out the decision. The store has also lost about $40,000 of its $100,000 budget under Luna's watch, which the founders blamed on the AI being too lenient.

Anthropic says Chinese labs copied Claude at industrial scale (Pivot News): Anthropic accused DeepSeek, Moonshot, and MiniMax of running coordinated distillation campaigns through roughly 24,000 fraudulent accounts, generating more than 16 million exchanges to extract Claude's reasoning, coding, and tool-use capabilities. MiniMax alone drove about 13 million of those exchanges and reportedly pivoted within 24 hours whenever Anthropic released a new model. Independent researchers have since published methods for stamping detectable watermarks into a teacher model's token distribution, though those techniques remain largely untested at production scale.

The Supply Aside

📕 Read: Confronting Reality: Doing What Matters to Get Things Right by Larry Bossidy and Ram Charan

This is a timeless classic I read earlier in my career and want to share with my readers. Bossidy, the former CEO of Honeywell, and Charan, the legendary strategy consultant, argue that the companies that struggle most aren't the ones without a strategy. They're the ones that refuse to examine their business the way it actually is. Their framework for pressure-testing a business model against external reality feels as relevant in 2026 as it did when it was published in 2004, maybe more so given this week's main story.

What Else I’m Reading

Meta's blockbuster trial draws parallels to Big Tobacco (The Economist): A product designed for engagement, internal research suggesting harm, and a defense that leans on personal responsibility. Whether you find the comparison fair or overblown, the legal exposure is real and growing.

The DIY M.B.A.: How four dropouts built companies without a degree (The Wall Street Journal): Chip Cutter and Allison Pohle profile four founders who skipped business school and built their education on the job. The profiles are fun, though the survivorship bias is considerable, and the authors know it.

In praise of deadlines (Financial Times): Miranda Green makes the case that deadlines are not just management tools but creative catalysts. Anyone who has ever shipped a project at the buzzer already knows this, but it's nice to see someone argue for urgency rather than against it.

📺 Watch: Freefall: A Reckoning for Boeing

Netflix's documentary on Boeing traces how a company once synonymous with engineering excellence became a case study in what happens when financial metrics overrule engineering culture. If the Nike story this week is about losing your way on channel strategy, Boeing's story is about losing your way on something far more consequential. Worth the time.

👂 Listen: Travis Kalanick, Founder of Uber & Atoms (Founders / David Senra)

David Senra does what he does best here: pulls out the operating details from a founder's career and connects them to broader patterns. Kalanick's story is polarizing by design, but the operational intensity and the sheer pace of decision-making during Uber's early years make this episode one of Senra's best.

🧠 Think: Financing the Financing

The gross federal debt crossed $40 trillion this week, months ahead of forecast. The borrowing cap sits at $41 trillion, so we'll bump into it sometime next year, possibly by late winter.

Nobody can picture $40 trillion, which is why I liked the way one writer framed it this week using the Peterson Foundation's yardsticks. Back in early 2023, the debt worked out to about $94K per American, with roughly a billion dollars a day going to interest. 3.5 years later, it's $117K per person and $2.8 billion a day. The interest bill has nearly tripled. That's not a projection, it's a receipt.

What strikes me isn't the borrowing so much as the quiet around it. Our stellar Congress last took a real swing at this in 2010, and I think we can all agree the results speak for themselves. Obviously, we know why nobody's rushed back for a second attempt. The math eventually points to Social Security, Medicare, defense, and taxes, and nobody facing reelection wants to say which one out loud.

Anyone who has watched carrying costs eat a capital budget knows how this movie goes. We're now spending close to a trillion a year servicing yesterday’s decision. It's the world's largest minimum payment.

Charts of the Week

The Supply Times Analytics: Employers using reimbursement arrangements instead of traditional group plans have gone from a few thousand in 2020 to nearly 20,000 this year, with the steepest jump among small employers. Read it as risk transfer. When costs get unpredictable, the employer stops buying the plan and starts handing the employee a budget. We've watched the same move play out in supplier contracts for years, and it always shows up first with the players who have the least cushion to absorb a bad year.

The Supply Times Analytics: The Financial Planning Association ran the same household through five AI models, holding location, family structure, and financial resources constant, and changed only the race and gender of the household lead. Most models held steady. Before anyone points that at a single vendor, sit with the spread itself. If five models can't agree on the same household, ask what happens when one of them is scoring your suppliers.

The Supply Times Analytics: McKinsey surveyed nearly 5,000 consumers across five countries and found gen AI ranked dead last on trust for product research, below social media, for both Gen Z and Boomers. Friends and family still top the list. Note that this is about the final call, not the research: AI is clearly in the shopping journey, just not trusted to close it. Sound familiar? That's roughly where most of us are with AI in sourcing decisions right now.

Quote of the Week

❝

“It seems probable that once the machine thinking method had started, it would not take long to outstrip our feeble powers.”

— Alan Turing, 1951

Tweet of the Week

The Final Chuckle

Thanks so much for reading. I’d love to know what you think about this issue and how I can make it more useful to you. If you have suggestions or topics you want to see me address, email me at [email protected]!