Everything About AI
NextFin News - Broadcom is assembling more than $60 billion in debt to finance AI chips for Anthropic and other customers, while Nike cuts jobs and warns that China will get worse before it gets better. Two of the market's most watched names moved in opposite directions this week, and the split lays bare the central tension of this market: capital is still cheap for the AI infrastructure buildout, but it is unforgiving for legacy brands still paying for years of underinvestment.
Two Movers, One Market Divide
Nike shares fell as much as 9.7 percent in New York on Friday after the sportswear giant reported fiscal first-quarter results that beat on earnings but missed on revenue and issued full-year guidance pointing to a high-single-digit revenue decline — steeper than the roughly 2 percent drop analysts had expected. The company also unveiled "Pace," an operating-model transformation that will cut jobs beginning in 2027, reorganize its geographies, modernize its supply chain, and establish a new corporate campus in Bengaluru, India.
Broadcom, by contrast, gathered a Wall Street syndicate to raise more than $60 billion of fresh financing — a $42 billion senior-secured tranche plus an $18 billion junior tranche led by Blackstone — to fund custom AI chips it will lease to Anthropic and other companies. The structure builds on a June partnership with Apollo and Blackstone that raised $35 billion, and it is part of a broader plan to finance more than 20 gigawatts of AI compute capacity by 2028, roughly the output of 20 nuclear plants.
The juxtaposition is not incidental. One company is shrinking its cost base to survive a demand slump; the other is helping customers borrow tens of billions to keep buying its chips. The question underneath both moves is the same: which of these stories is cyclical, and which is structural?
Nike's Turnaround Hits the Bill for Past Mistakes
Nike's first-quarter numbers were a study in contradiction. Revenue fell 4 percent on a reported basis and 5 percent on a currency-neutral basis to $11.2 billion for the quarter ended August 31, 2026, missing the $11.32 billion analysts had expected. Yet diluted earnings per share came in at 48 cents, ahead of the 43-cent consensus, and gross margin expanded 60 basis points to 42.8 percent, helped by lower warehousing and logistics costs. Selling and administrative expense declined 3 percent to $3.9 billion, and the company returned approximately $610 million to shareholders through dividends, up 3 percent from the prior year.
The beat, however, did not matter. Investors focused on what management said next. Chief financial officer Dave Denton told analysts the company's guidance range "assumes that China actually gets worse from a revenue perspective for the balance of this year." Chief executive Elliott Hill added that turning around the company's troubled regions "will take time," after deliberately cutting Dunk production by about half — a decision Hill said created roughly a $200 million headwind in the Sportswear division during the quarter.
"Our Nike performance business is not yet large enough to offset the pressure we're seeing in Nike Sportswear, Jordan Brand, and Greater China," Hill said on the earnings call.
The geography breakdown explains the urgency. Greater China revenue fell 26 percent on a currency-neutral basis to $1.18 billion — the ninth consecutive quarter of year-over-year declines and the steepest yet. EMEA fell 5 percent, APLA was flat, and Converse dropped 28 percent to $263 million. Only North America grew, up 2 percent to $5.13 billion, helped by momentum in running, football, and basketball. Nike Sportswear declined in the low double digits and Jordan Brand fell in the mid-teens. NIKE Brand wholesale revenue fell 1 percent to $6.8 billion, while NIKE Direct fell 8 percent, dragged by a 13 percent decline in digital sales.
The company's response is Pace, a multi-year program expected to deliver approximately $2.5 billion in cumulative savings through fiscal 2031. That sounds substantial until you look at the timing: the savings are back-loaded, with the bulk not ramping until fiscal 2029 and 2030, while the costs arrive now. Nike expects roughly $1.0 billion of pre-tax charges through fiscal 2031, primarily employee-related, plus approximately $0.3 billion of severance already recognized in fiscal 2026 and another $0.3 billion expected in fiscal 2027. The program will also trim fiscal 2027 adjusted earnings per share by about 15 cents, with guidance set at $1.15 to $1.35 — well below the roughly $1.68 analysts had modeled.
The job cuts themselves — reported at around 1,400 roles, mostly in technology — are the third round of reductions this year, following 775 cuts at U.S. distribution centers in January. Management said decisions on affected roles will begin in 2027, and CEO Hill acknowledged in a memo that "news like this creates uncertainty." Notably, Nike declined to confirm a specific headcount, calling media reports speculative because the company does not yet know the number of roles or locations.
There is a deeper structural problem underneath the cyclical demand weakness. Nike is actively revoking online selling rights from several major retail partners in China beginning in January, steering customers toward Nike-controlled digital channels and branded storefronts on Tmall, JD.com, and Douyin. The goal is to fix inconsistent pricing and restore brand equity. The cost is near-term revenue and profitability that management says will feel the strain across several selling seasons. In other words, Nike is choosing to shrink before it can grow — a painful but arguably necessary reset of a distribution model that had become promotional and brand-diluting.
This is the crux of the cyclical-versus-structural call on Nike. The demand weakness in Sportswear and Jordan is cyclical: product cycles turn, silhouettes come back into favor, and consumer sentiment in the lifestyle category recovers. But the distribution reset in China is structural — once you revoke the online rights of major partners and consolidate onto owned channels, you do not go back. The old wholesale-and-marketplace model that fueled Nike's growth in the 2010s is being deliberately dismantled. That means the next few years are not a simple "wait for the cycle to turn" story; they are a transition period where revenue contracts before the new model can scale. The Pace savings, arriving mostly in fiscal 2029 and 2030, are explicitly designed to bridge that gap — cost cuts today funding a business model that pays off later.
Broadcom's $60 Billion Bet: Who Really Carries the Risk?
While Nike retrenches, Broadcom is doing the opposite: helping its customers spend. The chip designer is working with banks to syndicate more than $60 billion in debt, split into a $42 billion Class A senior-secured tranche and an $18 billion Class B junior tranche. Blackstone is leading the junior piece, committing $9 billion of its own capital and planning to syndicate the remainder. A representative for Broadcom declined to comment.
The financing is anchored by Anthropic, which is expected to become Broadcom's largest compute customer in 2027. The deal extends an April agreement in which Anthropic, Broadcom, and Google committed to roughly 3.5 gigawatts of computing capacity built on Google's TPUs, with that capacity coming online starting in 2027. It also builds on the AI XPV partnership Broadcom struck with Apollo and Blackstone in June, whose opening deal raised $35 billion. The partnership's stated ambition is to finance more than 20 gigawatts of computing power for AI labs by 2028.
The structure matters as much as the size. Rather than Broadcom taking the capex onto its own balance sheet, the debt sits in a special-purpose vehicle that buys Broadcom's chips and leases them back to the AI labs. This keeps Broadcom's balance sheet clean while locking in long-term customers — a financial-engineering solution to the central bottleneck of the AI boom, which is no longer chip supply but the financing required to deploy it.
The market read the news as validation. Broadcom shares ticked up 0.6 percent in pre-market trading on Thursday after an initial report of a $42 billion lending arrangement, and the stock has been bid higher on the broader financing news. With a market capitalization around $1.66 trillion, Broadcom is now one of the most valuable companies in the world, and management has guided for AI revenue to quadruple from 2026 levels and non-GAAP earnings per share to nearly triple to $30 by 2028.
The scale of the ambition is difficult to overstate. Anthropic's revenue run rate reportedly exceeded $65 billion by the end of July 2026, up from about $9 billion at the end of 2025 — a more than sevenfold increase in under two years. The company confidentially filed for an initial public offering in June 2026 and has projected 2028 revenue of $190 billion to $200 billion. If those numbers hold, the $60 billion financing is not aggressive; it is conservative relative to the revenue base that is supposed to service it. The bet is not really on Broadcom's creditworthiness. It is on Anthropic's ability to keep growing fast enough to make the lease payments.
The Second-Order Question: Is AI Capex Just Moving Off Balance Sheet?
The conventional read of the Broadcom deal is bullish: demand for custom AI chips is so strong that customers need dedicated financing vehicles to afford them, and Broadcom is positioning itself as the tollbooth on that traffic. That is true as far as it goes. But the second-order question is whether this financing structure is solving a real capacity problem or simply moving credit risk from the chipmakers' balance sheets into the broader debt market — where it is less visible and potentially more fragile.
Consider the mechanics. A special-purpose vehicle borrows $60 billion, buys chips from Broadcom, and leases them to Anthropic. The debt is serviced by Anthropic's lease payments. If Anthropic's revenue trajectory holds and its planned public listing proceeds smoothly, the structure works. But the junior tranche — $18 billion, led by Blackstone — absorbs the first losses, and the pricing of that piece will tell you how much risk lenders actually think they are taking. There is precedent for caution: in May 2026, an $18 billion financing tied to OpenAI's custom-chip project with Broadcom hit a snag when lenders struggled to reliably value a private company, and the deal was scaled back. The lesson is that the market is not writing blank checks, even in this cycle.
This is where the cyclical-versus-structural call becomes critical. The demand for AI compute is structural: model training and inference needs are growing faster than any other technology cycle in recent memory, and custom silicon delivers efficiency advantages that off-the-shelf GPUs cannot match. Broadcom's position as a designer of application-specific integrated circuits for hyperscalers is a structural advantage, not a cyclical one. But the financing of that demand has a cyclical component — it depends on credit markets staying open, on AI labs continuing to raise capital, and on revenue from deployed models actually materializing. If any of those links breaks, the lease payments falter and the SPV structure gets stress-tested in real time.
There is also a third-order channel that equity investors in semiconductor names have not fully priced in. If AI infrastructure increasingly moves into debt-financed special-purpose vehicles, the chip cycle becomes less sensitive to the chipmakers' own balance sheets and more sensitive to credit spreads. A widening in high-yield or private-credit spreads could slow AI deployment even if chip demand remains strong. In that scenario, Broadcom's order book stays full on paper while the financing that converts orders into deployed capacity dries up. The semiconductor cycle, in other words, becomes a credit cycle in disguise.
The counter-thesis is straightforward and deserves its due: this is exactly how infrastructure gets built. Data centers, cell towers, and aircraft have all been financed through asset-backed structures that separate the operator from the owner. AI compute is simply the next asset class to be securitized, and Broadcom is wisely refusing to carry the debt on its own books. Under this view, the $60 billion is not a red flag but a sign of market maturity — the moment when a speculative boom develops the plumbing of a durable industry. The presence of Blackstone and Apollo, two of the largest alternative-asset managers in the world, as financing partners is itself a signal that institutional capital has done the diligence and is willing to underwrite the risk.
That argument is strong, but it depends on one assumption: that the underlying tenants can pay. Aircraft and cell towers generate cash flows from diversified customer bases with long contracts and essential-service demand. AI labs are concentrated, largely pre-profitability, and dependent on a technology cycle that could shift faster than a multi-year lease. The falsifying signal to watch is the pricing of the junior tranche and the pace of Anthropic's revenue growth into its public listing. If the junior piece prices above roughly 7 percent, or if Anthropic's revenue trajectory slips ahead of its planned IPO, the market is quietly telling you the risk is being underwritten at too low a price. A third signal: if Broadcom's guidance for AI revenue quadrupling from 2026 is walked back in any quarterly update, the financing narrative loses its anchor.
What Comes Next: Two Timelines, Two Trades
The near-term path for Nike is clear and painful. Management has signaled that China gets worse before it gets better, and the Pace savings do not meaningfully ramp until fiscal 2029 and 2030. The stock, down more than 40 percent this year and trading near $35 against a 52-week high of nearly $75, is pricing in a multi-year turnaround. The key signal to watch is whether the Greater China decline narrows in the second quarter — a sequential improvement would suggest the distribution cleanup is working faster than management expects. If instead the decline stays at or beyond 26 percent and Sportswear continues falling in the double digits, the market will keep marking the stock lower. A narrower China decline, combined with stabilization in Sportswear, would be the first real evidence that the structural reset is landing.
Broadcom's timeline runs the other direction. The financing is still being ironed out, and the deal's final terms will set the tone. If the syndication completes smoothly and Anthropic's customer anchor holds, Broadcom's position as the AI infrastructure banker strengthens — and the stock's premium valuation gets another fundamental leg. The risk is concentration: Anthropic as the largest customer is a bet on one company's execution, and any stumble in the AI spending cycle would hit Broadcom harder than a diversified supplier. The base case is that the deal closes and the stock grinds higher on continued AI revenue visibility. The downside case is a credit-spread widening that delays deployment and forces a repricing of the entire AI capex complex. The upside case is that the $60 billion syndication is oversubscribed, signaling that institutional capital is still chasing AI exposure at any price.
For investors watching these two names, the lesson is not simply "AI up, retail down." It is that this market is rewarding companies that can externalize the cost of their growth while punishing those still internalizing the cost of their past. Nike is paying today for a distribution and product strategy that drifted over the last cycle. Broadcom is being paid today for a future that its customers must borrow to build. Both stories are defensible. Only one of them requires you to believe that cheap debt will last long enough for the bet to pay off.
The Bottom Line
Nike's job cuts are a cyclical symptom of a structural reset: the company must shrink its cost base and clean up its distribution before the "Sport Offense" can scale, and the next several quarters will likely disappoint. Broadcom's $60 billion financing is a structural shift in how AI capex gets funded — but it moves the credit risk into the shadows, where it is harder to see and harder to price. The market is telling you it trusts the chipmakers more than the sportswear giant. Just remember: the chipmakers' customers are the ones signing the loans.
Explore more exclusive insights at nextfin.ai.
