Wall Street is getting nervous about the bill. After earnings reports from Alphabet and Tesla—the first of the so-called "Magnificent Seven" megacap companies to report this season—sent U.S. stock futures tumbling, investors are openly questioning whether the trillions being poured into artificial intelligence infrastructure will ever pay off.
Dow futures fell 142 points, or 0.27%, while the S&P 500 futures dropped 20.25 points, also 0.27%, and Nasdaq-100 futures sank 81.5 points, or 0.28%. When the market opened, the declines accelerated. The Dow Jones Industrial Average fell 463.0 points, or 0.89%, to 51,755.54. The S&P 500 dropped 80.7 points, or 1.08%, to 7,418.29, while the Nasdaq Composite fell 445.4 points, or 1.73%, to 25,245.542.
The anxiety reflects a growing gap between who's winning and who's losing in the AI investment cycle. Charu Chanana, chief investment strategist at Saxo in Singapore, crystallized the divide: "U.S. megacaps may face more scrutiny because they are writing the cheques, while chipmakers, memory suppliers and infrastructure companies get paid earlier in the investment cycle."
That dynamic played out across global markets on Thursday. Asian markets surged as oil prices climbed toward $100 a barrel, buoyed by demand from the infrastructure buildout. In Seoul, the KOSPI jumped more than 4%, with SK Hynix gaining 4.8% and Samsung climbing 3.7%. Tokyo's Nikkei and Hong Kong's Hang Seng also ticked higher. These chipmakers and semiconductor suppliers are positioned earlier in the spending chain, capturing revenue before the companies actually deploying the AI systems see returns.
The Spending Gamble
The scale of AI investment is staggering. Advanced Micro Devices announced plans to sell up to two gigawatts of its Instinct MI450 chips to AI lab Anthropic beginning in the first half of 2027, a deal that also includes an investment of as much as $5 billion in the Claude maker. Nine months ago, AMD announced a multiyear deal with OpenAI that would bring in tens of billions of dollars in annual revenue while giving the ChatGPT creator the option to buy up to roughly 10% of the chipmaker.
On Thursday, AMD was set to launch a raft of AI hardware designed to rival Nvidia at an event in San Francisco. The company is attempting to capture market share from Nvidia in the fast-growing data center chip sector, especially for so-called inference computing—the data crunching that occurs when a user queries a chatbot such as OpenAI's ChatGPT. AMD was expected to show off its data center hardware, including its first-generation server racks called Helios, and formally launch its Venice central processing unit for data centers.
Nvidia, meanwhile, released technical details about its Vera CPU, which it said, when combined with its Rubin graphics processing unit, will do the best job at maximizing how much work AI agents can do with a given amount of electricity. At the Moscone West convention center on Wednesday, hundreds of executives and engineers gathered to take in technical presentations and mingle on a showroom floor. AMD displayed the Helios data center rack amid booths from cloud computing providers such as Vultr and TensorWave, both of which operate data centers with AMD hardware.
Who Bears the Risk
But the question haunting markets is whether the companies writing the massive checks—Alphabet, Tesla, and other megacaps—will ever recoup their investments. These firms are committing enormous capital to infrastructure that's still unproven in terms of commercial returns. Meanwhile, the chipmakers and suppliers get paid upfront, shifting the financial risk onto the tech giants and their shareholders.
In Europe, the picture was more complex. Government borrowing costs moved to long-term highs as the European Central Bank held rates steady at 2.25% after lifting them in June. The ECB warned that "the full inflationary impact of the energy shock has yet to play out," noting that "uncertainty remains high." Germany's 10-year bund yield climbed above 3.2% for the first time since 2011. Markets now see a four-in-five chance of another interest rate hike in September if energy prices remain elevated. Morgan Stanley analysts said the ECB's "messaging remained unchanged and open," and maintained their call for another rate hike in September.
The energy shock stems partly from Middle East tensions. The Iran-aligned Houthis said on Thursday they had struck two Saudi oil tankers as part of a naval blockade on Saudi Arabia, threatening to create a second chokepoint on global oil supplies alongside Iran's near-closure of the Strait of Hormuz. The U.S. military carried out a new round of strikes on Iran, marking a 12th successive night of American attacks and prompting further Iranian retaliation. Brent crude jumped almost 5% to more than $98 a barrel.
In Europe, semiconductor stocks diverged as investors weighed AI demand against the reality of higher borrowing costs and energy prices. The divergence reflected broader uncertainty: some chipmakers benefit from AI infrastructure demand, while others face headwinds from economic slowdown and inflation.
Why This Matters:
The AI investment cycle is creating a two-tier market where risk is distributed unevenly. Chipmakers and infrastructure providers capture revenue early and with certainty, while the companies betting billions on AI deployment face uncertain returns and mounting pressure to justify their spending. This structural imbalance matters because it concentrates financial risk on a handful of megacap tech firms—and ultimately on their shareholders and employees—while benefits flow to a narrower set of suppliers. Meanwhile, energy shocks linked to geopolitical conflict are pushing up borrowing costs globally, making it more expensive for governments and businesses to invest in other priorities like infrastructure, education, and climate transition. The question isn't just whether AI will be profitable—it's whether the current investment model is sustainable, and who pays if it isn't.