The AI trade has started to wobble, and the cracks are not limited to fringe players. Advanced Micro Devices is walking into its next earnings report against that backdrop.
AMD is scheduled to report fiscal third-quarter 2026 results on November 3. Analysts expect revenue of around $13.08 billion. That would mean a chipmaker still growing quickly while parts of the AI ecosystem show signs of strain.
The timing is awkward. Anthropic has filed to go public, credit markets tied to data center buildouts are flashing warning signs, and some infrastructure stocks are sliding.
What AMD needs to show
The setup comes from a strong second quarter. AMD posted Q2 2026 revenue of $11.54 billion, a 50% increase from a year earlier.
Net income for the quarter came in at $2.3 billion.
The engine was the Data Center segment. Revenue there more than doubled year over year to $6.7 billion, which accounts for more than half of the company’s total sales.
Two product lines did the heavy lifting. EPYC server processors and Instinct AI graphics processing units drove the bulk of that growth.
The roughly $13.08 billion consensus for Q3 sits more than $1.5 billion above the Q2 figure.
CNBC identified AMD as a strong growth stock heading into the print. It also recommended an options strategy designed to limit downside risk ahead of the report.
The cracks in the AI trade
The stress is showing up in places that usually move slowly, starting with the bond market.
Yields on one AI infrastructure borrower’s 2032 notes touched 15%. That is a level typically associated with borrowers investors consider genuinely risky.
Credit default swap spreads for the same name nearly doubled in October. A credit default swap works like an insurance policy against a borrower failing to pay its debts, so a wider spread means the insurance got more expensive.
Equities are feeling it too. IREN, an AI infrastructure operator, is down more than 60% over the past year, a reminder that riding the AI theme does not guarantee a smooth ride.
Then there is Anthropic. The AI developer filed its S-1 registration for a Nasdaq IPO on October 26, 2026, stepping into public markets just as financing conditions for AI and data center projects turn choppier. Anthropic reported first-half 2026 revenue of $16.3 billion alongside a net loss of $80.7 billion.
Why the divergence matters
The interesting split here is between companies that sell AI hardware and companies that borrow heavily to build AI capacity. AMD sits on the selling side.
Chipmakers like AMD collect revenue when customers place orders. Data center developers, by contrast, often take on significant debt upfront and bet that demand will be there to fill the buildings later.
That difference becomes crucial when capital gets more expensive. Rising borrowing costs squeeze the builders first, because their business models depend on cheap financing to fund long construction timelines.
For AMD, the risk is indirect but real. If financing pressure forces data center operators to slow expansion, fewer servers get built, and fewer EPYC and Instinct chips get ordered.
So far, AMD’s numbers do not show that slowdown. Data Center revenue doubling in a single year suggests customers were still spending aggressively as of Q2.
What to watch on November 3
The headline revenue number will grab attention, but investors will likely care more about the Data Center line. Whether it keeps climbing at anything close to its recent pace will say a lot about real AI demand.
The Instinct GPU business deserves particular scrutiny. It represents AMD’s most direct bet on AI accelerators, a market where buyers have plenty of incentive to diversify suppliers.
Anthropic’s IPO process adds another variable. Pricing, demand, and early trading will offer a real-time read on how much risk investors are still willing to take on AI.
That is likely why CNBC paired its bullish view with a risk-limiting options approach. Confidence in the company’s growth, combined with caution about the market around it, may be the most honest stance available heading into November 3.
Disclosure: This article was edited by Diego Almada Lopez. For more information on how we create and review content, see our Editorial Policy.