If you’re a short-term trader maybe you can make a case for buying AI and chip stocks on the recent rout.
If you’ve got a time horizon that stretches for longer than the next few weeks or months, I think the dip you want to buy is the big one in gold and silver.
Chip stocks fell again on Friday, with Philadelphia Semiconductor Index (SOX) down 5.78%.
Nvidia (NVDA) shares rose in premarket trading but could not hold the gains and were down about 4% in the regular session.
Advanced Micro Devices (AMD) fell about 10%, while Intel (INTC) and Qualcomm (QCOM) each lost about 6%. Broadcom (AVGO) shed about 6.5%.
Memory and storage companies also took heavy losses, with Micron (MU) dropping about 5% and SanDisk (SNDK) falling about 8.5%. Western Digital (WDC), Seagate (STX), and South Korea’s SK Hynix (SKHY) all declined between 5% and 10%.
There is also speculation that Nvidia’s interim CEO might be less willing to back the large, risky circular financing deals that past CEO Jensen Huang promoted. Companies like CoreWeave (CRWV), which build and run AI infrastructure, were hit. CRWV lost 11%; IREN (IREN) and Applied Digital (APLD) both dropped 10%, while Cipher Digital (CIFR) shed 8.5%.
Maybe you can make a case that when Nvidia hires a new CEO, and when the company reports earnings on Tuesday, November 17, the tech sector will be primed for a bounce after so much selling.
To which I say:
The quarter that Nvidia will report on November 17 is the quarter that ended on October 25. Any good news the company reports will be backward looking.
I expect that when Nvia hires a new CEO, he or she will will want to clesr the decks after Huang. That’s likely to mean resetting growth projections lower to correct some of Huang’s more aggressive revenue accounting, and a recalibration of Huang’s policy on circular financing of customers’ chip purchases and funding for AI startup companies. In other words, there’s a good chancee for a technology version of a kitchen sink quarter that hits the stock while potentially setting the stage for gains in 2027 and 2028.
The widely anticipated Anthropic IPO will produce a lot of buzz in the AI sector, and it’s likely to rocket out of the gate. (Anthropic is scheduled to begin trading on Thursday, November 19 under the ticker ANT, subject to regulatory clearance and completion of the offering. Pricing is Wednesday evening, November 18. The proposed IPO price range is $70–$78 per share; the final offering price has not yet been set.) Preliminary reports say the offer is 6-times over subscribed so there will be a large number of investors and traders who wanted in, but who didn’t get in, and who will now buy. But there will also be a lot of folks who did get in, and some early round investors who will be able to sell because not all their shares are locked up, who will be looking to turn paper appreciation into real money. (Up to approximately 316 million shares are potentially eligible for resale from the first trading day—about 38% of the proposed 825-million-share offering.)
I think a likely pattern would be for Anthropic’s IPO to follow the path of the SpaceX (SPCX) IPO. SpaceX (SPCX) began falling after its June 16, 2026 post-IPO peak—its third trading session. At the November 13 closing price, the shares are down 6.86% from the $150 first public trading price.
The summary conceals a lot of volatility. On June 12, shares opened for public trading at $150, above the $135 IPO offering price. On June 16 the stock reached an intraday record of $225.64—approximately 50% above its first trade. Its highest closing price was $211.39 that day.
The loss from that record high to the November 13 close is about 38%.
So the most likely pattern in the short term is a decent pop from the offering price followed by a drop below the offering price as more shares are available for trading.
You want to trade that in the short run? If you can get in before the first trade, or at close to the opening price, I think you’ve got a goodchance to make some quick money. After that I think the odds are against you.
Which is about how I feel about buying on this technology dip. If you’re going to make money from the short-term bounce this week–and maybe next–get in early. Because the longer-term problems in the AI and technology sector haven’t gone away.
And the doubts haven’t been answered. And have actually become more acute since the election and with the prospect of more interest rate increases from the Federal Reserve.
Basically, the medium-run question boils down to this: Will AI companies generate significant profits before the bill for building out the AI infrastrucure comes due?
Again CoreWeave is a good example of the problem.
CoreWeave’s September filing identifies approximately $13.4 billion of principal repayments due between October 2026 and December 2027, before interest. The company’s June filing has a fuller list of principal payments due. July–December 2026 $4.413 billion. 2027 $6.184 billion. 2028 $4.416 billion. 2029 $2.421 billion. 2030 $3.221 billion. After that $14.896 billion. These are scheduled payments on debt outstanding at June 30, excluding interest and subsequent borrowing. (Much of CoreWeave’s equipment financing is amortizing debt: principal is repaid in installments as customer contracts generate cash. A facility with a final maturity in 2031 or 2032 can therefore require substantial payments in 2027. This differs from a conventional corporate bond whose principal is largely payable at maturity.) Interest is additional.
But CoreWeave isn’t alone. Oracle’s published debt schedule shows $7.21 billion of principal due in fiscal 2027, followed by $10.15 billion in fiscal 2028. (Its fiscal years end on May 31, so fiscal 2027 covers June 2026–May 2027.) And then 2029 $5.500 billion. 2030 $7.250 billion. 2031 $9.750 billion. After May 31, 2031 $90.250 billion. (This is principal only: it excludes interest and lease payments.)
Needless to say, but I’ll say it anyway, anything that delays the completion of a revenue-producing data center–like let’s say a state moratorium on new permits–makes the cash flow to meet these obligations more uncertain. (And I’m not sure that’s the most dangerous outcome. Many AI compute contracts offer customers the opportunity to walk away from paying. Think an OpenAI or Anthropic or Microsoft will honor promises to pay Oracle or SpaceX billions if a data center is late going online?)
Now contrast this picture with dip in gold and silver, and the medium- to long-term prospects for gold and silver.
The extent of the dip is of the same dimension as the technology dip. Both gold and silver reached their 2026 intraday highs on January 29. At the November 13 close, gold had fallen 30.68%–certainly in the same league as the AI dip–and silver had plunged a shocking 59.82% from that peak.
I don’t think we can call a bottom here with absolute certainty. For U.S.-dollar spot prices going into the week of November 16, the immediate support zones are approximately $3,850–$3,870 for gold and $48.50–$49.00 for silver. Both metals ended Friday close to those areas, so the question is whether buyers can defend them–not whether a durable bottom has already formed. Gold is testing support now. A sustained break below approximately $3,850 would put $3,800, then the broader $3,780 area, in focus. A rebound that stalls below $4,000 could still be only a recovery within a downtrend. For silver, the question is whether $48.50 holds, followed by a recovery above $50 and then $52–$53.
With the Federal Reserve almost certain to raise interest rates by at least 25 basis points at its December 9 meeting–93.6% odds as of the CME FedWatch tool on Friday, November 13–and again at its January 27 meeting–with 76.3% odds of an additional increase, according to FedWatch–there is substantial near term downward pressure on the metals that could lead to a move below support.
You could argue that in the very short term–a few weeks to two months–buying the AI and technology dip is more attractive than buying the precious metals dip. I think that’s a supportable argument. Supportable, but I don’t agree. Global central bank buying provides reasonably strong support at current levels and is likely to balance out, roughly, the downwarrd pressure from higher interest rates at the Fed and elsewhere. For example, the People’s Bank of China has accelerated its reported gold buying: it added 25.8 metric tonnes in October 2026, its largest monthly addition in about three years, bringing purchases for January–October to approximately 129 tonnes. (October is the latest available reporting month; the figures were released November 7.)
Go out a little further, however, and the risk/reward balance shifts decisively in favor of buying the precious metals dip. Rather than a precarious hope that everything will go right in AI economics, the case for gold and silver rests on very solid demographic trends–the world’s population of expensive-to-care-for older folks is rising–and the demonstrated track record of self-interest and short-sightedness aming the world’s politicians. If you wanted to be mean but honest, you could say “veniality” and “stupidity.” The odds of losing money betting on those qualities are very, very low.
We are headed into a global government debt crisis that is extremely likely to result in the devaluation of most if not all of the world’s paper currencies. (You can add in the effects on government budgets and national economies from global climate change exacerbated by the unwillingness of those same politicians to spend money now to head off bigger future costs–but that sort of feels like piling on.)
I think a smart long-term strategy is to use dips in gold and silver to bet on the likelihood of that crisis. And we’ve got one of those dip-buying opportunities right now.
You can find my recimmendations for gold and silver ETFS and for gold and silver mining stocks in my online portfolios on JubakAM.COM (subscription of $249 a year required) or a somewhat shorter list on JubakPicks.COM (Free if you sigN up as a user by giving us your email address.)
