Week ended Oct 16, 2026: tech shares hammered, Treasuries rally

• Tech and AI sell-off dragged down NASDAQ and semiconductor indices.

• Flight to safety boosted "old school" equities and US Treasuries.

• Hotter US inflation data keeps pressure on December Fed decision.

• CoreWeave liquidity fears sparked broader credit contagion worries.

• Portfolio protection maintained by trimming blue chips and holding QQQ puts.

The drivers of last week’s performance

Old school investors “won” last week hands down.  The “trendy” stock markets and indices that are concentrated in tech shares, especially semiconductor shares (e.g. the NASDAQ and Asian exchanges like the KOSPI), took it on the chin, as investors finally got skittish on the broader A.I. trade.  The winners last week were what I view as the “boring places to have your money”, including US Treasuries, European stocks, and indices like the DJIA that are not heavily concentrated in emerging technology names.

I am not that surprised that A.I.-related shares have come under pressure, although this has largely been confined so far to the semiconductor sector.  The Mag 7 giants performed fairly well on the other hand, at least relatively speaking.  This reflects their “moats”, including attributes like diverse revenue streams, relatively low leverage, global presence/dominant market shares, juicy operating margins, and reasonably fair (relatively speaking) valuations aside from TSLA.  Of course, given the massive capex in the pipeline for data centre buildout by these tech giants, a fragile value chain involving data centres is also a concern for the hyperscalers.  However, as the story unfolded as might have been expected last week, the pain started with semiconductor company stocks and those companies travelling in that orbit, especially the ones losing money and in need of huge amounts of financing before they reach profitability. 

The poster child of the A.I. unwind last week was CoreWeave (CRWV), a cloud infrastructure provider that has been financed largely with reasonably-easy-to-obtain debt financing in the capital markets, as well as from banks and private debt funds.  I have provided a more detailed discussion for CRWV at the end of this update, in that it might not be interesting to all of my readers.  This situation is worth keeping an eye on because of the potential collateral damage into the markets more broadly.

The other things that mattered last week included inflation data in the US (CPI and PPI), Fed Chairman Warsh’s comments at the IMF meeting in Bangkok (the last such commentary from a Fed governor before the next FOMC meeting), and S&P 500 earnings.

• US inflation data:  CPI for September, released Wednesday, came in slightly higher than expected, illustrating the ongoing price pressure in the US.  Headline CPI was 3.7% YoY, and core CPI was 2.6% YoY.   PPI released the following day was also higher than expected, with headline at 6.3% YoY, and core PPI at 5.0% YoY.  Yields in the US moved sideways on the news, surprising to me given how elevated inflation is, and the fact that PPI is normally a harbinger to future changes in CPI.  The price action makes me believe that many investors view the market as oversold, and this might have contributed to yields falling rather than rising. However, I know from speaking to family and friends in the US that prices feel completely out of control Stateside.

• Fed Chairman Warsh at the IMF meeting in Bangkok (overnight Thurs):  Mr Warsh stuck squarely to his mantra of providing not a whiff of forward guidance.  Nonetheless, investors seemed to view his commentary as suggesting that the Fed would not increase its policy rate at the FOMC meeting at the end of the month.  Bond investors breathed  a sigh of relief as bonds rallied into this, although I remain somewhat perplexed that in light of the inflation data releases earlier in the week, investors have become more passive.  As far as I am concerned, this is going to make the final FOMC meeting of the year in December a “must act” decision – the Fed Funds rate will be increased then, assuming as expected that the FOMC does not tighten policy at the October meeting.

• Earnings:  The major US banks released earnings last week, a mixed bag but generally OK in my opinion.  JNJ (beat), ASML (beat) and TSM (beat) also released earnings, but the general downdraft seemed to pull most stocks lower last week, especially anything and everything semi-conductor related.  This occurred in spite of earnings beats and favourable guidance from the likes of ASML and TSM.

The table below summarises market performance for last week and YtD.  As mentioned already, “old school” equities (DJIA, European indices) and Treasuries were the place to be last week, while anything sniffing of tech concentration (especially semiconducters) got hammered.  The other beneficiary of this tilt last week was a flight-to-quality trade into US Treasuries, which pushed prices up and yields down across the curve.  Corporate bonds took it on the chin as concerns over A.I. related financing in both the investment grade and high yield markets caused corporate spreads to widen.  Based on years of experience, stress often first appears in the corporate credit market before engulfing equities.  Let’s hope this is a temporary blip in spreads, as the U.S. economy remains on solid economic footing for the time being.   

There are more detailled tables at the end of this update in the section "Market Tables." 

What matters ahead

• Central bank meetings: The next FOMC meeting takes place October 27–28, with a policy decision to be announced on October 28. The ECB (October 28–29), the Bank of Japan (October 29–30), and the Bank of England (November 5) closely follow the FOMC.

• Earnings: 75 or so S&P 500 companies release earnings this coming week, including GM, GE, SPCX and NFLX on Tuesday; and TSLA, META, NEE, TXN, IBM and T on Wednesday.   As far as other Mag 7 companies, AMZN, GOOG and MSFT release earnings the following week, and AAPL releases earnings on Nov 2.

• Economic data: A slew of economic data for China will be released on Monday (3QGDP, retail sales and industrial production data);  global flash PMI data for the US, Eurozone, and the UK will be released on Friday, providing a perspective on the manufacturing and services momentum in each country; and CPI will be released for the UK (Wednesday) and Japan (late week) ahead of each country’s respective upcoming central bank meetings.

  
  

With economic data rather scant in the US next week, there will be increased attention on corporate earnings, as well as the broader reaction to A.I. weakness this past week.  

What I did last week

I sold some blue chip names around the edges into strength this past week, wanting to take a bit off the table at these levels.  I unwound some QQQ (NASDAQ index) Dec puts (well out of the money), but still hold March and June QQQ puts as insurance at more attractive strike prices.  I also unwound some Nov 235 puts on CRWD (Crowdstrike) on Wednesday (at a loss), which were being affected by time decay.  The stock was moving in my direction (lower), but I wasn’t sure it would decline enough in the coming weeks to provide value to the position.  In retrospect, I pulled the trigger too early on this series, as the shares reached an intraweek low on Friday of around $225/share, down massively from its October 6 intraday high of nearly $287/share.  However, I used the weakness late Friday to unload some Nov 255 puts for CRWD that I was also holding, a very profitable trade. Lastly, I wound down my near-expiry TSLA puts Thursday (at a loss).  I did not reload on the short side of either CRWD (still massively over-valued) or TSLA (puts CRWD to shame), but must admit that I was itching to reload on Elon Musk’s flying cars rhetoric late in the week.  Mr Musk is an exceptional entrepreneur, but he also has an exceptional record of rarely delivering on his promises in a time frame he espouses. 

CoreWeave: in distress

CoreWeave (CRWV) has $7.75 billion of high yield bonds in four tranches due in 2030 (two) and 2031 (two).  These bonds got hammered last week on concerns about the company’s ongoing ability to tap the capital markets to fund its growth. CRWV loses money on the bottom line, but has been operating cash flow positive the last few quarters.  The company also has over $5 billion of cash on its balance sheet.  Having said this, CRWV has spent over $20 billion in capex in the last four quarters, and it will clearly need to continue to access debt capital to continue its growth since it is losing money hand over fist. 

The bonds traded last week in the high 70s / low 80s context, reflecting yields of around 14% to 15%, clearly in distressed territory.  The stock was down 25% last week, and closing Friday at $61.39/share, is less than half of its 52-week high of $141.79/share. 

The broader concern – and this is one that has been the overhang for the entire infrastructure for chips companies and data centre buildouts – is vendor financing.  It is the complex inter-relationships between companies like CoreWeave and its customer base, banks and the private debt community that raises more red flags, fanning fears of broader contagion. 

Nvidia (NVDA) has provided over $2 billion of direct investments into CRWV, and has also agreed to purchase over $6 billion of compute capacity through April 2032, essentially a backstop facility.  Many large banks (JPM, GS, MS and others) provide bank credit.  And several private equity firms, including Blackstone as the largest, provide well over $10 billion of financing via securitisations, direct loans and future draw-down facilities.  

Should this gravy train of “circular financing” – meaning Nvidia invests equity cash into CoreWeave → CoreWeave uses debt/equity to buy Nvidia chips → Nvidia books revenue and backs CoreWeave's compute capacity – be interrupted, the collateral damage into a broad array of industry players and financing sources would have dire effects on markets, and certainly would delay the build-out of data centres. 

The dark side of me thinks that this was coming and is in fact just what is needed to clear out the “hanger-ons” around the periphery that have been flying high but perhaps ought not to be.   It is a situation to watch carefully. 

My bet is that CRWV will be fine because it has such a diverse financing base.  For players with the risk tolerance, the bonds could represent real value at these beaten-down levels, but this depends on the priority of the bonds vis-à-vis other creditors involved in the company. If I were to get involved after doing much more work on the name, I would likely play the stock not the bonds, because the stock offers more upside. 

Market tables