The AI Selloff Meets the Midterm Calendar
Oct 22, 2026 · Don Schreiber, Jr.
The AI selloff looks overdone, and the midterm-year calendar is turning in its favor. A durable rebound, though, likely needs relief on interest rates and AI financing, not just strong earnings.
Chip stocks just went through a historic slide. The Philadelphia Semiconductor Index fell for an eighth straight session through October 16, matching its longest losing streak since 2006, after a 12–13% drop that erased about $2 trillion in value. But the worry is about how AI data centers are being financed, not whether demand for AI chips and software is real.
Meanwhile, history is on the bulls' side: midterm-year fourth quarters have been the strongest of the four-year presidential cycle. What sets this year apart is the macro backdrop, with Treasury yields above 5%, firm inflation and a Fed that won't rule out a hike. October 28 is the pivot, when the Fed decision lands the same day Microsoft, Meta and ServiceNow report earnings.
The week in review
Technology led the market lower in the week ending October 16, while investors rotated into bonds and defensive sectors. On Friday the S&P 500 fell 1.1% to 7,659, breaking a 55-day streak without a 1% daily decline, the Nasdaq dropped 2.1% to 26,140 and the Dow slipped 0.3% to 52,001. For the week, the S&P 500 lost 2.0% and the Nasdaq 4.5%, their biggest weekly drops since June, while the Dow added 0.7% for a second straight weekly gain.
Inflation ran hot. September CPI rose 3.7% from a year earlier (0.6% for the month). Producer prices rose 1.3% for the month, the largest monthly gain since 2022, and 6.3% year over year.
The Fed kept a hike on the table. Chair Kevin Warsh declined to rule out a rate increase at the October 28 meeting and said he would follow the data.
Oil stayed elevated. The U.S.–Iran tanker conflict around the Strait of Hormuz kept crude climbing: WTI settled at $99.02 on Friday, its highest close since mid-September, and Brent settled at $109.86, its highest since May.
Money moved to safety. Treasuries had their best week since mid-February, with the 10-year yield falling 14.6 basis points to 5.10%, as investors rotated out of tech and into bonds and defensive sectors.
Software joined the selling. On Friday, Okta fell 8.9%, CrowdStrike 7.4% and Palo Alto Networks 7.1%, showing pressure on growth stocks beyond chips.
Market | Friday close (Oct 16) |
S&P 500 | 7,659 (−1.1%) |
Nasdaq Composite | 26,140 (−2.1%) |
Dow Jones Industrials | 52,001 (−0.3%) |
Russell 2000 | 2,721 (−1.7%) |
10-year Treasury yield | 5.10% |
WTI / Brent crude (Dec futures) | $99.02 / $109.86 per barrel |
Gold (Dec futures) | $4,181.10 per ounce |
Closing levels for Friday, October 16, 2026.
What's behind the AI selloff
The selling is about how the AI build-out is financed, not whether the demand is real. Credit markets are questioning the debt behind AI data centers, and that worry has spread to every stock tied to the AI trade.
Credit stress is the trigger. The cost of insuring CoreWeave's bonds against default hit a record, implying roughly 60% odds of default within five years.
Debt-heavy names are falling hardest. On Friday, CoreWeave fell 9.3%, IREN and Applied Digital 7.5% each, and Super Micro 7.4%.
Chip fundamentals have not broken. Strong results haven't protected the stocks: Micron fell to about $898 despite delivering strong results and guidance on September 30, and optical supplier Coherent has lost nearly 30% in the selloff despite 34% revenue growth last quarter.
The selling is crowded. Short sellers have made about $32 billion on chip stocks in seven sessions, and 20 of the chip index's 30 members have fallen more than 10% since the selloff began.
Perspective matters. Even after the drop, the chip index remains up roughly 57% this year.
Our read: falling prices while reported results hold up is the profile of a sentiment-driven correction. That is typically where rebounds start, once the specific worry eases.
The seasonal backdrop
We are entering what has historically been the strongest stretch of the four-year presidential cycle. Since 1937, the S&P 500 has averaged a 6.6% gain in the fourth quarter of midterm years, more than twice the 2.9% fourth-quarter average across all years.
The year after has been reliably positive. In every midterm cycle since 1946, the S&P 500 was higher one year after Election Day — 20 for 20 — with an average 12-month gain of about 15%.
The first three quarters were flat. In midterm years, Q1 through Q3 have averaged −0.5%, −0.6% and −0.1%, so gains have tended to bunch up at year end.
This overlaps with the usual year-end rally, which typically runs from mid-November into January. This year, AI earnings season runs through that same window.
Why this year could be different
Seasonality is a tailwind, not a guarantee. Four conditions set 2026 apart from most past midterm-year rallies.
Rates and inflation are working against stocks. A 10-year yield above 5%, CPI at 3.7% and PPI at 6.3% leave little room for the Fed to ease. Most strong midterm fourth quarters came with the Fed easing or on hold.
The Fed may hike on October 28. Chair Warsh has kept that option open. A hike would weigh most on the long-duration growth stocks that lead AI rallies.
Earnings beats don't fix credit problems. A strong Microsoft quarter does not repair a debt-funded data-center operator's balance sheet. Credit spreads, not earnings per share, are the gauge to watch.
Investors are favoring defensives. Staples, healthcare and utilities are leading, and Treasuries are drawing inflows. Momentum stocks can trail the broad index while that rotation lasts.
The midterm figures also rest on a small number of historical observations and track the S&P 500, not AI technology stocks specifically.
The AI earnings calendar ahead
AI earnings season runs from October 28 to early December, overlapping the strongest part of the seasonal calendar. October 28 is the pivot: the Fed decision and three AI bellwethers in one day.
Estimated date | Event / reporting companies |
Oct 28 | Fed decision; Microsoft, Meta, ServiceNow |
Oct 29 | Atlassian, Cloudflare, Twilio |
Nov 2–9 | Palantir, Arista, RingCentral, Qualys, Freshworks, Tempus AI, AMD, Super Micro, Paycom, Dynatrace, Fastly, Datadog, Five9 |
Nov 3 | Election Day |
Nov 16 into Jan | Typical year-end rally window |
Nov 24–25 | Workday, Dell, Veeva |
Dec 1–3 | CrowdStrike, GitLab, Zscaler, Okta, Rubrik, HPE |
Dec 8 | SentinelOne, SailPoint |
No company has confirmed its date yet. These estimates are based on last year's schedule and will firm up through October and November. Veeva, SailPoint, Five9 and Zscaler carry the widest uncertainty.
How we're approaching it
Our AI-focused strategies emphasize businesses that monetize AI, such as software and security companies, rather than the borrowers building its infrastructure. Those cash-generative businesses have little exposure to the debt-funded data-center build-out at the center of this selloff, and many report after the October 28 catalyst, when the Fed decision and hyperscaler spending plans will be known.
Risk management stays rules-based. WBI strategies are managed by the Power Factor Engine's disciplined ranking and stop process, and our Trend Switch models moved into a defensive bond position in late September. We don't change positioning on seasonal patterns or headlines; the process responds as prices and fundamentals change.
Our base case: if the Fed holds and the hyperscalers reaffirm their spending, quality AI software is well placed to lead a rebound into year end. If the Fed hikes or credit stress spreads, we would expect defensives to keep leading through November, and a year-end rally to start later and be narrower.
What we're watching
These five signals, in order, will tell us whether the seasonal rally takes hold.
The October 28 Fed decision. A hold with balanced language clears the way. A hike, or a clear signal of one, delays the rebound.
Hyperscaler spending plans. Capital-spending guidance from Microsoft and Meta on October 28, and from Alphabet and Amazon around the same time, will show whether the AI build-out is intact.
AI credit spreads. The cost of insuring CoreWeave's debt is the clearest gauge of the financing worry. A peak and reversal would be an early all-clear.
The 10-year Treasury yield. A sustained move back below 5% would ease pressure on growth stocks.
Breadth in chip stocks. We want to see many chip stocks rising off their lows, not just Nvidia.
Sources
Benzinga: Nasdaq 100 slides as Micron, SanDisk, CoreWeave sink
Stocktwits: Chip stocks on track for 8-day slide erasing $2 trillion
Stocktwits (citing Bloomberg and S3 Partners): Short sellers have earned $32B amid chipmaker selloff
The Next Web: Huang says CoreWeave needs no backstop, as its credit swaps hit a record
Reuters via BusinessDay: AI credit fears push tech default insurance to record highs
Longbridge: Chip stock selloff hands short sellers $32 billion in paper profits
First Trust: September CPI and September PPI
investingLive: Warsh won't prejudge rest of 2026
OilPrice: HSBC raises 2026 Brent forecast to $90 as Hormuz crisis drags on
Barchart via Webull (citing J.P. Morgan and Bespoke): S&P 500: Q4 strength or election risk?
The Wealth Advisor (citing Carson Group): Midterm years have produced the same stock-market pattern since 1946
24/7 Wall St. (citing J.P. Morgan Asset Management): Midterm-year quarterly S&P 500 averages
Investrade / Hammerstone Markets: Market Review: October 16, 2026 (closing index, yield, oil and gold levels)
Earnings dates: company investor-relations history and aggregator estimates (MarketBeat, TipRanks), compiled September 28, 2026
Important disclosures
This commentary is provided by WBI Investments, LLC for informational purposes only. It is not a recommendation to buy or sell any security or an offer of advisory services. Opinions reflect our judgment as of the date shown and may change without notice. Market data are approximate, as of October 16, 2026, and taken from third-party sources believed reliable but not guaranteed. Historical and seasonal patterns may not repeat. Past performance does not guarantee future results, and all investing involves risk, including possible loss of principal. Companies named are for illustration only and should not be considered recommendations.
