By Dave Harden – CEO & CIO

State of the Markets

The past month

October saw a hotter inflation print and the “AI wall of worry” collide with a Federal Reserve still raising rates, and November opened with a political earthquake and the sharpest equity de-risking of the year.

SGI’s house view is that the fundamentals did not break: growth is robust, inflation remains elevated and sticky, the Fed hikes in December, and the post-election sell-off reflects positioning rather than a verdict on the economy. But….

• Growth is robust. Q3 real GDP grew 3.4% (consensus 3.2%), with consumer spending up 3.8% and business capex up 12.7% (equipment +20.9%). AI accounted for nearly 60% of the quarter’s capex gain.
• Inflation is sticky. September core CPI rose 0.33% (0.2% expected) on broad-based gains, and core PCE is running about 3.1% year over year. Diesel is up roughly 67% since the Iran war began. Trust me, I love/hate driving the truck!
• The Fed is raising rates. The Fed held in October, with two dissents in favor of a hike. We expect a hike in December and another in March; market odds of a December hike fell from ~90% to 68% after the election sell-off. Not sure that moves the needle enough to change our view. Higher rates are already hitting housing.
• The AI trade is evolving. Hyperscaler results disappointed, 2026 capex estimates stopped rising for the first time in 12 quarters, and financing stress surfaced among the most leveraged AI borrowers. In one survey, we noted, 73% of those who cut risk did so based on price action, not fundamentals. Regardless, more negative news and volatility have entered the AI trade.
• The midterms delivered a Blue Wave. Democrats are looking at ~55 Senate and ~255 House seats and won in Texas. But will federal policy change? Locally, more likely; nationally, less likely.
• Markets de-risked sharply. The S&P 500 traded in a 7,600-7,800 range for the month, then every sector fell on Nov 5: momentum had its worst day since 2020, semiconductors are 38% off their June high, and the S&P 500 is back to its 15-year average multiple (17.9x).
• Market corrections happen. If you needed your investments for Santa in 2026, that was not the best timing. If you are investing for next year and beyond, we believe the markets remain constructive. De-risking over days or weeks is a normal part of investing, and in most cases, when you work with us, we have planned for it.1

Growth & the economy: robust, and increasingly an AI-capex story

Third-quarter GDP (released Oct 29) grew 3.4% annualized vs. a 3.2% consensus, and the composition was the message: consumer spending +3.8%, business capex +12.7% (equipment +20.9% on computers and industrial machinery, structures +5.4%), and federal spending +15.8%. Real final sales to private domestic purchasers, the cleanest read on underlying demand, rose 4.9%. Housing (-3.4%) and a 3.3-point drag from net exports, as AI-related imports surged, offset the gains, partly offset by a 1.6-point inventory build.

The market has a vulnerability: the economy is now exposed to the AI buildout. We estimate AI-related investment at roughly 2% of GDP, accounting for about half of nominal capex growth and 20-30% of GDP growth recently; in Q3, it accounted for nearly 60% of the capex gain, and domestic AI capex growth accelerated to a 55% annualized pace. The risk is that a slowdown in the investment cycle would “suddenly” expose a later-cycle economy, meaning the market could turn over more than just the negative day on Thursdays (11/5). If AI capex growth slowed from ~40% to 20%, it’d likely subtract about 0.4 points from GDP growth. Add a 20% correction, and it would cost nearly 1 point; a more severe correction of down 25% could subtract roughly 2 points. That’s before any contagion to credit, employment, and confidence. To put this into perspective, if AI investment kept growing at 40% (which is the current best estimate by analysts), it could reach about 5.5% of GDP by 2030, comparable to the railroad boom, above the dot-com peak, but still below housing.

The softer edges are the interest-sensitive ones. Real personal spending was flat in September (core goods -0.3%, services slowing to +0.14%), the savings rate fell to 3.75%, and the retail-sales control group rose just 0.15%, below the 0.3% expected. In general, a 5% 10-year yield is not very restrictive for the aggregate economy, but the pain is concentrated among leveraged smaller companies and rate-sensitive consumer categories.

Inflation, the Fed & rates: a hot September, a December hike still the base case

Inflation. It appears to me that underlying inflation is firm but not re-accelerating. The print (Oct 14) was 0.33%, with the headline at 0.65%, and the problem was breadth: owners’ equivalent rent +0.35%, lodging +2.0%, airfares +1.2%, used vehicles +0.7%, recreation goods +0.9%. Producer prices followed, with supercore PPI at 0.49% (0.3% expected), and September core PCE landed at 0.32% month over month. Keep an eye on core CPI to see if it will step back down to 0.20%, as airfares fall (-0.5%, the first monthly decline since the Iran war began). This would still leave core PCE near 3.1% year over year, well above the Fed’s target.

The energy channel is diesel, not gasoline. Diesel is up about 67%, and gasoline 46%, since late February. The diesel crack spread (~$2.60/gal) is above its 2022 high, and freight costs feed into core goods. Trust me, I don’t need those specifics. When I filled up the other day, it was over the credit-card gas limit! I know energy prices are high. Fading tariffs may offset this component of inflation, but we need more time to see a clearer picture. 

The Fed. Under Chair Warsh the Fed is in a hiking cycle, and October was a skip rather than a pivot. The Oct 28 statement upgraded activity from “solid” to “strong” and called for a “timely return” of inflation to target; two dissented in favor of a hike (Cleveland’s Hammack and Governor Cook). The press conference was more dovish in tone (“we removed a dose of accommodation”), and Warsh pushed back on the idea that AI investment is insensitive to rates, noting that “interest rates get in the cracks.” A December hike is near-certain (okay, it’s dropped from 90% to 68%, but it’s still pretty high), provided financial markets do not deteriorate further. Another is expected in March.

There does seem to be a growing risk that the market is underpricing rate hikes. For context, the median Fed tightening cycle since 1972 has lasted 19 months and totaled about 394 basis points.

Rates. The two main drivers of rates seems to be the IRAN war and the deficits, not policy or politics.

Labor market: steady, slightly loosening, and not the Fed’s problem

The labor market is the quiet part of the story. For October payrolls, most estimates were 55K to 72K, but the actual number was exceptionally strong at 165K! Most forecast unemployment to tick up, but it actually fell to 4.1%. It’s good, because we all want more people to be employed and to have good jobs. But wages must outpace inflation, or real wages decline. As I write this, the Fed’s Mary Daly calls October’s jobs report “good news,” expressing optimism about the US workforce. However, she stresses that wage growth isn’t booming and that real wages are falling. This implies that this may have a residual effect on inflation. 

The employment angle on the AI buildout matters for the policy debate: a large data center employs roughly 1,500 workers at peak construction but only 50-100 in operation, while manufacturing employment in the data-center supply chain is up 13% (about 158,000 jobs) since 2019, compared with a 1% decline for manufacturing overall. The bottom line is that a soft payroll print would have fed the risk-off narrative (which is one reason the market stayed positive today, at least while I write this note), but it may not change the December setup for a rate hike.

Washington & policy: a Blue Wave, but has anything changed?

The Nov. 3 midterms produced a wave not seen since 1991. Democrats are projected to hold roughly 55 Senate seats and about 255 House seats. From a market perspective, one could attribute the results to the IRAN war, tariffs, and ICE. Each has created uncertainty, and the market vehemently hates uncertainty, regardless of whether the policies are right or not.

Most consequential for markets is the unseating of Texas Governor Abbott. This may affect the new Texas Stock Exchange and local data centers. Today (Nov. 6^th), the new Texas governor-elect criticized data centers for exploiting local resources and said the relationship with Elon Musk must be on Texas’s terms. That sounds like a change to me. Although I think local change will and does happen, it appears different at the federal level.

What actually changes: less than the headline. Assuming a peaceful transition of power, the President keeps the veto and his executive authorities, so a Democratic House and Senate produce nearly the same federal policy. The lame-duck session (Congress returns Nov 9 for about five weeks) is expected to go out quietly: low odds of a final reconciliation bill, a defense top-up, or a debt-ceiling increase, with permitting reform and China-related NDAA items likely slipping to next year.

The pivotal 2027 event is a debt-ceiling showdown, which may be the catalyst for any late-2027 healthcare deal on Medicaid cuts and ACA credits.

Data centers are the market’s real post-election worry. Voters approved nearly every data-center moratorium on the ballot, and the Texas result put the AI buildout in question. The tone is changing from build-now to build-right. Moratoria are by design temporary, new conditions will raise costs and extend timelines, and projects already under construction are exempt. Roughly 75% of the 34 GW planned across the top 20 states for 2027 is already under construction; of the 276 GW planned overall, 41% sits in states whose governors or governors-elect have voiced support for a moratorium or pause, with Texas the largest share. Delays are not cancellations, and the bigger risk is to 2028-and-beyond investment decisions, not to 2027 growth.

Data-center projects facing opposition or legal challenges, by state. Source: Visual Capitalist, “Mapped: U.S. States With the Most Data Center Opposition,” Oct 29, 2026; data from SueDataCenters.org as of Oct 20, 2026.

Current narratives that may be overdone (Nov 5): “sell everything”; the rotation out of semis into hyperscalers; “buy healthcare” (a 2027 deal is plausible but late and partial, and the wave adds drug-pricing risk for pharma); “buy clean energy” (no realistic path to reversing the IRA credit cuts); “sell oil” (no line of sight to a resolution that durably lowers prices); and “buy tariff losers” (guardrails on presidential tariff authority are plausible but would not remove sector tariffs).

IRAN is the key. It appears the policy is oscillating between “low-keying it” (blockade plus sanctions enforcement) and escalation-or-deal threats. An escalation before the transition of power is my worst-case scenario. If there is no escalation, at least not until after the transition of power, then energy prices most likely stay range-bound, and yields remain unchanged. This creates a clearer picture, and the market can have room to expand. 

The Markets:

The S&P 500 spent most of the past month pinned in the 7,600-7,800, but yesterday it had the sharpest de-risking of the year.

1. Narrow and range-bound (early-to-mid October). Tech had driven roughly 70% of the index’s ~15% YTD total return. The top 10 contributors accounted for ~61%, and cap-weight was beating equal-weight for a fourth straight year (+2.4% YTD). The forward P/E had fallen 4.3 turns from its 23x peak a year earlier to 18.7x, close to the 19.1x ten-year average, with the median stock at 17.1x and all 11 sectors cheaper on the year. The week of Oct 12 brought a “trifecta” of AI worries, hot CPI/PPI, and higher oil, and the first -1.1% days since July.
2. Disappointing Earnings (week of Oct 26). Disappointing hyperscaler results triggered an “anything but tech” rotation: S&P 500 -0.6%, Nasdaq-100 -2%, Russell 2000 +1.3%, with tech down 6% over two days and Alphabet down 14%. For the first time in 12 quarters, calendar-2026 AI capex estimates were not meaningfully revised higher, and consensus now sees spending peaking in 2028. This also aligns with the election results. Investors shifted from capex to AI profitability. Earnings season was solid on the growth front (2026 operating EPS remains on track for one of the fastest growth rates since 1990; margins expanding in most sectors), but dollar-weighted bottom-line surprises were the weakest since 2022.
3. Post-election de-risking (Nov 4-5). Thursday’s sell-off hit every S&P 500 sector, with tech hit the worst. This appears to be a clear de-risking with possible year-end positioning.  It’s amplified by leverage and crowding rather than a fundamental break.  It’s just hard to step in a invest when the market falls so quickly.
4. Watch IRAN. If there is no escalation in IRAN, I am cautiously optimistic that market volatility may continue, but not necessarily to the downside. Thus, I am cautiously optimistic about positive returns through year-end.

Anatomy of the sell-off. The moves during the de-risking were extreme. The momentum factor had its worst day since 2020, and the market-beta factor is roughly two standard deviations stretched to the downside. Semiconductors sit 38% below their June highs, not seen/felt since 1995. On the other side of the trade, Staples just had its best month since October 2000, which could be one reason Walmart sold off today, down 3% at one point. This defensive rotation will probably continue, as peak-fear gauges aren’t there yet: the VIX rose sharply but hasn’t reached the key 40 level, although the VIX curve did invert. The silver lining is overall market valuation: the S&P 500 now trades at 17.9x forward earnings, right at its 15-year average of 17.8x, and the Nasdaq-100 at 18.3x is below its 15-year average of 20.8x.

Largest intra-year drawdown in the SOX semiconductor index each year; 2026 is 38% year-to-date vs. a 30% average and 26% median since 1995

Next-twelve-month P/E for the S&P 500 (17.9x vs. 17.8x 15-year average) and Nasdaq-100 (18.3x vs. 20.8x).

Short-term trade. Though I’m cautiously optimistic about Tech, Industrials, Energy, and Healthcare, Healthcare looks to be the favorite. It seems to offer both upside potential and defensive characteristics. The year-end rally call holds, but it depends on no escalation in IRAN. On November 17th, NVDA’s report will tell us a lot. But we don’t yet know when Anthropic will IPO. As for positioning into year-end (at the moment), we are slightly underweight in tech and overweight in earnings growth.

Risks & what to watch

The bear case is a continued slowdown in AI, the economy following suit, and inflation remaining elevated as the Fed raises rates. The base case is the opposite: robust growth, a more measured buildout, a December hike, and a positive year-end.  All conditioned on the IRAN war not escalating.

DISCLOSURE: The information contained in this document is for informational purposes only and should not be construed as investment advice. Summit Global Investments (SGI) makes no representations or warranties as to the accuracy or completeness of the data provided. Forward-looking statements are based on current expectations and involve inherent risks and uncertainties that may cause actual results to differ materially. Past performance is not indicative of future results, and all investments carry risks, including potential loss of principal. SGI’s investment strategies are subject to change without notice.

1 If you don’t have a financial plan that considers your financial goals, objectives, and risk tolerance, please let us know.  We would be more than happy to work with you to create one.