Sector Detector: Contradictory signals challenge Fed policy, while AI suffers from misguided opposition

 

  by Scott Martindale
  CEO, Sabrient Systems LLC

 

Quick note: Sabrient’s annual Forward Looking Value 14 portfolio just launched last week on 8/26 with a diverse mix of 31 stocks across 9 business sectors and a 50/50 Large/SMID-cap mix. Think of it as a less-concentrated and more value-oriented version of Baker’s Dozen. Last year’s FLV 13 terminates 11/17 and is up +38.4% vs. +19.4% for S&P 500 Value (SPYV) and +20.2% for S&P 500 (SPY) as of 9/2. I personally like the UIT structure as a diversifier for a client’s equity portfolio because they are unmanaged—while most other products are managed to some extent, even quarterly rebalanced rules-based indexes—and most active managers tend to underperform their benchmarks. Also, value stocks tend not to be as volatile, so value or dividend portfolios tend to fit well with the unmanaged, buy-and-hold-for-15-months structure.

Overview:

The US economy is changing faster than the traditional macro indicators can explain it. On the one hand, Q2 GDP (just +1.5%) and jobs growth look sluggish, inflation is sticky, debt and bond yields are surging, and the bifurcated “K-shaped” economy is causing poor consumer sentiment and public dissatisfaction—to the point that socialism(!) is gaining traction. The top 10% by net worth own 87% of stocks and 68% of total net worth, while the bottom 50% own just 1% of stocks but carry 52% of consumer debt. Consumer debt is rising while the personal savings rate remains low to support spending. Real (inflation-adjusted) consumer spending was essentially flat in July, while the personal savings rate finally edged up slightly to 3.0% in July after a steady decline from 6.4% in January 2024 to as low as 2.6% in June 2026.

On the other hand, the official Q2 GDP reading was held down by high AI-related imports and inventory drawdowns, which actually reflect robust economic activity. Private domestic demand (considered a truer signal of underlying economic health) was quite strong in Q2 at +4.2% annualized rate, consumer spending was up +3.4%, business investment excluding housing was up +8.5%. Indeed, the Atlanta Fed’s GDPNow now forecasts Q3 GDP at 4.7% (as of 9/3). Corporate profits, productivity, margins, cash flow, and capex are massive. Jobless claims in our “low-hire, low-fire” labor market are near their lowest level in decades, which Fed chairman Kevin Warsh calls “an empirically robust real-time indicator” constrained only by flattening labor supply (a lack of willing workers).

Perhaps the low personal savings rate is to be expected as Baby Boomers retire and draw down their wealth, and household debt as a percentage of disposable personal income has stopped rising in Q1-Q2. Stocks are near all-time highs even though interest rates have become punitive (compared to what the broad economy had become accustomed to), and market breadth is improving.

While these data points might seem like contradictions, they may in fact be a predictable reflection of a structural transition toward a more capital-intensive, AI-driven, productivity-led economy.

Regarding inflation and Fed policy, Warsh stated in his Jackson Hole speech, “Certain sectors—like housing and agriculture—are showing strains. But, on balance, I would be hard pressed to describe broad financial conditions as restrictive…. We must be confident that underlying inflation is moving to our objective, clearly and at sufficient speed. Otherwise, we have work to do.” But you likely heard it paraphrased as simply, “Inflation is running too hot…financial conditions are not restrictive…we have work to do,” which is clearly more hawkish than his actual statement and seemingly a fait accompli that rate hikes are imminent. Fed funds futures now imply 2+ rate hikes over the next 12 months, with 50% odds of a 25-bp hike at the upcoming FOMC meeting this month.
 
However, I humbly disagree. I think Warsh’s actual statement cleverly balanced between sounding sufficiently hawkish to appease the “bond vigilantes” while leaving himself an opening to avoid hiking rates. Most of the non-supply-shock components of inflation are relatively subdued, and by some key metrics, underlying inflation is indeed moving toward the 2% objective—for example, Warsh’s preferred Trimmed Mean PCE has held steady at +2.3% in June/July, and the real-time, blockchain-based Truflation CPI is also +2.3% (as of 9/3). Meanwhile, all is not well in this K-shaped economy. The labor market is stagnant, the lower/working class consumer and wage earner is struggling, and monetary policy is clearly restrictive in the most rate-sensitive industries and demographics.

We can’t continue to rely solely on massive AI capex and retiring Baby-Boomer spending (neither of which are particularly rate sensitive) to power our economy while watching the lower leg of the K languish. All it takes is for one hyperscaler to announce a significant reduction in capex and we’ll likely see economic growth fall—along with the stock market.

So, I continue to believe a rate cut is appropriate—certainly not a hike (as fed funds futures and Polymarket confidently predict). A hike would only further hinder those struggling cash-poor segments (many of whom are warming up to the socialist “free stuff” rhetoric) while doing little to restrain the cash-rich AI trade or resolve the event-driven supply shock. I feel alone in the woods on this, but I see the neutral rate at 3.0-3.25%. Unfortunately, such an action apparently would be quite unpopular in the bond market and thus is unlikely to happen anytime soon, thus keeping the economy bifurcated. However, in speeches this week, Fed governor Christopher Waller and NY Fed president John Williams pumped the brakes on the tightening narrative (“Give disinflation a chance”).

Kevin Warsh has suggested in the past a preference for fulfilling his mandate of price stability through a combination of: 1) a lower fed funds rate, 2) continued balance sheet reduction to constrain money supply growth (including a revised Treasury-Fed relationship that gives the Treasury more say over balance-sheet decisions), and 3) structural reform of how inflation is measured (replacing the lagged and imputed components). This approach would ensure that liquidity is cheaper to access, preventing credit markets from locking up without flooding the entire financial system with inflationary liquidity. Less Fed demand for bonds might steepen the yield curve somewhat and keep 30-year mortgage rates elevated, but a lower fed funds rate reduces rates on short-term construction loans and adjustable mortgages.

According to global liquidity expert Michael Howell of CrossBorder Capital, the “Fed does not need to cut Fed funds to loosen US monetary conditions. It can hold the policy rate steady while Treasury bill issuance, reserve management and bank balance-sheet expansion deliver the liquidity impulse. This is the essence of ‘Treasury QE’: fiscal expansion financed at the short end, supported by enough reserve liquidity to keep funding markets orderly.”

By the way, regarding the sudden tariff flare-up with Canada, our neighbor’s retaliation could have targeted industrial inputs but instead it’s notable that they only apply to food and consumer products in an (admittedly) targeted effort to pressure Republican candidates in competitive US midterm election races in swing states like Pennsylvania, Michigan, and Wisconsin, and red states like Ohio, Indiana, and Kentucky.

In my full commentary below, I discuss:

1. Broadening continues as small caps, value, and the S&P 493 carry the baton
2. High earnings vs. high yields and the impact on valuations
3. The NVIDIA juggernaut and AI Phase II: proving the economy-wide ROI
4. Materials and power constraints on the AI buildout
5. K-shaped economy, inflation, and the Fed’s policy dilemma
6. Runaway Federal debt and the need to grow our way out of it
7. Crude oil inventory drawdown and oil company windfall profits
8. My Final Comments essay: Is AI opposition a nuclear power redux?
9. Sabrient’s sector rankings, positioning of our sector rotation model, and some top-ranked ETF ideas

According to Morgan Stanley, AI’s diffusion across the global economy is a $40 trillion opportunity that will rely on affordable compute. Indeed, there is so much demand for compute, some say it is becoming a new asset class—like stocks, bonds, commodities, real estate, private credit, crypto, collectibles, royalties, cash-flowing private businesses, or carbon credits. Pending regulatory review, CME Group and Silicon Data have announced plans to launch futures contracts on compute by tracking daily and hourly on-demand GPU rental cost indexes for NVIDIA’s H100 and next-gen Blackwell B200 chips. This allows a company reliant on AI to lock in a future price and ensure delivery from a new datacenter that takes 2-3 years to build.

Some are worried about falling token prices hurting the projected ROI on capex. But at the Google I/O 2026 Conference, CEO Sundar Pichai said that AI is using 3.2 quadrillion tokens per month, which is up 7x YoY. Indeed, rapidly falling inference costs can create demand that didn’t exist at previous prices. This is largely due to Jevons Paradox, which says efficiency gains lower production cost, which gets passed on to customers, thus driving up demand, which increases total resource usage over time—i.e., demand for a product or resource rises as price falls (aka demand elasticity). So, if the hyperscaler can produce and sell compute tokens cheaper, usage surges, and the hyperscaler makes more money overall. It seems this trend can only be disrupted by misguided (or deliberately subversive) politicians hell-bent on obstructing this broad, multi-layered, truly all-in, capital-investment cycle, as I discuss further in my Final Comments essay below.

This is just one of the many reasons I expect to see, through the end of this decade and likely beyond, smaller government and less low-ROI government spending in favor of more high-ROI capital allocation from an unleashed private sector as the primary engine of organic economic growth through fiscal support like favorable tax policy, deregulation, and other supply-side incentives for reshoring/onshoring to increase productive capacity.

As such, I still think the S&P 500 could reach 8,000 by year end. However, I also think a further market pullback this month is likely, perhaps to test support at the 50-day moving average, and as a reminder, September is the only month since 1975 in which the S&P 500 has averaged a net loss (-0.8%). But assuming continued healthy market broadening beyond the Big Tech titans, and praying the Fed does not start tightening, I continue to see opportunities in active stock selection, equal-weight indexes, value stocks, cyclical sectors, small caps, and bond-alternative dividend payers. I also continue to believe the Healthcare sector, which finally came alive this year, offers tremendous growth opportunities as it leverages AI. Sector earnings for Healthcare Select SPDR (XLV) are projected to rise by 22% YoY in 2027, which is second only to Technology.

Sabrient’s quant-based, actively selected (“quantamental”) Baker’s Dozen, Forward Looking Value, Dividend, and Small Cap Growth portfolios have been largely outperforming their benchmarks—several by substantial margins. And as a reminder, our Earnings Quality Rank (EQR) is licensed as a quality prescreen to the actively managed, low-beta First Trust Long-Short ETF (FTLS), which now has more than $2.5 billion in AUM.

Here is a link to this full post in printable PDF format. As always, I’d love to hear from you! Please feel free to email me your thoughts on this article or if you’d like me to speak on any of these topics at your event.  Read on….

Market Commentary:

For your holiday weekend reading pleasure, this month’s commentary is on the long side. But I promise you won’t regret taking the time to get through it all!

The cap-weight S&P 500 (SPY), equal-weight S&P 500 (RSP), Dow Jones Industrials (DIA), S&P MidCap 400 (MDY), and Russell 2000 (IWM) all surged to new all-time highs last month—and small caps are outperforming them all YTD. As shown in the chart below, only Nasdaq 100 (QQQ) among the major indexes has not returned to its high from early June, as investors rotate capital into other high-potential opportunities rather than selling and going to cash.

According to DataTrek Research, the cap-weight S&P 500 is now 25% more volatile than its equal-weight counterpart, the highest in 16 years, so the equal-weighted S&P 500 is a reasonable way to achieve US large cap stock exposure with less volatility. Investors seem to agree, as the S&P 500 ETF (RSP) is approaching $100 billion in AUM, with performance led by many of the “S&P 493” rather than the MAG7.

Remarkably, the earnings beat rate in Q2 hit 86%, perhaps the highest ever. But that barely begins to explain the magnitude of the YoY earnings growth, which was an incredible 52%! However, it should be noted that unusually large investment gains (beyond operating earnings) at Alphabet (GOOGL) and Amazon (AMZN) inflated the headline figure, but even after excluding those two outliers, earnings still grew an exceptional 34% YoY. And encouragingly, it wasn’t just the MAG7 as the “S&P 493” also performed quite well, growing earnings about 32%. Moreover, the strength extended well beyond the S&P 500, as BEA data showed economy-wide corporate profits from current production surged by $401 billion (9.1%) versus Q1—up roughly 23% YoY versus Q2 2025 and marking the second-largest quarterly dollar increase on record. Corporate profits now equal roughly 14% of GDP, which is near an historic high, led of course by the AI trade.

Also notable is the impressive recovery in the software industry from the “SaaSpocalyse,” which posited that AI would destroy enterprise software. The chart below shows the ratio of iShares Expanded Tech-Software (IGV) to iShares Semiconductor (SOXX) along with the Technology Select Sector SPDR (XLK). You can see the recovery in software relative to semiconductors since mid-June, even as the broader Technology sector has only gone sideways.

So, underlying breadth in earnings growth was unusually strong, and earnings revisions are still moving up quickly, rather than following the normal pattern of being marked down as the year progresses. For example, at the beginning of April, consensus estimates for the S&P 500 were $323 for CY2026 and $377 for CY2027. Now we’re around $360/$410 for 2026/2027. And some on Wall Street are even more bullish, like JP Morgan (JPM) which raised its forecast to $365/$420.

Applying these numbers to valuations, while the S&P 500’s FY2026 and FY2027 EPS estimates are rising, the forward P/E has fallen from about 23.2x at the start of the year to 21.5x today, indicating the market has rallied solely on earnings growth even as multiples have contracted. Assuming FY2027 EPS indeed reaches $420, and assuming the 10-year bond yield is nearing a peak and pulls back to around 4.5% (implying a 22x forward P/E on the S&P 500, according to the Fed Model), then the index could approach 9,250 by late next year, or about 20% above the current index price.

The 30-year Treasury yield has now spent 57 days above 5.0% in 2026—the most in any year since 2006. The UK’s 30-year hit its highest level since 1998. Japan’s 10-year hit 3% for the first time since 1996. It is the uncertainty of how high yields might ultimately reach that is holding down P/E multiples. In other words, it’s not just the current yield that influences today’s forward P/E but fears of how much higher it might go. As Nick Colas of DataTrek opined, “If there was a way to guarantee that the 10-year yield would go no higher than 5.0% for the rest of the decade, the forward multiple on the S&P 500 would immediately jump by 2 points.” That’s how critical the yield curve can be.

Some commentators (like respected economist Ed Yardeni) believe yields are simply normalizing after a long period of abnormally low yields following the Global Financial Crisis.

The NVIDIA juggernaut and AI Phase II: proving the economy-wide ROI:

Of course, the most anticipated earnings report came last week from juggernaut NVIDIA (NVDA), and it was incredible, resoundingly reinforcing the structural AI capex narrative with +106% YoY growth in revenue ($96.2 billion) and +124% YoY growth in operating income ($63.7 billion)—which implies a stunning 66.2% operating margin. Also, it showed +117% YoY growth in its Data Center segment ($89.0 billion, or 92.5% of total sales, including $48.7 billion from hyperscalers). While total sales to hyperscalers rose +102% YoY, non-hyperscaler sales rose even more at +138% YoY as demand broadened into sovereign-government cloud builds (so they can own their AI compute power instead of relying on foreign Big Tech), traditional enterprise IT and SaaS, and AI-native startups.

Then the firm announced a huge guidance beat on next-fiscal-year revenue growth of 70% (versus 45% consensus). The chart below illustrates the astounding growth in revenue and operating income, as well as the extraordinary operating leverage of the AI buildout.

The next chart further illustrates the huge operating margin it commands. In four years, quarterly revenue increased from $5.9 billion to $96.2 billion, while operating margin expanded from 10.1% to 60.8% one year ago to 66.2% today.

Even more incredible is that the company insists its massive 70% revenue growth projection is lower than underlying demand due to the shortage of memory chips (even assuming no sales at all to China, due to uncertainty from both US export restrictions and China regulatory hurdles). NVIDIA’s inability to meet the full level of demand is also due to datacenter permitting and construction timelines, power grid expansion, and supply-chain bottlenecks that hinder its ability to expand physical manufacturing capacity as fast as product demand is growing. According to McKinsey, there is as much as $1.5 trillion in infrastructure awaiting federal approval, including power plants, pipelines, and factories.

But the main constraint is limited capacity in advanced node chip packaging and tight supplies of HBM and DRAM memory chips from the likes of Micron (MU), SK Hynix (SKHY), and Samsung—a situation they say might not ease until 2028 (aka “RAMageddon”). Add in higher manufacturing costs for its next-generation production and these constraints will compress NVIDIA’s margins a bit, from today’s 75% to perhaps 72% (still amazing).

Regardless, NVIDIA’s underlying message was clear: it has a buyer for every chip it can make. According to CEO Jensen Huang, “AI infrastructure buildout is at full steam… We’ve got a huge year coming up next year, and it’s going to be pretty extraordinary.” Indeed, total capex among the top five hyperscalers is expected to increase from $800 billion this year to $1.3 trillion next year, and NVIDIA is a primary recipient. According to the WSJ, Anthropic is expected to tell its IPO investors that its total addressable market (TAM) is $30 trillion—which happens to match current US annual GDP. As such, it is considering a $100 billion raise at a $2 trillion valuation. This is similar to SpaceX’s (SPCX) IPO claim of a $28.5 trillion TAM.

As DataTrek Research opined, “Big Tech now treats AI spending as existential rather than optional. Markets have historically underpriced compounding innovation, and betting against US large cap Tech has been a losing trade for decades… Big Tech is doing what it always does, following a ‘build it and they will come’ [Field of Dreams] strategy, albeit at unprecedented cost.”

Does that relieve investor fears of a growing AI bubble? Not really. Author and AI-skeptic Ed Zitron questions with solid data and sound logic the circular financing, depreciation, utilization rates, inference costs, and hyperscaler capex returns. He wonders where the ultimate return on all this spend is going to come from. He argues that if NVIDIA sells $100 billion of GPUs to hyperscalers, and those hyperscalers spend hundreds of billions (if not trillions) of dollars building AI capacity, eventually the end users in aggregate must generate substantially more than all of that in economic value to justify it all, which he sees as unlikely.

However, history displays numerous transformational technologies where the investment bubble was real but the technological revolution was also real. Railroads in the 19th century to fiber-optic networks and the dot-com boom at the end of the 20th century leading into the mobile/cellular revolution to start the 21st century are obvious examples in which investors massively overbuilt physical infrastructure, many companies went bankrupt, and capital was destroyed, but over time society ended up making full use of the infrastructure with enormous ROI going to the surviving companies. Today, labor productivity is quite high, but total factor productivity (combining both labor and capital) has remained modest. The Federal Reserve Bank of San Francisco observed that this divergence mirrors the early 1990s leading into the Digital Revolution (desktop computers and Internet followed by high-speed fiber-optic networks and smartphones). Today, massive capital investments into AI infrastructure mean workers have better tools, but the systemic, macro-level efficiency gains and toolkits of AI have not yet been fully realized across the broader economy.

We won’t know for some time what ROI the hyperscalers will achieve longer term, nor can we quantify how much boost to productivity and profits AI implementers across the broad economy ultimately will achieve. But the possibilities are vast and exciting, ranging from cloud services, enterprise SaaS, on-device integration, and robotics, to national security, medical research, and military capability. So, the “day of reckoning” on AI spend is still years away, but as I discussed in my August post, our economic health depends on the promise of AI to come true.

The day after its earnings call, NVDA shares rose 8% and has sights on its all-time high from 5/14. Valuations on the MAG7 are reasonable, relatively speaking. Nevertheless, the overall market reaction suggests that rather than continuing to bid up share prices for NVIDIA and the hyperscalers, investors at this phase of AI buildout are not worried about whether AI demand is still growing—because it’s growing like a weed. Rather, it’s about high-ROI capital allocation, physical constraints, competitive economics, and who captures the ultimate return on what is perhaps the largest private-sector infrastructure buildout in history.

Materials and power constraints on the AI buildout:

A key constraint on this buildout is that, although AI is digital at the point of use, it is enormously physical at the point of scale. The unprecedented capex flowing into datacenters requires not only GPUs and electricity but massive amounts of copper, aluminum, grid equipment, cooling infrastructure, and critical minerals. Copper may become particularly problematic given the combination of rapidly rising datacenter demand, declining ore grades, limited US mining and refining capacity, and long lead times for new mines. And datacenters must compete for these same materials with those building grid expansion, EVs, robotics, defense, solar, and other forms of electrification.

This creates both economic and geopolitical vulnerability. China dominates the processing and refining of many critical minerals—including rare earth elements (REEs), gallium, and germanium—while the US must import copper and other materials. Thus, even if we can maintain the lead in frontier AI models and advanced chips, our ability to deploy AI infrastructure at the scale envisioned could increasingly depend on access to commodities and processing capacity controlled elsewhere, including adversarial countries.

As a result, the AI trade now extends beyond hyperscalers, semiconductors, and software into minerals mining and processing capacity and producers of copper, aluminum, rare earths, and other critical materials in allied countries. That means everything from power utilities and electrical component makers. In fact, some power companies have already sold out their production through 2030. High-voltage transformers and heavy-duty gas turbines are now even harder to obtain than the newest generation of NVIDIA chips—simply because none are available.

Indeed, today’s best investment opportunities have broadened from advanced semiconductors toward memory [e.g., Micron and SK Hynix], AI hardware and servers [e.g., Dell (DELL), Hewlett Packard (HPE), and Cisco Systems (CSCO)], networking and optical components [e.g., Broadcom (AVGO), Coherent (COHR), and Marvell (MRVL)], power suppliers [e.g., Vistra (VST), NextEra Energy (NEE), and Bloom Energy (BE)], electrical equipment and turbines [e.g., Eaton (ETN) and GE Vernova (GEV)], cooling and UPS [e.g., Vertiv (VRT) and Johnson Controls (JCI)], and all the many commodities and critical metals it takes to build them. Huge order backlogs abound.

K-shaped economy, inflation, and the Fed’s policy dilemma:

Q2 corporate earnings reports also laid bare the worsening K-shaped economy, as lower/middle-income consumers, wage earners, small businesses, and the housing market are weakening while corporations continue spending aggressively on strategically essential AI infrastructure. Revolving consumer debt has eclipsed $1 trillion, as wage growth has stalled and personal savings rate fell to 2.6% in June (although it ticked back up a bit to 3.0% in July), down from 6.4% in January 2024. Home equity lines of credit balances rose by $13 billion to $459 billion. Credit card balances rose by $21 billion to $1.26 trillion, and auto loan balances increased by $28 billion to $1.71 trillion. And although delinquency rates are stable with 4.7% of outstanding debt in some stage of delinquency, the “severely derogatory debt” category rose from 1.8% to 2.0% in Q2, the highest since Q3 2020, and 10.6% of student-loan balances were 90+ days delinquent, the highest since Q1 2020.

On the other hand, the better news is that household debt as a percentage of disposable personal income has flattened out at just above 11.0% after rising rapidly from about 9.0% during the pandemic stimulus, and household debt as a percentage of GDP remains in a long-term downtrend since the GFC (other than a brief pandemic blip when GDP fell -9%). Also, regarding the low personal savings rate (3.0%), this would be expected as Baby Boomers retire and draw down their wealth, thus lowering the national savings rate. So, the signals are conflicting.

July CPI fell to +3.4% YoY from June’s +3.5%. Not surprisingly, Core CPI excluding food & energy came in much lower at +2.5%, versus +2.6% in June. July PPI was +4.7%, down substantially from +5.5% in June. Shelter remains the largest component of core inflation, but it is finally showing persistent signs of abating, while healthcare continues its upward march due to aging demographics and structural service demands (offset by falling drug prices). Obviously, Fed policy has more impact on shelter than on healthcare.

The July inflation reports finished with the PCE release on 8/26. Headline PCE came in at +3.7% (no change from June), and former Fed chair Jay Powell’s preferred metric of core PCE was +3.3%. But new Fed chairman Kevin Warsh’s preferred metric of “Trimmed Mean PCE” came in at +2.3%—which is the lowest level since November 2020, a full percentage point below core, and only slightly above the Fed’s “official” inflation target of 2%.

Trimmed Mean PCE filters out the most extreme price swings that can distort the total, offering a more stable and relevant perspective. Moreover, it has been quite stable at 2.3–2.4% all year long, while headline PCE jumped around between 2.9–4.1%. So, in my view, if you ignore the event-driven energy and supply-shock impacts upon which Fed policy has little impact (short of inducing a recession), other inflationary pressures haven’t changed much this year.

The two charts below compare trends in these key metrics over the past five years. The top chart shows YoY comparisons, and the lower chart shows rolling 3-month annualized averages (which quantifies the current trend in inflation). In both charts, the abrupt surge in headline PPI and CPI would be frightening if indicative of a structural problem in the global economy. However, we know this was event-driven due to disruptions to supply chains and the spike in oil, gas, and fertilizer prices from the Iran conflict and blockade of the Strait of Hormuz. As these supply chain pressures have begun to ease, crude oil (WTI futures) fell from a high around $108/bbl in mid-May to below $80 but are back up around $90 now. But it looks much better when you exclude food & energy prices, with Core CPI at +2.5%.

Even more encouraging is the lower chart showing 3-month rolling annualized averages. Although headline PPI and CPI annualized 3-month trends were both startlingly high just a few months ago, they have fallen hard. In particular, although 3-month annualized Core PCE is +3.0%, Core CPI is just +1.6%, headline PCE +2.2%, headline CPI +0.49%, PPI +1.2%, and Trimmed Mean PCE +2.2%. These trends seem quite encouraging.

And then we have the alternative metrics I like to follow, including the real-time (i.e., unlagged) blockchain-based Truflation, which is published daily and tracks over 15 million live price points (vs. CPI’s 80,000), including real-world housing components (e.g., mortgage rates and home prices) rather than delayed imputed metrics like Owner’s Equivalent Rent (OER) used in CPI. Truflation CPI came in at +2.1% at end of July but rose slightly to +2.3% at the end of August.

Looking ahead, the Cleveland Fed’s Inflation Nowcast model for August predicts CPI will be +3.38%, Core CPI +2.38%, PCE +3.80%, and Core PCE +3.40%, as of 9/3. Also, the 5-year breakeven inflation rate is +2.35%, as of 9/2, reflecting the implied forward expectations of investors in 5-Year Treasury Constant Maturity Securities and 5-Year Treasury Inflation-Indexed Constant Maturity Securities.

So, Trimmed Mean PCE, Core CPI, the 3-month annualized trends, 5-year breakeven, and Truflation all put underlying inflation much closer to the Fed’s 2.0% target. Meanwhile, the labor market is stagnant, the lower/working class consumer and wage earner is struggling, and monetary policy is clearly restrictive in the most rate-sensitive industries and demographics (despite what Warsh intimates in his official comments).

Furthermore, we can’t keep relying solely on massive AI capex and retiring Baby-Boomer spending (neither of which are particularly rate sensitive) to power our economy. Ed Zitron and Josh Brown on The Compound podcast summarized the multiple potential threats that could spell the end to the AI boom with any of these five events: 1) a hyperscaler “throws in the towel” by significantly scaling back on capex, 2) a hyperscaler, most likely Oracle (ORCL), has its huge debt downgraded, 3) a major AI startup runs out of money, 4) an AI IPO does not go well, or 5) one of the many neoclouds [specialists in renting out AI compute, like CoreWeave (CRWV) or Nebius (NBIS)] can’t raise debt. Capex cycles typically falter when willing lenders dry up, and don’t forget that Oracle was cut to BBB- (right above junk status) on 7/9, largely due to its OpenAI concentration risk.

So, I continue to advocate for a rate cut—but certainly not a hike, which only would be recessionary and further hinder those struggling segments in the lower leg of our K-shaped economy while doing little to restrain the cash-rich hyperscalers’ capital spending or resolve the event-driven energy/supply shock or tariff-related price increases (which are one-off, like a new tax, and not rising).

Once the energy/supply shock ends, secular disinflationary trends can take full control once again, including aging demographics, slowing global population growth, accelerating disruptive innovation like AI and automation, rising productivity, falling shelter and energy costs, modest liquidity growth, a stable dollar, the deflationary impulse on the world of a struggling China dumping goods, and the redundancy and additional industrial capacity (aka “duplicative excess capacity, for national security) of America’s factory reshoring and re-industrialization.

The current event-driven supply shocks are reflected in the New York Fed’s Global Supply Pressure Index (GSCPI), which computes a z-score (number of standard deviations from the long-run average). Back in April, it surged to +1.84, which suggested elevated supply chain pressures (but still well below the December 2021 high of +4.44), but it has retreated to +0.79 (as of July).
 
So, I believe inflation likely has seen its peak from the event-driven disruptions, and rising longer-term Treasury yields likely have been driven more by hawkish signals from the Fed and bond vigilantes demanding higher real yields than from any actual concerns about structural inflation becoming embedded into the economy. Indeed, rising productivity, moderating event-driven supply chain disruptions, and a resumption in disinflationary secular trends promise to make inflation recede back below +2.5%—and by some metrics, as I pointed out earlier, it already has.

Economist Ed Yardeni, who coined the term “bond vigilantes,” believes those folks should not be feared unless the 10-year Treasury yield exceeds nominal GDP growth. Today, the 10-year is below 4.8% while nominal GDP in Q2 2026 grew 6.6% YoY. Moreover, he observed that the 10-year TIPS yield is historically a proxy for real GDP growth, as illustrated in his chart below. So, he believes the 10-year yield is “where it should be relative to the fundamentals” and that concurrent stock market strength “suggests that investors read higher yields as evidence of economic growth rather than a threat to it.”

While Warsh wants to make the Fed less relevant to daily monitoring among market players (by moving away from providing forward guidance and the let the markets price themselves), Treasury Secretary Scott Bessent is making the Treasury increasingly more relevant.

As James E. Thorne of Wellington-Altus explained on X.com, “The Fed controls the overnight rate. Treasury controls the supply, maturity, liquidity and increasingly the downside risk of the debt that the world uses to price duration. That power now matters more. America is carrying nearly $40 trillion in federal debt. Mortgage rates, corporate borrowing costs, infrastructure finance and the valuation of every long-lived asset depend not on the fed funds rate but on the term premium embedded in the long end. Treasury issuance, buybacks and maturity management therefore shape monetary conditions more directly than another quarter-point move from the Federal Open Market Committee.”

Indeed, that was the message of Bessent’s recent announcement of a doubling of the Treasury’s liquidity-supporting buyback operations for notes and bonds with 10 or more years to maturity through the end of its refunding cycle in November, from $2 billion to $4 billion per operation. It will purchase longer-maturity, less liquid bonds from the market, using cash raised through new shorter-maturing debt issuance—reminiscent of previous “Operation Twist” tactics by the Fed—to smooth out anomalies in the yield curve. Bessent says it is a response to institutional investors’ strong eagerness to unload such illiquid bonds at a discount.

Although it looks a lot like Federal Reserve QE, the net impact is no change in the total debt or money supply since it is an asset swap of shorter-term debt used to buy longer-term debt, which serves to reduce long-term supply and depress rates—even while Warsh and the Fed continues to seek to shrink its balance sheet by allowing bonds to mature without reinvesting the proceeds (instead, the Treasury must transfer cash to the Fed by issuing replacement bonds bought by banks and investors, which drains liquidity from the financial system). On the other hand, if Bessent chooses to draw down the nearly $1 trillion TGA to buy back T-bonds instead of “twisting” (issuing T-bills as a “duration swap”), it injects liquidity into the financial system.

As for the Fed, it is no longer expanding the money supply via new RMPs, it is still doing $17 billion in reinvestment purchases. This means they are letting longer-term assets (like mortgage-backed securities) roll off their balance sheet and replacing them with short-term T-bills to keep the overall balance sheet size steady rather than actively growing it. Wall Street strategists expect this pause to last through autumn, with a potential restart only if banking liquidity starts to tighten again toward the end of the year.

Also, Secretary Bessent’s intervention alongside Japan to support the yen and his subsequent push to expand the Fed’s Foreign and International Monetary Authorities repo facility would give foreign central banks, including Japan, another way to raise dollars against their Treasury holdings rather than selling those securities into the market. Japan is the largest foreign holder of U.S. government debt.

James E. Thorne went on to say, “The yen intervention was the appetizer. It demonstrated that currencies, liquidity and market structure are now instruments of strategy. The next phase is more consequential: active management of the Treasury curve while constructing the rails for a dollar-centered digital financial order. That is where the Genius Act fits. Properly understood, it is not simply crypto regulation. It is an attempt to bring dollar stablecoins, reserve standards and digital-payment infrastructure within an American legal and financial perimeter…. This is fiscal dominance with a digital dimension. Treasury increasingly sets the incentives; the Federal Reserve reacts within constraints shaped by debt service, market functioning and dollar strategy. Wall Street keeps parsing buybacks, currency intervention, stablecoin legislation and Bitcoin as unrelated events. They are not. Debt is the constraint. Yield management is the response. Digital dollars are the distribution channel.”

Gold has responded to all of this by seemingly having ended its correction and resumed its climb (now back above $4,500), particularly given the nascent softening of the Fed’s hawkishness. Central banks around the word already were quietly buying the dip, with a record 289 metric tons bought during Q2 alone, led by Poland, China, and Czechia, and in fact gold recently surpassed US Treasuries as the largest reserve asset. The intense buying is due to a variety of factors—including as a geopolitical safe haven and as a hedge against debasement of fiat currencies, rising US debt and perceived fiscal irresponsibility, and US weaponization and counterparty risk of bonds or bank deposits (via sanctions)—as part of a broader de-dollarization and diversification strategy. Similarly, bitcoin seems to have caught a bid (now back above $80,000) in a resumption of its previous role as “digital gold”—like it used to do before it started trading along with high-beta stocks.

The “debasement trade” is rooted in the fear that the federal government’s continuing fiscal irresponsibility will erode the value of traditional currencies, especially the U.S. dollar, over the long term. As Phil Rosen of The Opening Bell Daily opined, “The debasement trade seems to be finding new momentum today, but it’s a forever trade that has never gone away.”

And by the way, the massive bond issuance isn’t just from our federal government but also other governments around the world as well as many of our largest corporations—in fact, US corporate bond issuance just eclipsed $1.7 trillion YTD (more than 2025’s full-year total), flooding more supply onto the same buyer pool as government debt, thus driving up yields to attract buyers. As for sovereign borrowing, approximately 78% of new debt is purely for refinancing existing debt, with new debt going toward raised spending on defense, energy security, and aging populations.

No, the bond market isn’t happy about our country’s deteriorating fiscal situation and seems to be signaling a belief that postponing hard choices on spending is no longer an option—perhaps partly driven by the rising tide of socialism and all the growth-stunting, debt-surging tax-and-spend plans its candidates espouse. I liken the debt-and-deficit situation to the reckoning with Iran that President Trump decided was exigent given its 47-year reign as the world’s largest state-sponsor of terrorism coupled with its nuclear weapon/ICBM development. Both have reached a point that could be deferred no longer.

Runaway federal debt and the need to grow our way out of it:

Unfortunately, Republicans—who historically were the party that at least talked about reducing spending and debt (although Democrat President Bill Clinton worked with Republicans to balance the budget in the 1990s)—have seen that the electorate has become enamored with all the “free stuff” promised by other candidates and thus are showing little appetite for budget tightening. So, folks, we get what we vote for, and that means all politicians across parties promise to bring home the booty if you elect them. This of course is the source of our debt spiral. It’s not just the Iran War, which of course has worsened it. It is the combination of growing entitlements (“Nanny State”) and rising interest rates, which causes the deficit to growth such that we must spend more than 50% of the budget deficit on interest payments alone.

We have seen outsized investor focus on inflation and the Fed’s response to it. But while monetary policy gets most of the headlines, the much bigger long-term economic threat is our rapidly growing federal debt, which just hit $40 trillion. Moreover, the deficit is nearly $2 trillion out of a $7 trillion federal budget, which is about 30% of spending and nearly 6% of GDP (compared to a 50-year average of 3.8%).

Of course, the growing deficit must be financed through Treasury issuance, and we face an impending $18 trillion “debt wall” to refinance over the next three years (not to mention another $1.8 trillion in corporate debt and total household debt near an all-time high of $18.8 trillion). Indeed, the federal government is an unsustainable “compounding cycle” in which rising debt and interest rates produce ever-higher interest expense, requiring further borrowing. In fact, more than half of the budget deficit now reflects non-productive interest payments on debt—which means that interest now accounts for more than 3% of GDP.

So, obviously interest rates matter quite a bit. Let’s look at some numbers. Estimates show the average interest rate on federal debt was just 1.70% in 2021 versus 3.44% today. According to the Congressional Budget Office (CBO), gross interest expense has risen from $565 billion in FY2021 to $1.22 trillion in FY2026 on $10.8 trillion in added debt (and growing fast). Moreover, nearly half of GDP is eaten up by the combination of government spending at all levels plus the costs of regulatory compliance.

In my view, the debt problem can be addressed in three ways: 1) cutting away the debt by keeping spending growth below the rate of GDP growth, while rooting out waste and fraud; 2) inflating away the debt through modest inflation of perhaps 2.0–2.5%, which gradually reduces the real value of existing debt; and 3) growing away the debt through robust real GDP growth, given that tax receipts historically average 17% of GDP since 1960 no matter the tax rates (i.e., higher tax rates suppress GDP while lower tax rates raise GDP, so higher GDP generates more tax receipts).

We cannot tax our way out of this problem, nor is drastic austerity or high inflation politically realistic or even economically desirable. And Federal Reserve rate hikes can’t fix the event-driven supply shocks and inflationary pressures, short of inducing recession. In fact, I believe inflation has seen its peak, and elevated Treasury yields likely are driven more by hawkish signals from the Fed and the massive flood of bond issuance (government and corporate) than by actual concerns about entrenched structural inflation. So, growing our way out of debt is both the most important and the hardest approach.

This is where AI becomes much more than just an investment theme. Warsh and Bessent are optimistic about AI improving the outlook for both the economy and inflation, with investment in datacenters, software, and infrastructure raising productivity and increasing non-inflationary growth potential. In the view of Eric Peters, CEO of One River Capital, “The only practical way to forestall [a sovereign debt crisis in the US] is through a productivity boom that materially lifts non-inflationary growth and repays our policy sins. This is why the US government is all in on the AI buildout. If AI fails to deliver on its vast promises, the risk is far wider than the tech sector. It will lead to a deep recession and massive budget deficits, which will catalyze a debt sustainability crisis.”

Indeed, a sustained productivity boom boosts real GDP, raises real wages, widens profit margins, and expands the tax base—without raising tax rates, which would only stunt growth. This is precisely what our indebted nation needs. Like the internet, electrical grid, highways, railroads, and sanitation systems before it, AI is becoming foundational infrastructure that can make life easier, healthier, safer, and more productive. It reduces both physical and cognitive friction and expands our individual and collective capabilities.

Suppose real GDP can sustainably grow at 3–4% annually rather than the 2% pace to which we are accustomed. Add modest inflation, and nominal GDP—and associated federal tax receipts—might expand at 5–6% annually. If Washington can simultaneously keep spending growth well below nominal GDP growth, deficits and the debt-to-GDP ratio (now at an unacceptable 125%) will shrink.

This is why we need policies that encourage and streamline the AI buildout and the energy to power it—the best hope for a productivity boom that energizes non-inflationary economic growth. This is no time for AI backlash and NIMBYism. And we aren’t likely to cut, tax, or inflate our way out of a $40 trillion debt burden. No, our best hope is to grow our way out of it by steadily growing the denominator of the debt-to-GDP ratio faster than the numerator.

As such, I continue to pound the table on: 1) fiscal policies that keep federal spending growth below the rate of GDP growth while pursuing policies that maximize private investment, productivity, and real economic growth; and 2) monetary policies that lower the fed funds rate closer to 3.0% to provide relief to the lower leg of our bifurcated K-shaped economy (small business, housing, lower/middle-income consumers, and the low personal savings rate) while gradually reducing interest expense as federal debt rolls over. While the 10-year yield has greater impact on mortgage rates, the overnight fed funds rate impacts interest rates on short-term debt, credit cards, and variable-rate loans.

Something else the president might do is to organize, incentivize, and mobilize the private sector (i.e., “strategic orchestration” rather than another government-run program) to build affordable single-family homes and perhaps transition some of our struggling college that have shrinking attendance into trade schools that teach the trades and skills needed today and in the future. Both would address and excite the lower leg of the K.

Let me know what you think!

Crude oil inventory drawdown and oil company windfall profits:

According to Goehring & Rozencwajg Associates, refinery runs fell by 5 million bbls/day between January and May, with 2 million bbls/day resulting from China protecting its home market from high oil prices, plus war disruptions in Ukraine and Middle East. And if demand holds steady while refinery runs are down by so much, then product inventories collapse—a price soon confirms it: the burp depressed crude, yet products stayed elevated. The WTI 3-2-1 crack spread (gasoline and diesel premium over WTI crude oil) is averaging around $65/bbl on the US Gulf Coast, compared to an historical norm of $20-30. In addition, the diesel crack spread (premium over crude oil price) has eclipsed $100/bbl for the first time ever. Diesel is the key fuel for the global supply chain, including harvesting of food and how goods are transported.

The US Strategic Petroleum Reserve (SPR) fell to 287 million barrels, its lowest level since 1982, and down from 588 million barrels prior to Russia’s attack on Ukraine. SPR helps stabilize oil markets during times of natural disasters or geopolitical turbulence.  The current drawdown is set to continue until reserves hit just over 240 million barrels, at which time the integrity of the underground salt caverns would start to deteriorate.

Today’s surge in oil company profits resulting from the energy shock—as Iran’s “Tehranical,” theocratic regime reacts with suicidal obstinance to President Trump’s line in the sand regarding nuclear warheads and ICBM capability—has understandably generated accusations of profiteering. I spent the first half of my career in the oil industry, so I am somewhat biased against this narrative, but I also have a gut-level discomfort with the juxtaposition of the industry’s record profits with consumers struggling to pay at the pump.

Of course, there is an important distinction between so-called “windfall profits” and “price gouging.” Oil and refined products trade in global markets, so when geopolitical disruption constrains supply, market prices rise—even though a producer’s cost of extracting an existing barrel may change very little. Likewise, refiners have benefited from exceptionally strong refining margins as overseas supply has been disrupted. Yet with refineries running near maximum capacity (95.6%), and with very little in the way of capacity upgrades on the docket (e.g., coking, hydrocracking, catalytic cracking), there is little evidence that producers or refiners are deliberately withholding supply to manufacture higher prices.

Thus, today’s record profitability is better characterized as an extraordinary geopolitical windfall than systematic consumer exploitation. Gasoline pricing simply follows the formula—for the most part—reflecting global crude prices + refining margin + transportation and distribution costs + taxes + retail margin, which are all market-based. As such, market-based components work in both directions: crude prices and refining margins can soar when supplies are constrained, but they can also collapse during periods of oversupply or weak demand, producing thin margins or even losses.

Nevertheless, free market principles don’t preclude some measure of accommodation when an external geopolitical event dramatically alters those market inputs. But rather than government imposing price controls or a broad windfall-profits tax, which could distort market signals and discourage investment, or expecting competitive companies to get together on their own and agree to collectively cap profits in the spirit of altruism, a reasonable compromise might impose a modest, temporary levy only on profits exceeding an exceptionally high historical profitability threshold, while incentivizing companies to offset it through incremental investment in production, refining, pipelines, storage, and other energy infrastructure (rather than just giving special distributions to shareholders at the expense of consumers). Companies would still retain substantial outsized profits, and importantly, the market-based pricing formula would remain intact.

Any proceeds from this levy should then be returned directly to American households through a temporary “energy dividend” or tax credit rather than disappearing into the vast federal budget. The objective would not be to decide how much profit an oil company “deserves,” nor to penalize an industry for benefiting from market forces that can just as easily move against it. Instead, the policy would preserve the price signals and profit incentives essential to capitalism while recognizing the unusual circumstances of a geopolitical supply shock. Oil companies would retain the bulk of their extraordinary profits if they reinvest them to expand America’s energy security, while temporarily sharing a modest portion of purely windfall profits with the consumers who would otherwise bear the full cost burden of the crisis.

Let me know what you think!

Final comments: Is AI opposition a nuclear-power revolt redux?

As I discussed at length in my August post, the AI backlash/alarmism continues to grow—seemingly in tandem with the rise in socialism, aka “parasites” (in my view), who see any successful capitalist (whether small business owner, founder, entrepreneur, investor, property owner, or corporate leader, i.e., the visionaries, leaders, risk-takers, and builders) as a “host” to suck dry rather than to encourage, admire, learn from, invest in, work for, or seek a business relationship with. The socialist narrative, amplified on social media, is that capitalists get richer at the expense of the common man.

But the truth is, capitalists are not hoarding a fixed “economic pie”; they are growing the “pie” for everyone’s benefit by investing in PP&E, creating jobs, and hiring suppliers and contractors, even if some of the old jobs and suppliers are phased out (“creative destruction”), and even if workers aren’t gaining wealth quite as fast as the entrepreneurs and business owners. Entrepreneur and investor David Sacks enthused, “The AI boom is the catalyst for America’s re-industrialization, which is creating lots of blue-collar and union jobs.” Moreover, our country’s economic growth—and indeed, our continued standing as the world leader in technological innovation—largely depends on winning the AI race.

And yet, so many policymakers seem hell-bent on obstructing this broad, multi-layered, truly all-in, capital-investment cycle—whether due to, at best, a misguided fear of the future, or at worst, a pathological desire to destroy the “tech oligarchy” and/or undermine the Trump Administration, no matter the cost to our country. Scottish author and documentarian Neil Oliver lamented, “We have created an ecosystem that has enabled to thrive the most frightened, psychopathic, parasitic, kakistocratic [the least-qualified, most-unscrupulous] leadership the world has yet seen… We’ve got to take responsibility for the fact that they are our fault. You get the government you deserve.”  And that means we often elect opportunists who play on or pander to our emotions to get elected and then make short-term decisions in reaction to the shifting whims of a fickle electorate in order to retain power, rather than focusing on long-term goals and (sometimes existential) realities.

Renowned Tech futurist/optimist George Gilder’s team opined in The Gilder Report, “The central contradiction is that voters generally support American leadership in advanced technology but oppose the industrial infrastructure required to produce it… That contradiction is faithfully duplicated in government policy. Washington is spending billions of dollars to expand domestic semiconductor production while some lawmakers seek to stop construction of the facilities consuming the most advanced processors, memory, networking devices, and optical components… In a masterpiece of contradiction, New York has offered substantial support for Micron’s proposed semiconductor complex while restricting data-center development. Yet Micron’s high-bandwidth memory and server DRAM are installed in the AI systems housed inside those data centers. The manufacturing project and the data-center projects belong to the same economic chain.”

In my view, this is why our country’s founders explicitly intended to avoid pure democracy in favor of a constitutional representative republic—because voters cannot be trusted to make logical, strategic voting decisions without falling victim to the manipulation and gaslighting of demagogues and mob mentality, as well as “short-termism.” Moreover, demagoguery today is amplified by social media, including professional provocateurs and foreign malign influence campaigns (especially from China, which seeks to win the AI race) to convince a susceptible electorate to vote against their own best interests.

For example, X.com just uncovered a massive Chinese bot farm consisting of roughly 200,000 inauthentic accounts designed to stoke public fear, manipulate public policy, and inflame domestic social tensions around the world—with 200 of them specifically targeting the AI datacenter narrative in the US. These AI bots use generative tools to pump out inflammatory memes, fake news formats, and AI-generated cartoons that depict American tech executives enriching themselves at the public’s expense, with claims about massive AI datacenters straining the American power grid and driving up electricity prices for consumers. They take legitimate discussions and hyper-escalate polarization, with the main objective of slowing down the blistering pace of American AI development.

Turns out, putting Americans at each other’s throats like useful idiots has worked well for our adversaries. And it’s all exacerbated by our elected representatives, who are mostly career politicians with a self-preservation instinct—too afraid of losing their jobs to not shift positions to accommodate the mob’s latest histrionics, thus effectively aiding and abetting these nefarious foreign influence campaigns. If you find yourself reflexively opposed to anything related to AI adoption, datacenter siting, crypto, or President Trump, perhaps you should stop to self-reflect.

Some of our representatives (from both sides of the aisle) have introduced legislation for a federal moratorium on new datacenters until all the impacts are studied and evaluated. But this would do nothing to stop the advancement of AI in China. The Gilder Report asserted, “China would continue constructing computing clusters, electrical generation, semiconductor capacity, and model-development infrastructure. [Senator John] Fetterman summarized the geopolitical consequence directly: ‘America must lead in the development of AI, otherwise we live under China’s rules.’ …The United States can regulate data centers, protect local communities, and require developers to pay their infrastructure costs. It cannot declare a unilateral timeout while China continues building. The fight moving through the midterm elections may therefore determine more than which candidates win in November. It may determine whether America’s semiconductor advantage is converted into lasting AI leadership—or stranded by the inability to build the infrastructure those chips require.”

This is similar to what happened to nuclear power in the US. According to Stephen McBride of the Rational Optimist Society on Substack, “Uranium is ridiculously dense with energy. One gummy bear-sized fuel pellet holds as much energy as a ton of coal or 149 gallons of oil. A golf-ball-sized lump of uranium would cover your entire life’s energy use.” However, in a useful history lesson he explains, “We went from one reactor in 1957 to 112 by 1990. At the peak of the buildout the world was adding about 20 gigawatts of clean nuclear energy every single year. In 1966 the New York Times previewed the 21st century: a world without deserts, without smog, billions living ‘in unprecedented luxury thanks to nuclear power.’ …The atom was going to desalinate oceans, green deserts and power flying cars. Then the villains arrived. Bureaucrats. Regulations pushed up the cost of atomic energy ninefold. Utilities cancelled more than 100 planned reactors. And for the next 30 years America didn’t greenlight a single new plant… [as onerous regulations] essentially made building new power plants in America illegal.”

As a result, the US ceded nuclear energy dominance to China.

Will the same thing happen with AI? By similarly framing AI infrastructure as an environmental and economic threat, foreign influence operations successfully weaponize legitimate domestic discourse to achieve a geopolitical outcome: stalling or even halting (in the case of nuclear) our rapid technological progress.

Perhaps SpaceX teaming up with NVIDIA on Project Starmind for orbital datacenters offers a ray of hope. This concept has quickly moved from mere “pie in the sky” to “pie in the oven.” Employing the Space-1 module (Vera Rubin NVL72 rack-scale platform, with Vera CPUs and Rubin GPUs), its AI1 satellite using Starlink V3 bus technology is designed to handle an average compute payload of 120 kW with laser links through Starlink. The massive 11-million-square-foot (sf) SpaceX Gigasat factory in Texas is purpose-built to mass-manufacture thousands of these specialized AI satellites, with initial commercial constellation deployment targeted for late 2027, assuming prototype testing goes well. Other companies that might play a role in orbital datacenters include Redwire (RDW) and Planet Labs (PL).
 
Or perhaps it will be ocean-borne, self-propelled, untethered, floating steel datacenters like those proposed by privately owned Panthalassa, powered solely by continuous wave energy with natural cooling. Or perhaps Microsoft’s (MSFT) concept of a subsea, watertight, nitrogen-atmosphere pod.

In addition, Elon Musk envisions a new 100-million-sf “Terafab” in Texas for making advanced microchips. In effect, he is proposing a complete, vertically integrated, alternative industrial infrastructure for AI compute, extending from advanced semiconductor manufacturing on Earth to solar-powered computing in orbit. The vision is audacious.

More broadly, it seems that no matter what technological challenge arises, there is never a shortage of American entrepreneurial innovators and private capital coming up with a plan to resolve it—and it’s not just Elon Musk. Check out this article by the Rational Optimist Society about the many companies doing truly awe-inspiring things in the new space race. These are the visionaries, risk-takers, builders, and doers that the socialists want to demonize and tax, aided and abetted by our adversaries.

So, let’s reaffirm and embrace American exceptionalism, entrepreneurship, meritocracy, property rights, rule of law, and our rugged, resourceful, can-do spirit—and reject all calls for it to be devalued, cast aside, or wither away in deference to the newly fashionable “warmth [or should I say shackles] of collectivism,” in which compliance must be coerced because it cannot be maintained for long by free choice.

Gain access to Sabrient’s proven models:

You can gain access to several of Sabrient’s proven proprietary models and scores for all stocks in our system through our Sabrient Scorecard subscription. It’s an easy-to-use investor tool in multi-tab Excel format, downloadable weekly with fresh scores, to make the process easy for idea generation and portfolio monitoring. It includes our Top 30 stocks each week for 4 distinct investing strategies—Growth, Value, Dividend, and Small Cap—or you can input your own list of stocks. I welcome you to learn more by visiting: https://www.moonrockstopowerstocks.com/sabrient-scorecard

Also, Sabrient’s process-driven, growth-at-a-reasonable-price methodology is revealed by Sabrient founder and former NASA engineer (Apollo Program) David Brown in his book, Moon Rocks to Power Stocks: Proven Stock Picking Method Revealed by NASA Scientist Turned Portfolio Manager. It teaches how to methodically and strategically build wealth in the stock market for those four distinct investing styles. You can download the book and two Bonus Reports (on the history and new opportunities for 1) Energy and 2) Space Exploration—all in PDF format. [Note: David’s book is also available in both paperback and Kindle formats on Amazon.]

Latest Sector Rankings

Relative sector rankings are based on Sabrient’s proprietary SectorCast model, which builds a composite profile for each of over 1,500 equity ETFs based on bottom-up aggregate scoring of the constituent stocks. The Outlook Score is a Growth at a Reasonable Price (GARP) model that employs a forward-looking, fundamentals-based multifactor algorithm considering forward valuation, historical and projected earnings growth, the dynamics of Wall Street analysts’ consensus earnings estimates and recent revisions (up or down), quality and sustainability of reported earnings, and various return ratios. It helps us predict relative performance over the next 3-6 months.

In addition, SectorCast computes a Bull Score and Bear Score for each ETF based on recent price behavior of the constituent stocks on particularly strong and weak market days. A high Bull score indicates that stocks held by the ETF recently have tended toward relative outperformance when the market is strong, while a high Bear score indicates that stocks within the ETF have tended to hold up relatively well (i.e., safe havens) when the market is weak. Outlook score is forward-looking while Bull and Bear are backward-looking.

As a group, these three scores can be helpful for positioning a portfolio for a given set of anticipated market conditions. Of course, each ETF holds a unique portfolio of stocks and position weights, so the sectors represented will score differently depending upon which set of ETFs is used. We use the iShares that represent the ten major U.S. business sectors: Financials (IYF), Technology (IYW), Industrials (IYJ), Healthcare (IYH), Consumer Staples (IYK), Consumer Discretionary (IYC), Energy (IYE), Basic Materials (IYM), Telecommunications (IYZ), and Utilities (IDU). Whereas the Select Sector SPDRs only contain stocks from the S&P 500 large cap index, I prefer the iShares for their larger universe and broader diversity.

The table below shows the latest fundamentals-based Outlook rankings and our full sector rotation model:

The latest rankings display a bullish bias, in my view, given that cyclicals and secular growth sectors dominate the top of rankings, and defensive sectors are at the bottom. Next 12 months (NTM) analyst earnings forecasts still look quite strong across the board, and investors seem optimistic that impacts from the Iran conflict and shipping blockade are either shrinking or effectively worked around. All but Utilities have seen recent upgrades to forward estimates.

Technology (dominated by the mega-cap Big Tech titans and AI-driven highflyers) remains firmly at the top with an Outlook score of 86. Its continued strength reflects the high quality of those juggernaut cash machines despite having the highest forward P/E at 25.8x, which remains well below the 31x it reached last fall as multiples contract—and it continues to fall. The consensus NTM EPS growth estimate of +28.1% has risen faster than price such that the forward PEG (ratio of P/E to EPS growth) for Tech is a modest 0.92 (the lowest of all sectors). Tech also displays strongly positive sell-side analyst earnings revisions and the highest profit margins, return ratios, and insider buying. Although the market is broadening, investors continue to be willing to pay up for strong, reliable growth.

Because many Tech stocks are riding secular growth trends (i.e., little cyclicality), no other sector comes close to the consistent sales growth, margins, operating leverage, and ROI. And Tech not only benefits from its own product development and productivity gains, but those products help companies in other sectors with their product development, product delivery, and productivity. So, Tech benefits by helping all sectors grow and prosper. The main concern is whether spending of over 100% of hyperscalers’ massive cash flow on capex will generate sufficient ROI in the near term to placate investors.

Others with a low forward PEG include Telecom (1.12), Energy (1.02), Financials (1.07), and Basic Materials (1.17). As for forward P/E, Energy (13.0x) and Financials (14.5x) are by far the lowest, which is reflective of their cyclicality.

After Tech, the rest of the top 6 sectors by Outlook score are Telecom, Energy, Industrials, Financials, and Basic Materials. Firmly at the bottom of the rankings are noncyclical defensive sectors Utilities and Consumer Staples, which have the lowest forward EPS growth rates (the only ones in single digits), among the lowest consensus analyst revisions to EPS estimates, and very little insider buying.

Keep in mind, the Outlook Rank does not include timing, momentum, or relative strength factors, but rather reflects the consensus fundamental expectations at a given point in time for individual stocks, aggregated by sector.

To learn more about how you can access our weekly Stock and ETF Scorecards, please visit:
https://www.moonrockstopowerstocks.com/sabrient-scorecard

Sector Rotation Model and ETF Ideas

Our rules-based Sector Rotation model, which appropriately weights Outlook, Bull, and Bear scores in accordance with the overall market’s prevailing trend (bullish, neutral, or defensive), returned to a bullish bias when the S&P 500 leapt back above its 50-day moving average last Monday. (Note: In this model, we consider the bias to be bullish from a rules-based trend-following standpoint when SPY is above both its 50-day and 200-day simple moving averages, but neutral if it is between those SMAs while searching for direction, and defensive if below both SMAs.)

As highlighted in the table above, the Sector Rotation model suggests holding Technology (IYW), Telecom (IYZ), and Financials (IYF), in that order. Or, if you prefer to take a neutral stance, it suggests holding Technology, Telecom, and Energy (IYE). However, if you prefer a defensive stance, it suggests holding Energy, Healthcare (IYH), and Consumer Staples (IYK).

Here is an assortment of other interesting ETFs that are scoring well in our latest rankings: Inspire Fidelis Multi Factor (FDLS), VistaShares Artificial Intelligence Supercycle (AIS), Xtrackers Semiconductor Select Equity (CHPS), iShares Technology Opportunities Active (TEK), Zacks Small/Mid Cap (SMIZ), Guinness Atkinson Smart Transportation & Tech (MOTO), First Trust DBA Deregulation (DGLO), iShares Future AI & Tech (ARTY), Lazard Next Gen Technologies (TEKY), Aztlan Global DM SMID (AZTD), Franklin Intelligent Machines (IQM), First Trust Active Factor Small Cap (AFSM), Monarch Volume Factor Dividend Tree (MVFD), Global X AI Semiconductor & Quantum (CHPX), NestYield Visionary (EGGQ), JPMorgan Active Developing Markets Equity (JADE), Pacer Data Digital Revolution (TRFK), Roundhill Generative AI & Technology (CHAT), and Touchtone Sands Capital US Select Growth (TSEL). All score in the top 5% (95-100) of Sabrient’s Outlook scores.

As always, I welcome your thoughts on this article! Please email me anytime. Any and all feedback is appreciated.

Also, please let me know of your interest in any of Sabrient’s new indexes for ETF investing, such as High-Quality Energy, High-Quality Healthcare, Defensive Equity, High-Quality SMID Growth, High-Quality Growth & Income, and High-Quality Value, as well as the actively managed Space Exploration & Off-Earth Sustainability, Future Energy, and Sabrient Select High-Conviction Portfolio (similar to our Baker’s Dozen portfolio, but larger). Visit Sabrient.com for more information on the six passive indexes.

IMPORTANT NOTE: I post this information periodically as a free look inside some of our institutional research and as a source of some trading ideas for your own further investigation. It is not intended to be traded directly as a rules-based strategy in a real money portfolio. I am simply showing what a sector rotation model might suggest if a given portfolio was due for a rebalance, and I do not update the information on a regular schedule or on technical triggers. There are many ways for a client to trade such a strategy, including monthly or quarterly rebalancing, perhaps with interim adjustments to the bullish/neutral/defensive bias when warranted, but not necessarily on the days that I happen to post this article. The enhanced strategy seeks higher returns by employing individual stocks (or stock options) that are also highly ranked, but this introduces greater risks and volatility. I do not track performance of the ideas mentioned here as a managed portfolio.

Disclosure: At the time of this writing, of the securities mentioned, the author held positions in DIA, QQQ, FTLS, SPCX, GEV, MRVL.

Disclaimer: Opinions expressed are the author’s alone and do not necessarily reflect the views of Sabrient. This newsletter is published solely for informational purposes only. It is neither a solicitation to buy nor an offer to sell securities. It is not intended as investment advice and should not be used as the basis for any investment decision. Individuals should consider their personal financial circumstances in acting on any opinions, commentary, rankings, or stock selections provided by Sabrient Systems. Sabrient makes no representation that the techniques used in its rankings or analyses will result in profits. Trading involves risk, including possible loss of principal and other losses, and past performance is no guarantee of future results. Investment returns will fluctuate, and principal value may either rise or fall. Sabrient disclaims liability for damages of any sort (including lost profits) arising from the use of or inability to use its rankings or analyses. Information contained herein reflects our judgment or interpretation at the time of publication and is subject to change without notice.

 

Copyright © 2026 Sabrient Systems, LLC. All rights reserved.