Where the Rotation Went

The Setup

Two forces are driving this rotation, and both have names. The first is the chip trade cracking: the AI-semiconductor complex that carried the first half of the year rolled over hard through July, a slide punctuated by Intel's report yesterday, which we covered in Thursday's Avoid Thesis. The second is Kevin Warsh. The new Fed chair's June meeting held rates at 3.50% to 3.75% but delivered the hawkish shock of the year: a 130-word statement stripped of its easing bias, nine of eighteen officials penciling in a 2026 rate hike, and a two-year yield that jumped the most on a Fed day since 2008. Bank of America's economists now forecast three hikes this year. As of yesterday, market odds of at least one 2026 hike sit near 72%, with the next meeting five days away. Higher-for-longer rates are a tax on richly valued growth stocks and a tailwind for lenders' interest margins, and capital has responded by moving from one to the other.

The price evidence, measured from the semiconductor index peak in late June through yesterday's close: chips -10.3%, the Nasdaq 100 -4.4%, the S&P 500 -0.4%, financials +3.9%, and industrials, utilities, and healthcare all within half a percent of flat. One sector caught the outflow. But the detail that shaped this list is inside that winning sector: the regional bank index gained only 0.5%. The rotation found the giant financials and stopped, even though the regionals just reported some of the best numbers of the season. That gap between where the money went and where the data points is the story of this edition.

The screen: we pulled current platform scores for 50 names across two groups, the financials that are catching the rotation, and the AI-power industrials, the transformer, switchgear, and electrical equipment makers, that many expect to catch it next. Standing exclusions applied. Names whose calculated upside exceeds our 80 to 100 percent reliability band came out, which removed several apparent bargains. Insurance carriers required special handling, because the platform's valuation models systematically overstate insurer discounts through their treatment of float, so all but one came out, and the survivor carries that caveat in writing. What remained is five names, every one carrying a Bullish outlook in eight consecutive readings, presented from largest to smallest discount against the tape that has and has not found them.

JPMorgan Chase (JPM)

Start with the anchor. JPMorgan Chase (JPM) opened the earnings season on July 14 with a quarter that will be in the textbooks: net income of $21.2 billion, reported as the largest quarterly profit in American banking history, on revenue of $58.0 billion against expectations near $50.2 billion, one of the biggest beats the company has ever printed. Equities trading revenue rose 86% to $6 billion. The quarter also carried a one-time flavor of the moment: June's $86 billion SpaceX offering, the largest IPO ever priced, handed underwriting fees to nearly every major bank, with JPMorgan among the primary co-underwriters, part of a sector-wide investment banking surge.

The platform's read is what makes the record quarter a list entry rather than a headline: Company Strength of 68, Strong, with a Moat of 10/15, and a fair value of $389.51 that sits 11% above the current price near $350, with a Bullish outlook in eight straight readings. An 11% discount is the smallest gap on this list, and that is the point of putting it first. This is the quality anchor: the fortress balance sheet, the diversified fee machine, the bank that just demonstrated what this rate and capital-markets environment does for the industry's best operator. The sector trades near 12 times earnings against roughly 22 times for the S&P 500, and JPMorgan is the reason that multiple gap is hard to dismiss as a value trap.

What the data cannot do is make the discount large. The market has never been confused about what JPMorgan is, and even after lagging the sector rally earlier in the year, it is not being given away. The honest description is a great business at a modest discount in a regime built for it: hawkish rates supporting interest income, capital markets running hot, and credit still quiet. On a list that gets more speculative as it goes, this is the entry that requires the fewest things to go right.

Bank of America (BAC)

Bank of America (BAC) is the accountability entry, because we wrote it down before the report. In our July 10 Earnings Season Watchlist, published four days before the print, the platform scored it a Moderate 50 at a fair value of $72.37, Undervalued at $59.25, with the note that the bank had made a new high and still had not caught up to its own earnings power. Then it reported: net interest income of $16.2 billion, up 9%, a ninth consecutive earnings beat in a quarter where all five major banks cleared expectations, and the customary post-stress-test capital returns on deck.

Here is what happened to the scores, because this is exactly the conversion process the watchlist franchise exists to track. Since the report, Company Strength has climbed from 50 to 59. Fair value has risen from $72.37 to $78.59, an 8.6% increase. The stock, meanwhile, has advanced about 3% to $61.28. Read those three numbers together: the price went up, the fair value went up faster, and the discount widened from 22% to 28%. The market gave Bank of America a polite nod for a record-adjacent quarter; the models, fed the new financials, concluded the business is worth meaningfully more than they thought two weeks ago. The Bullish streak is now eight for eight.

The soft spot in the data is the one we flagged in the watchlist and will not hide now: Financial Health scores just 47/100, the weakest on this list, a reflection of the leverage inherent to a balance sheet this size and a reminder that big banks earn their discounts partly through opacity. What has changed since July 10 is that the quarter arrived, the credit line stayed quiet, and the scoring system moved toward the stock rather than away from it. When the evidence updates in your direction, the position of this publication is to say so with the same volume we used going in.

M&T Bank (MTB)

Now the part of the trade the money skipped. M&T Bank (MTB), the Buffalo-based Eastern-seaboard lender, reported on July 15 the highest quarterly earnings per share in its history: $5.32 against expectations of $4.66, up 25% from a year ago, its fourth consecutive beat, on net income of $818 million. The engine was exactly what a hawkish-Fed regime is supposed to produce: net interest income up 4.6% to $1.79 billion, net interest margin expanding to 3.70% from 3.62%, and average loans growing to $141.4 billion, driven by a $5 billion increase in commercial and industrial lending. The bank bought back $465 million of stock in the quarter and guides full-year net interest income to as much as $7.35 billion.

The platform's response is the largest clean discount our rotation scan produced: Company Strength of 66, Strong, with the best Moat among the banks here at 11/15, and a fair value of $371.14 that sits 51% above the price near $246, Deep Value, with a Bullish outlook eight for eight. And unlike the deep discounts our energy scan surfaced last week, this one is not riding a commodity spike. It is built on a lender printing record earnings, at a record-adjacent share price, that the valuation models still cannot reconcile with what the business earns.

Which returns us to the strange fact from the setup: the regional bank index is up 0.5% in the four weeks financials gained 3.9%. The rotation, so far, is an index-level trade, flowing to the mega-caps through ETFs and futures while the individual regionals posting the actual record quarters sit still. Commercial lending is the regionals' disproportionate business, and commercial lending, including the loans financing the data center construction boom, is precisely what is accelerating. The data cannot tell you when the money makes that second stop. It can tell you the discount waiting there is twice the size of the one the crowd already took.

Citizens Financial (CFG)

Citizens Financial (CFG) is the improvement story of the regional group. The Providence-based lender reported second-quarter earnings of $1.30 per share, up 15% sequentially and 41% from a year ago, on net interest income growth of 14% and more than 600 basis points of positive operating leverage, meaning revenue is growing dramatically faster than expenses. Margin expanded again and management guided to further gains, with a stated path to a 16% to 18% return on tangible common equity by the end of 2027, an efficiency ratio headed from 61% toward the mid-50s, and $920 million returned to shareholders so far this year. Heading into the season, Citizens was one of the two regionals leading the group, up more than 20% on improving loan growth and steady credit.

The platform scores the trajectory: Company Strength of 67, Strong, actually a shade above M&T, with a Growth score of 13/15, a fair value of $91.80 that leaves the price 28% below, and the now-familiar eight for eight Bullish streak. The discount matches Bank of America's, but the profile is different: this is a smaller balance sheet earning its rerating through self-help, expense programs, a growing private bank, and margin repair, layered on top of the same rate tailwind lifting the whole industry. Forty-one percent earnings growth at eleven-ish times earnings is the kind of arithmetic that eventually attracts a crowd.

The risk profile is also different, and worth stating. Turnaround math depends on the turnaround: the operating-leverage story requires expenses to stay disciplined and the 2027 return targets to survive contact with whatever the Fed does next week and beyond. A regional bank promising expanding margins is implicitly forecasting a rate environment it does not control. The scores grade what has been reported, and what has been reported is four quarters of acceleration. The market, holding the regional index flat, is effectively saying it wants to see a fifth. That skepticism is the discount.

The Hartford (HIG)

The Hartford (HIG) finishes the list with the highest composite score our entire 50-name rotation scan produced, and an asterisk we will spend a full paragraph on. First the business: a 200-plus-year-old property and casualty insurer focused on small commercial and middle-market business insurance, group benefits, and personal lines, a franchise that earns underwriting profit on premiums and investment income on the float those premiums create. Rising rates are a direct tailwind to that second engine, a tailwind that arrives without the credit risk a lender must carry to capture the same rates. The platform scores the company Strong at 73, the best on this list, with Financial Health of 77/100, also the best here by thirty points over Bank of America, a Moat of 11/15, and a Bullish outlook in eight straight readings.

Now the asterisk, stated as plainly as we can because it is a standing rule of this publication: the platform's valuation models systematically overstate insurance discounts. The way float accounting flows through book value and earnings power calculations produces Deep Value labels across the entire carrier group, and today's scan was a caricature of the pattern, with several insurers printing calculated upside of 100% to 160% that we excluded outright as artifacts. The Hartford's stated fair value of $245.10, a 72% gap against the price near $142, sits below our hard exclusion band but well inside the zone where the float distortion inflates the number. Treat the direction as meaningful and the magnitude as unreliable. On conventional measures, the stock trades at an undemanding multiple of earnings that have compounded steadily for years; that, not $245, is the defensible claim.

Why include a name that needs a warning label? Because the non-valuation evidence is the strongest in the scan, and it does not depend on the distorted model: the best strength score, the best health score, a Bullish streak matching every bank above, and a business whose economics improve mechanically as the Fed stays hawkish, without taking the credit risk a lender takes to get the same exposure. The Hartford is on this list for what the scores can measure cleanly. The fair value figure is printed because we print what the platform says, and flagged because we tell you when not to take it literally.

What the Wealth Engine Scores Say

Here is the platform's systematic read on all five, current as of Thursday. Every outlook below has read Bullish in eight consecutive readings.

The pattern across the boxes: four Strong ratings and one Moderate, five Bullish streaks at eight for eight, and discounts that grow as the market capitalization shrinks, from 11% at JPMorgan to 51% at M&T. That gradient is the article's thesis in numeric form. The rotation has been an index-level trade, and index-level trades reach the biggest names first, which is why the smallest discount on this list belongs to the stock that has already been found and the largest clean discount belongs to a regional bank the rotation has not touched.

These scores are systematic. They evaluate companies on reported financials, balance sheet quality, moat characteristics, and valuation models including discounted cash flow, peer comparison, and earnings power. They measure what a company is today, not what it might become. For lenders, that carries a specific nuance: bank earnings are a function of a rate regime the banks do not set, and the fair values above are computed from results the current hawkish regime produced. They are honest photographs of a favorable environment.

And one methodological note in the interest of full transparency: these models have known blind spots by asset class, and we route around them in public. Insurance fair values run hot, so The Hartford's number carries a written warning. Several insurers and consumer lenders screened even higher and were excluded entirely because their figures crossed our reliability band. The scores you see are the ones we are willing to stand behind, and the ones we are not, we name in the next section.

What Didn't Make the List

The most important miss is a whole thesis. If the rotation's first leg is financials, the widely expected second leg is the AI-power industrials, the transformer, switchgear, and electrical equipment makers wired into the data center buildout. We scanned them alongside the financials, and the data rejected the leg twice over. In price: industrials are flat since late June while financials gained 3.9%, so the money has not moved. And in value: every AI-power name screened Expensive with a Neutral or worse outlook, and the best of them, Hubbell, trades roughly 51% above its calculated fair value. The buildout is real and the platform is not blind to it; the scores simply say the sector's prices absorbed that future years ago. A rotation into those names at these levels would be capital leaving one crowded trade for another.

Inside financials, the exclusions tell their own story. Allstate, MetLife, and AIG screened with calculated upside of 107% to 160%, textbook cases of the insurance float artifact, and came out on the reliability rule, the same standard that removed Super Micro from our earnings watchlist and EQT from last week's energy scan. Ally and Synchrony printed 95% and 90% and followed them out. The exchanges are the mirror image: CBOE carries the scan's only Elite strength rating at 81, and CME and Nasdaq both score Strong, but all three trade 33% to 51% above fair value. Quality, fully priced.

The nearest misses were banks, fittingly. Regions Financial screened Bullish eight for eight at a 36% discount and was the last cut; Truist cleared the same bar at 26%. Wells Fargo looks cheap at a 21% discount but is the one major money-center bank whose outlook reads Neutral rather than Bullish, eight readings running, and on a list defined by aligned signals, that misalignment decided it. Anyone extending this list past five starts there, with Regions, and the platform will show you exactly what we saw.

What Could Go Wrong

Start with the engine of the whole trade: the hawkish Fed cuts both ways. Higher rates widen lending margins right up until they break something, and a Warsh Fed hiking into a slowing economy is precisely the scenario that turns bank tailwinds into a credit cycle. The provision lines were quiet this quarter and the private credit worries of the spring have gone dormant without new failures surfacing, but every bank discount on this list is implicitly a bet that quiet continues. The next test is five days away: the FOMC meets Wednesday, and both outcomes carry risk for this list, a hike that accelerates the margin story while raising recession odds, or a dovish surprise that unwinds the very rotation this article maps. We claim no edge on that meeting; we note that the trade's foundation gets re-poured next week.

The regional thesis carries its own specific hazards. The market holding the regional index flat while these banks print records is not necessarily irrational: regionals carry concentrated commercial real estate exposure the mega-caps dilute, their deposit costs are stickier, and the last two times investors chased regional bank value, 2023 taught the lesson about how fast that cohort reprices when confidence cracks. M&T's 51% discount and Citizens' operating-leverage story both assume the commercial lending acceleration is durable rather than a data-center-construction sugar high. And the rotation itself could simply round-trip: if the chip complex stabilizes after Intel's report and the AI trade reasserts, four weeks of financial sector inflows can leave as an index trade just as fast as they arrived.

Finally, the honest caveats about our own instruments. The Hartford's headline discount is inflated by a known model artifact, which we have labeled, and insurance underwriting carries catastrophe risk no financial screen captures; one bad hurricane season rewrites the group's earnings. Bank of America's 47 health score is the list's reminder that leverage is the sector's raw material. And all five names share a single macro sensitivity, which makes this the least diversified Five we have published: it is one thesis, expressed five ways, and if the rate regime that powers it reverses, the names will not take turns being wrong. They will be wrong together.

The Bottom Line

Follow the money and it leads somewhere specific: out of the chip complex that broke in July, into the financial sector, and, so far, no further than the names an index fund reaches. The platform's scan of the rotation finds the trade roughly one-third complete by its own logic. The mega-caps that caught the first flows now trade at modest discounts, 11% at JPMorgan, 28% at a Bank of America whose fair value rose faster than its price after a record-season quarter. The regionals posting the actual record earnings, M&T at a 51% discount and Citizens at 28% with 41% earnings growth, have caught nothing at all. And the widely assumed second leg into AI-power industrials fails both tests the data can run, with no price movement and no value on offer.

None of this is a forecast that the money finishes the trip. Rotations stall, reverse, and round-trip, and Wednesday's Fed meeting could re-price the entire premise in an afternoon. What the data offers instead is a map drawn before the outcome: where the flows have been, where the discounts are, and which parts of the consensus rotation story the numbers support and reject. The industrials leg, rejected. The insurance discounts, partly artifact, labeled as such. The regional bank gap, real, reported, and unclaimed.

That is the Wealth Engine Pro approach: measure the rotation instead of narrating it, publish the scores with their blind spots labeled, and leave a record specific enough to be graded. The watchlist entry this list carries forward, Bank of America, already got its first grade when the July 14 print moved the model in its favor. The other four now have their numbers on the record too. The rotation went where the index took it. The data says the value is one stop further down the line.