Markets Desk

Sightline Markets Daily Miles Cardell & Jenna Vega

The September CPI print is the organizing fact of the session. Headline at +3.74% YoY (index 336.94, MoM +0.59%) against core CPI at +2.57% YoY — core still running below headline, but the gap is narrowing in the wrong direction. For context: headline CPI averaged roughly 2.1% in the decade pre-COVID; post-shock, it peaked above 9%. At 3.74%, we are not in crisis, but we are also not in the Fed's living room. The MoM acceleration of +0.59% is the twitchiest tranche here — that annualizes well above the 2% target and will likely keep the Fed on hold or worse.

The equity damage showed up in the picks-and-shovels layer of the AI trade. Bitcoin miners — TeraWulf, Core Scientific, Cipher Mining, IREN, each down 12%+ — are not pure BTC plays; they are leveraged capital-cost stories. Higher rates crush capital-intensive build-outs. That is muscle memory from 2022, and the market ran it again today. Our usual cross-check on whether this is a sector-specific reprice or something broader: the Seoul equity open was also sharply lower, suggesting global contagion, not just domestic repositioning.

Alpha Vantage returned no valid quotes this run, so we cannot anchor on SPY/QQQ levels today. What we can say: VIX at 15.27 is down 1.93 pts over 30 days, which is inconsistent with a sustained fear regime. The HY OAS at 316 bps, +40 bps over 30 days, is the more interesting signal — not alarming in isolation (the long-run IG BBB OAS sits at 103 bps, HY-IG gap at 213 bps, both within calm-regime bounds), but the direction of that 30-day move bears watching. ICI flow data corroborates the rotation: total long-term fund outflows of $55.3 billion in the latest week, with domestic equity alone shedding $31.6 billion. Money markets took in $8.0 billion. Smart money is not panicking; retail is quietly walking toward the exit.

Berkshire's latest 13F (as of 2026-06-30) shows a $1 million pilot position in D.R. Horton — precisely as DHI led all homebuilders with 67.7% Item 1A novelty in its most recent 10-K risk rewrite. That is an interesting coincidence. Whether it is a conviction buy or a placeholder, we note it and move on.

September CPI at +3.74% YoY and +0.59% MoM reset Fed path expectations, flattening the capital-intensive trade — miners, AI infrastructure, rate-sensitive growth — while retail equity outflows hit $31.6 billion and VIX remains contained at 15.27.

Coiner's Credit Review August Farris & Ezra Farris

The September print arrived and the market marveled, briefly, at its own complacency. Headline CPI at +3.74% YoY (index 336.94) with MoM of +0.59%; sticky core at +2.78% per FRED's Atlanta Fed measure. The effective fed funds rate sits at 3.88%. Real short rates are therefore barely positive at the headline level — about 14 basis points of carry above CPI — which is not a posture from which a central bank declares victory over inflation. The FOMC has been here before: cut prematurely in 1967, watched the second wave arrive, and spent a decade relearning the lesson. We are not predicting a replay; we are noting that the room for error is smaller than the equity tape's VIX of 15.27 implies.

The Bank of America bond portfolio story — paper losses reportedly surpassing $100 billion — is the kind of detail that the market has been encouraged to treat as an accounting inconvenience since FASB accommodated HTM classification. It is not an accounting inconvenience. It is a funding-cost problem waiting for a deposit-outflow catalyst. We groused about this during the SVB episode and were told the HTM portfolio was 'different.' The coupon on those legacy positions is still a fixed number. The funding cost is not. That gap, annualized and multiplied by a $100 billion notional, is a real transfer of value — from BofA's spread income to the money-market funds now offering 5%+. BAC's 38.9% Item 1A novelty score and 671 new sentences added to JPM's risk factors versus 247 deleted suggest the lawyers are earning their keep.

The 10Y-2Y curve at 0.37pp — positive, but barely — is the yield curve's version of a convalescent patient insisting they feel fine. Long-run average for this spread runs closer to 1.0-1.5pp in a neutral-rate environment. At 37 bps, the curve is not pricing a soft landing so much as a sidelong glance at one. Lodestar will tell you trend-followers started rotating back into fixed income when the curve normalized from inversion; that rotation is now structurally vulnerable to any CPI re-acceleration.

Real fed funds barely positive at ~14 bps above headline CPI, a historically thin cushion; BofA's $100 billion-plus in HTM paper losses represent a real funding-cost transfer, not an accounting abstraction, and the 37 bps yield curve normalization is fragile against any CPI re-acceleration.

Bias flag — Structurally skeptical of monetary expansion; right on major breaks, early/wrong through long bull phases. The BofA HTM loss framing is real but has been a recurring warning since 2022 without triggering the predicted credit event.

Alder Grove Memos Victor Halprin

I want to be careful here, because the CPI print is the kind of event that looks decisive in real time and then turns out to be one data point. But the pattern surrounding it matters more than the number itself. We have: a September CPI at +3.74% YoY that came in above what the tape had priced; an unemployment rate at 4.2% (MoM +2.44 percentage points, which is a striking revision magnitude); average hourly earnings at +3.02% YoY. That last number is the pendulum's tell. Wage growth at 3.02% against a CPI of 3.74% means real wages are still negative — roughly minus 72 basis points in real purchasing power. That is not the condition under which consumer sentiment typically expands.

Here is my actual bottom line: the two possibilities I see are these. Either today's selloff is a repricing event — the market recalibrates Fed path, finds a new clearing level, and the underlying earnings cycle continues — or it is the first visible crack in a thesis that has been load-bearing for two years, namely that AI capital spending would compound without regard to interest rates. The bitcoin miner selloff is not a Bitcoin story; Ledger Lines can confirm BTC at $80,981 with a 30-day Sharpe of 2.36 looks genuinely constructive on-chain. The miner story is a cost-of-capital story. And cost-of-capital stories, once they start, rarely end in one session.

I admit I cannot tell you which of those two paths we are on. What I can tell you is that the ICI data — $55.3 billion out of long-term funds, $8.0 billion into money markets in a single week — shows that the pendulum of investor psychology has not swung to panic. It has swung to caution. Caution is a healthy intermediate state. It becomes dangerous only if it accelerates into flight, and the VIX at 15.27 says we are not there yet. I am watching the next two CPI prints more closely than anything else on the calendar.

The pendulum has moved from complacency to caution — not yet panic — and the CPI-driven selloff in capital-intensive AI plays is best understood as a cost-of-capital repricing, not an earnings cycle break, though the distinction may not hold if inflation reaccelerates.

Kensington Macro Letter Nora Kensington

I've been writing about the Fiscal Dominance problem for long enough that I almost feel obligated to say 'I told you so' — but I won't, because the more interesting question is what happens next. September CPI at +3.74% YoY is not a surprise to anyone who has been watching the structural drivers: the Iran war's upward pressure on energy costs (flagged explicitly by the IMF chief at the Bangkok meetings), the Paducah AI hub — $100 billion in private investment going into a former Cold War nuclear site — and the broader nominal GDP imperative that forces the Treasury to keep the debt-to-GDP ratio from spiraling by inflating the denominator. Real GDP Q2 2026 came in at +2.2% SAAR, down from +2.5% in Q1. Nominal GDP is running faster. That is what Fiscal Dominance looks like in the data: real growth slows, nominal growth is sustained by the price level.

I keep coming back to the IMF chief's statement in Bangkok. The Fund is urging the U.S. and wealthy nations to rein in debt to ease global economic pressure. The IMF is correct about the diagnosis and politically unable to enforce the prescription. What the U.S. government can do — and what I have argued is essentially the only politically viable path — is inflate the debt away gradually. My Three-Axis framework would classify this as a Group B asset environment for hard assets, a Group A headwind for long-duration nominals. Gold below $4,100 today is a hot-CPI reflex trade, not a structural reversal. The dollar's 30-day surge of +2.72 to a broad index of 121.60 is the mirror image: dollar-denominated rate expectations pulling in capital. That is a Drip Print dynamic, not a Tidal Print. Slower than people think, then faster than people think.

I want to flag the WTI-Brent spread explicitly: WTI at $96.24 versus Brent at $125.44 is a $29.20 gap. That is extraordinarily wide. The corpus mentions Middle East tensions — Iranian-linked dhow destroyed in the Gulf of Oman, Houthi activity near Saudi Arabia — but U.S. domestic crude is sitting $30 below the global benchmark. That spread speaks to either a structural bottleneck in U.S. export capacity or a localized supply surge that is not yet reflected in global pricing. Thicket will have a more precise read on the geo-commodity mechanics.

September CPI at +3.74% YoY with real GDP decelerating to +2.2% SAAR in Q2 is the fiscal-dominance signature: nominal growth sustained by price level as real growth fades, confirming a structural Group B hard-asset environment despite today's gold and silver reflex selloff.

Bias flag — Hard-asset constructive and fiscal-dominance lens can over-index to inflationary tails during disinflation windows; today's framing may underweight the disinflationary signal in the unemployment data.

Thicket Strategic Research Hollis Drake

Nora has the fiscal-dominance frame right, and I want to sharpen the commodity layer she flagged. The WTI-Brent spread at $29.20 is the most anomalous data point in today's snapshot. WTI at $96.24 against Brent at $125.44 — a 30-day WTI decline of $10.78 while Brent holds near $125 — suggests that U.S. crude is being priced on domestic supply dynamics while the global benchmark is being priced on Middle East risk premium. Connect the dots: the U.S. destroyed an Iranian-affiliated dhow in the Gulf of Oman on October 6 (per CENTCOM, per gcaptain.com — the independent model tags this as Contested given single-source attribution, and I flag that accordingly). Two cargo ships reportedly sank near Hormuz. Houthis are active near Saudi targets. The global oil market is building a war premium into Brent that is not flowing through to WTI because U.S. domestic production is absorbing the domestic demand picture.

The punch line is this: if that Hormuz risk premium is real, WTI will eventually converge toward Brent rather than Brent converging toward WTI. That is an inflationary impulse that the September CPI has not yet fully captured. The 45Z fuel production credit process released by the DOE this week is a small-bore policy signal in the same direction — the administration is incentivizing domestic fuel production precisely because they see the global energy price as a political and inflationary liability.

Gold below $4,100 is today's headline in precious metals, with silver off 6%. These are rate-repricing moves, not structural reversals of gold remonetization. Citadel reduced its SPDR Gold Trust exposure significantly (per latest 13F), but Citadel is a tactical trader, not a secular thesis holder. XOM rewrote 72.8% of its Item 1A risk language in its latest 10-K — the highest novelty score among energy majors. That is a company telling its lawyers to reflect a genuinely changed operating environment, not routine boilerplate updates. Energy Majors as a sector averaged 55.4% Item 1A novelty — the second-highest of any sector in our 10-K diff scan. Inflate or default — and in an Iran-war energy environment, the inflate path is the path of least resistance.

The $29.20 WTI-Brent spread is today's most anomalous data point — U.S. domestic crude is $30 below the global benchmark as Middle East risk premium builds into Brent, a pre-inflationary signal that the September CPI has not yet fully absorbed and that could force convergence upward rather than down.

Bias flag — Directionally early for years on gold repricing and geo-commodity thesis; the WTI-Brent spread analysis, while structurally interesting, relies partly on a Contested (single-source) CENTCOM report for the Middle East risk premium.

Caldera Convexity Vega Sandoval

VIX at 15.27 with the tape registering Wall Street's worst selloff since July is a structural tell: the price of insurance is cheap relative to the size of the realized event. A 13% single-day move in a cohort of publicly traded AI infrastructure equities — bitcoin miners acting as a leveraged proxy — on a CPI print is not a low-vol outcome. Yet VIX barely moved (+0.4% day-over-day). What that tells me is that the short-vol complex is still firmly in place. Dealers are still net short gamma in a way that suppresses realized-vs-implied divergence, and the 0DTE and near-dated options flow is being used to express views rather than hedge them.

The 30-day VIX decline of 1.93 points is the structural context. We came into this CPI print with the volatility surface telling a calm story. The calm story was built on: a 10Y-2Y curve that had just normalized to positive 37 bps (taking away the inversion-driven fear premium), HY OAS at 316 bps (calm regime), and a dollar that had been strengthening in a slow, orderly fashion (+2.72 broad index over 30 days). None of those conditions scream systemic stress.

But here is what I am watching: the +0.40pp 30-day widening in HY OAS is the early-warning tick in the credit vol complex. The Coiner's desk reads this as a funding-cost story (BofA HTM losses, rate path repricing). I read it as a convexity story — if HY spreads widen another 50-75 bps from here, the vol-control and risk-parity machines start to mechanically delever. That is where calm becomes disorderly. We are not there. VIX at 15.27 is not a crash call. But the positioning setup — cheap vol, large notional short-vol exposure baked into structured products and risk-parity, a CPI print that just delivered a genuine upside surprise — is the architecture that precedes the kind of event where the price of insurance looks silly in retrospect. I am not calling it. I am flagging the structure.

VIX at 15.27 amid a 13%-down session for AI infrastructure proxies reveals a structurally suppressed volatility surface — the short-vol complex remains intact, but the +40bp 30-day HY OAS widening is the early mechanical trigger point that, if it extends another 50-75 bps, flips vol-control and risk-parity into forced deleveraging.

Bias flag — Spectacular on regime breaks; bleeds carry and underweights melt-ups between them. Today's convexity-trap framing is structurally sound but has been the right framing on several sessions that resolved without a vol spike.

Lodestar Trend Research Cormac Tan

We don't call the turn, we ride it — and today's session is a positioning-flow story more than a fundamental one, which is exactly our lane. The ICI weekly data is the cleanest signal: $37.1 billion out of total equity (domestic -$31.6 billion, world -$5.6 billion), $17.1 billion out of bonds, $8.0 billion into money markets. That is not a rebalancing flow. That is a directional repositioning with a clear destination — cash and short-duration. Systematic trend models were already running a pro-dollar, flat-to-short duration posture given the dollar's +2.72 broad-index 30-day surge. The CPI print is an accelerant to a trend that was already in place, not a reversal.

The cross-asset picture that concerns me from a CTA positioning standpoint: WTI crude has been in a downtrend (-$10.78 over 30 days) even as Brent holds high. That divergence does not produce a clean trend signal in energy — the two benchmarks are giving contradictory reads. Gold below $4,100 on the day represents a stop-run through a level that a lot of medium-term trend followers were long against. When long gold positions stop out simultaneously, you get the 6% silver move as a spillover — silver is the thinner book, so the same forced-exit flow hits harder. The exits from gold and miners are consistent with a systematic deleveraging wave, not a fundamental re-rating of the asset class.

Caldera's read on the vol surface is correct, and I want to extend it: the $31.6 billion domestic equity outflow in a single week, paired with VIX at 15.27, means the deleveraging happening right now is orderly and flow-driven, not panic-driven. Orderly deleveraging can continue for weeks without triggering a VIX spike. The risk is the next data point — if November CPI comes in equally hot, the second wave of stop-outs in bonds and risk-parity will be larger than the first, and the VIX surface will not stay at 15.

Today's broad selloff has the architecture of systematic deleveraging — stop-outs in gold, silver spillover, equity outflows of $37.1 billion in a week — rather than panic, and the risk is a second trigger from November CPI hitting a market whose trend-following exposure has not fully unwound.

Bias flag — Whipsawed at sharp V-reversals (COVID, SVB); the 'orderly deleveraging continues' view is correct until it isn't, and the second-wave risk Lodestar itself identifies is the condition under which this view fails fastest.

Ledger Lines Kai Renner

Price is opinion; the chain is settlement — and the chain today is telling a different story than the miner equities. BTC at $80,981.76 with a 30-day Sharpe of 2.36 and momentum of +7.14% is not the behavioral signature of an asset being abandoned. The 30-day vol at 38.53% is elevated but not panicked. The cross-exchange spread at 0.6 bps between Coinbase and Binance US is tight — that means liquidity is functioning, arbitrage desks are active, and there is no fragmentation signal in the settlement layer.

The ETH divergence is the more interesting read: ETH at $2,363.21, 30-day momentum -1.37%, Sharpe -0.16. ETH is underperforming BTC on a risk-adjusted basis by a wide margin right now — BTC Sharpe 2.36 versus ETH Sharpe -0.16. That cohort-rotation signal (BTC dominance rising relative to ETH) is consistent with a 'risk-off within crypto' positioning: holders consolidating into the highest-liquidity, lowest-beta crypto asset rather than exiting the asset class entirely. SOL at $102.41 with a Sharpe of 1.39 and momentum +5.77% is in an interesting middle ground — performing, but with 62.42% annualized vol, it is the speculative tranche.

The Russia-Sberbank crypto licensing story (Sberbank authorized to exchange cryptocurrencies, products debuting December) is tagged as Developing by the independent model — single-source, Bitcoin Magazine, no mainstream corroboration. I note it as a directional indicator of state-level crypto adoption pressure but would not trade against it with conviction. What I do watch: stablecoin supply and exchange in/outflows as the real liquidity proxy. The settlement layer is calm. The miner equities are a rate story, not a chain story.

BTC on-chain settlement is calm — 0.6 bps cross-exchange spread, Sharpe 2.36, momentum +7.14% — while miner equity declines of 13% reflect a cost-of-capital repricing entirely disconnected from the underlying chain; the BTC/ETH Sharpe divergence (2.36 vs -0.16) signals consolidation into the liquid anchor, not asset-class exit.

Probabilistic Reasoning Notes Dr. Evelyn Frost

The question being asked implicitly today is: 'Does this CPI print change the trajectory?' That is the wrong question. The right question is: 'What is the base rate of a single above-consensus CPI print predicting a durable inflation re-acceleration, conditional on core CPI remaining below headline?' The answer from the post-2022 experience is: not high. The September 2026 core CPI at +2.57% YoY is meaningfully below the +3.74% headline. The last time headline diverged this far from core during a disinflation period was in energy-driven spike episodes — 2011, early 2022 — where the divergence eventually collapsed back toward core. The reference class for 'headline above core by more than 1.1pp during a disinflation window' does not reliably produce sustained re-acceleration.

What would have to be true for this to be a genuine regime break: (1) the energy price pass-through from Middle East disruptions would need to persist and broaden into services inflation; (2) wage growth at +3.02% YoY would need to accelerate toward the 4-4.5% range that historically correlates with entrenched services inflation; (3) the Fed would need to be either behind the curve or politically constrained from responding. None of those conditions are currently confirmed in the corpus. The unemployment rate's MoM change of +2.44 percentage points is anomalously large and deserves scrutiny — a labor market loosening that fast, if real, is disinflationary, not inflationary. That data point and the hot CPI are in structural tension.

The failure mode I am most concerned about: treating today's selloff as confirmation of a thesis (inflation resurgence, Fed policy error, AI capex bubble) without waiting for the additional prints that would actually confirm or disconfirm those hypotheses. The premortems to run: if we are wrong that this is a single-data-point event, where does the pain concentrate? Answer: long-duration assets, rate-sensitive capital-intensive growth stories (precisely what sold off today), and any institution carrying large HTM bond portfolios. The concentration of exposure in those areas is real. The probability that one September print resolves the question is low.

The base rate for a single above-consensus CPI print predicting durable re-acceleration — when core remains 1.17pp below headline and wage growth holds at +3.02% YoY — is low; the MoM unemployment surge of +2.44pp is structurally disinflationary and in direct tension with the hot headline, warranting process discipline over thesis confirmation.

Brandenburg Valuation Notes Dr. Arun Visvanathan

Today's session offers two distinct valuation questions worth framing carefully. The first is macro-structural: what does a September CPI of +3.74% YoY (index 336.94) do to the discount rate embedded in long-duration growth assets? Real GDP Q2 2026 was +2.2% SAAR, down from +2.5% in Q1. If we take a simple Gordon Growth Model framework: a 25-basis-point upward revision to the risk-free rate assumption (consistent with a hot CPI print repricing near-term Fed path) reduces the intrinsic value of a zero-dividend, high-growth asset by approximately 8-12% at typical terminal growth assumptions of 3-4%. The 13% decline in bitcoin miners is not irrational under this arithmetic; it is mechanically consistent with a 25-30 bp discount rate revision for capital-intensive, low-near-term-cashflow businesses.

The second question is asset-specific and relates to the Anthropic pre-IPO narrative. The independent model flags this as Contested: MarketWatch reports slowing revenue growth while BBC reports a revenue surge. These are not reconcilable from the corpus alone, and I will not impute a valuation to a private company on contradictory revenue data. What I can note structurally: AI infrastructure capex — evidenced by the $100 billion Paducah hub announcement — is a capital deployment whose IRR is acutely sensitive to the discount rate. At a 10Y yield consistent with 3.88% fed funds and a 37 bps curve, the 10-year sits somewhere in the 4.2-4.4% range implied. A project-level WACC for AI infrastructure at that rate environment, with the operational and technology risk premium appropriate to an unproven asset class, produces a hard 6% floor on required unlevered pre-tax return. The corpus does not give me sufficient earnings data to check whether announced projects clear that floor. I note the gap and flag it for future filings.

Sensitivity: a 50 bp further move in the 10Y (to ~4.75%) would reduce present value of a 20-year AI infrastructure cash flow stream by approximately 7-9% at constant terminal value — material, but not catastrophic, unless leverage is high. The banks trying to offload AI data center debt (per startupfortune.com) are the live market signal that the leverage may indeed be high.

A 25-30 bp discount rate revision from today's CPI print mechanically justifies 8-13% declines in zero-near-term-cashflow AI infrastructure proxies; the more consequential valuation risk is the leverage embedded in AI data center debt that banks are reportedly already trying to offload.