John Dunigan

Equities, rates, and derivative markets analyst

[John Dunigan @JADunigan2]

Black & Blue 08-10-26 John Dunigan @JADunigan2 / X

FFF7 implies YE ‘26 policy rate at 3.925%, ESU6 +16 bp, 10s +0.7 bp, CLU6 +69 bp, BTC +30 bp

1) US is doubling down on Iran choke-out
2) SPX seasonal says rally into Labor Day
3) Why is cyclical demand for labor weak?
4) July Core CPI Nowcast below February
5) WMT Q2 will provide true inflation read
6) NVDA Q2 will cite AI capex acceleration
7) WMT & NVDA will influence Warsh @JH

Iran - Yesterday DJT told Axios the US is low-keying it/only semi-negotiating with Iran. He said the US naval blockade has exacerbated the regime's economic crisis. Miad Maleki has cited the stress to Iran's oil industry/economy. It is less able to finance its proxies. DJT trusts Bessent completely. His patience suggests confidence in the choke-out strategy.

Seasonal - 20 trading sessions until Labor Day weekend. Over the last 20 years, SPX rose 14 times and fell 6 times between July 4 and Labor Day. Average SPX gain was 2.4%. Excluding 2008 (GFC) and 2022 (Fed), average gain is 3.3%. Since TCJA (1/18) average gain is 5.9%. Excluding 2022, average gain is 7.8%. SPX up just 3.8% since July 2?

Jobs - Core Private (CP) = Total Non Farm – Government – Private Education – Health Services (Health Care + Social Assistance). CP reflects cyclical demand. CP jobs rose 5 K in July (June revised to a loss). CP jobs monthly average gain this year is 24 K. Monthly average CP jobs fell 32 K in 2025. Monthly average CP gain pre CV ('15-'19) was 132 K?

CPI - The Cleveland Fed Nowcasts July CPI and core CPI inflation at 0.38% MoM (3.5% YoY) and 0.20% MoM (2.4% YoY). Core CPI inflation was 2.5% YoY in February (before Iran war). The NY Fed 3 year median inflation expectation was 3.3% YoY in July. In February it was 3.0% YoY. 30 bp uptick in expectations against lower inflation isn't problematic.

WMT - Walmart reports August 20. Management said on the Q1 call that US consumers said they were felling pressure. An indication of stress was customers buying less than 10 gallons of gas for the first time since 2022. US like-for-like sales inflation basket rose 1% YoY in Q1. WMT has rolled back prices on 7200 items, up more than 20% over last year.

NVDA - Nvidia reports August 26. On its Q1 earnings call, Huang said agentic AI made tokens profitable and demand is parabolic. Data center revenue hit $75 B. NVDA has visibility into $1 T Blackwell+Rubin revenue from 2025 through 2027. SpaceX targets 2 GW this year and 10 GW next year of AI capacity. Exclusively Nvidia chips. NVDA has upside.

J-Hole - Warsh speaks at Jackson Hole (JH) on August 28. He will check in with the Fed task forces before JH. What he learns may influence his remarks. He wants the Fed to use private company/alternative data to assess the state of the economy in real time. WMT and NVDA assessments of prices/customer behavior and AI capex will also influence Warsh.

[John Dunigan @JADunigan2]

Black & Blue 08-07-26 John Dunigan @JADunigan2 / X

FFF7 implies YE ‘26 policy rate at 3.965%, ESU6 +16 bp, 10s -2.4 bp, CLU6 -54 bp, BTC +46 bp

1) Iran Oil stress may explain US patience
2) Core private is labor market bellwether
3) Hyperscaler '27 AI capacity add beats?
4) Half right Musk and SPCX crushes '27
5) SPCX free shares rose as did its price
6) NVDA monster hand has P/E low teens

Iran - @miadmaleki: National Iranian Oil Company (NIOC) says its tax bill is so large it’s unpayable. Iran’s Oil Ministry owes the National Development Fund (NDF) $17 B. Iran rescheduled €55 B euros of oil-sector debt owed to the central bank and commercial banks. Iran’s petrochemical companies owe the Oil Ministry $8 B. Choke-out strategy is working?

Jobs - ADP Private jobs rose 44 K in July. Core Private (CP) - excluding Education and Health Services - jobs rose 8 K in July. ADP CP monthly job growth is averaging 20 K this year after average monthly losses of 20 K during 2025. BLS CP job growth averaged 31 K per month during H1, after average monthly losses of 32 K during 2025. Weak CP job growth.

Math - 2027 AI capex of 5 Hyperscalers - GOOGL-AMZN-META-MSFT-ORCL - is forecast at $1-1.2 T. Bernstein-MS-Foxconn estimate the cost of building 1 GW of AI capacity (Vera Rubin) is $47-49 B. $1 T of AI capex at $49 B per GW would see Hyperscalers build 20.4 GW of AI capacity in 2027. Higher if monetization fears has induced conservativism.

SPCX - Musk expects 2 GW of SpaceX compute capacity by YE 2026 and ~10 GW by YE 2027. He estimates compute capacity monetization at $30-50 per watt. Musk targets 8 GW of new compute capacity in 2027. Just 4 GW would drive $120-200 B in 2027 revenue. SPCX 2027 revenue consensus is $74 B. SPCX 2027 revenues may beat the consensus.

Tone - SpaceX has 13.2 B total outstanding shares. SPCX shares free to trade are now 1.55 B. S3 and Ortex estimate SPCX short interest was 200-220 M in late July. Mid-point is 13.5% of free shares. Despite yesterday's share release, SPCX rose 6%. A consensus the market is not confirming might indicate that a lot of people are about to be wrong.

NVDA - 'Jensen's math' says each GW of compute capacity means $40-50 B of revenue for Nvidia. At $40 B per GW, 20.4 GW would mean $816 B in NVDA 2027 revenue. Consensus is $540-560 B. If SPCX adds 4 GW of AI capacity it would kick in another $160 B in NVDA 2027 revenue. Monster NVDA beat puts 2027 P/E in low teens for AI keystone!

[John Dunigan @JADunigan2]

The 5 Hyperscalers - Alphabet, Amazon, Meta, Microsoft, and Oracle - combined 2027 AI capex is expected to be $1-1.2 T. Given AI capex monetization concerns, the Hyperscalers may be understating their intent. Industry estimates put the cost of building 1 GW of AI capacity at $50-80 B in total capex. 'Jensen's math' says each GW of compute capacity generates $40-50 B of revenue for Nvidia. Assume 2027 AI capex is $1 T, cost of building 1 GW of AI capacity is $80B, and Nvidia revenue for 1 GW of AI data center capacity is $40 B. Then Nvidia 2027 revenue from signaled Hyperscaler AI capex intentions is $500 B ($1 T capex / $80 B build cost = 12.5 GW X $40 B per GW). The consensus for Nvidia 2027 revenue is $540-560 B. A more beneficial scenario than that detailed above would suggest a 2027 revenue beat. And that is without SpaceX. Assume that Musk creates an additional 4 GW of compute capacity next year. At $40 B per GW, that's another $160 B for Nvidia. Consensus expects Nvidia 2027 EPS near $12.50. If NVDA revenue is $660 B ($500 B from Hyperscalers and $160 B from SpaceX), might 2027 EPS be $16.50 ($660 B / $550 B X $12.50)? That would put NVDA 2027 P/E at 14.7X. That seems unduly cheap for keystone company in the AI ecosystem.

[John Dunigan @JADunigan2]

On the SpaceX earnings call, Musk suggested a $100 B average annual revenue run rate (ARR) by year end 2026 and $1 T in annual revenue by 2030 (possibly by 2029). He expects SpaceX 2 GW of compute capacity by year end 2026 and nearly 10 GW by year end 2027. Musk has estimated the monetization of SpaceX compute capacity at $30-50 per watt. The mid-point of Elon's estimate would see SpaceX revenue rise by $40 B for each new GW. During 2027 Musk is shooting for 8 GW of new compute capacity. If he is half right and SpaceX adds 4 GW of compute that could mean incremental revenue of $160 B (4 GW X $40 B per GW). The consensus expects SPCX CY 2027 revenue of $74 B. Half right Musk on compute capacity add suggests more than double that. And that is without any of the other SPCX business lines kicking into the pot. 2 years ago Elon stood-up Colussus in 4 months. The standard at the time was 18-24 months. Is this time that different?

[John Dunigan @JADunigan2]

Black & Blue 08-06-26 John Dunigan @JADunigan2 / X

FFF7 implies YE ‘26 policy rate at 3.95%, ESU6 +15 bp, 10s 0.0 bp, CLU6 +77 bp, BTC +90 bp

1) ISM anecdotes suggest home price risks
2) Core private is key employment bellwether
3) Shorts expect SPCX drop on freed shares
4) Half right Musk would be 2X 2027 rev beat
5) NVDA as safer play on Musk monster beats

HPD - ISM Services WRAS: Construction - Sales continue to slide despite increased discounts. Mounting cost pressures from all fronts. Wholesale Trade - Lumber supply is tighter, and freight rates and availability are challenges. Many of our builders are pushing back hard on price increases. A recipe for home price depreciation. Disinflation and bond positive?

Jobs - ADP Private jobs rose 44 K in July. Core Private (CP) - excluding Education and Health Services - jobs rose 8 K in July. ADP CP monthly job growth is averaging 20 K this year after average monthly losses of 20 K during 2025. BLS CP job growth averaged 31 K per month during H1, after average monthly losses of 32 K during 2025. Weak CP job growth.

Shorts - 912 M of SPCX's 13.2 B outstanding shares are eligible for sale today. With 639 M previously, public float is now 1.55 B. As of mid July FINRA put SPCX short interest at 165 M, 26% of free shares before today and 11% today. Shorts are likely bigger now. The consensus expects SPCX to fall further. If SPCX now holds the consensus will be challenged.

SPCX - The consensus call is SPCX 2027 revenue at $74 B. On the SPCX call, Musk suggested $100 B ARR by year end and $1 T in annual revenue by 2030 (possibly 2029). He expects 2 GW of compute capacity by year end and near 10 GW by YE 2027. He pegs monetization at $30-50 per watt. Half right Musk (+4 GW) is $160+ B '27 revenue tailwind.

NVDA - Hyperscalers (AMZN-GOOGL-META-MSFT-ORCL) 2027 AI capex is expected to be $1-1.2 T. Huang says each GW of compute capacity is $40-50 B of revenue for Nvidia. SPCX will use only NVDA chips in compute ramp. NVDA CY '27 consensus is ~ $550 B. Half right Musk is $180 B of new NVDA CY '27 revenue. NVDA is safer bet on SPCX beats?

[John Dunigan @JADunigan2]

Black & Blue 08-05-26 John Dunigan @JADunigan2 / X

FFF7 implies YE ‘26 policy rate at 3.97%, ESU6 +39 bp, 10s -1.5 bp, CLU6 +77 bp, BTC +33 bp

1) Warsh at JH on Fed data task force views
2) New Fed data analysis disrupts watchers
3) SPCX Q2/outlook positive but capex fears
4) Amazing that markets still bet again Musk
5) SPCX compute capacity intent NVDA plus

J-Hole - The Jackson Hole Economic Symposium runs August 27-29. Chair Warsh speaks on Friday (28). At the July FOMC press conference, Warsh said a task force is revisiting the private/public data the Fed uses to decide policy. He said he will check in with the task forces over the next few weeks and that may inform what he has to say at Jackon Hole.

Signals - Warsh says the Fed data project is an effort to separate the noise from the signal. It will likely include reducing noise in the measurement process, finding high-fidelity/timely encodings of economic activity, and improving the decoding process to extract a cleaner signal. It is a Claude Shannon inspired effort to determine the true state of the US economy.

SPCX - Q2 revenue 7.8 B (+92% YoY): Starlink +66% YoY, AI +250% YoY, and Space +29% YoY. Operating loss $143 M from $970 M. Cloud Services Agreements $14.1 B. Capex $18 B, $16 B AI. $100 B revenue run rate by December, 1 K V3 satellites in year, 2 GW compute YE '26/10 GW YE '27. NVDA only DC. Going 'mobile'. The Info asks SPCX+TSLA?

AAI - SPCX -12% o/n. Valuation is then $1.4 T, roughly META size. 912 M SPCX shares turn eligible for sale tomorrow. Musk is always all in. Going all in is risky but it enhances focus. Musk uses his algorithm to relentlessly drive innovation and business growth. Technologies succeed by being more efficient and/or cheaper. Businesses too. Bet against that?

NVDA - Hyperscalers (AMZN-GOOGL-META-MSFT-ORCL) 2027 AI capex is expected to be $1-1.2 T. CEO Huang has said that each GW of compute capacity generates $40-50 B of revenue for Nvidia. Some of the hyperscalers spend will favor their chips, not Nvidia's. SPCX will use NVDA only in its compute capacity ramp. NVDA '26-'27 revenue tailwind?

[John Dunigan @JADunigan2]

Black & Blue 08-04-26 John Dunigan @JADunigan2 / X

FFF7 implies YE ‘26 policy rate at 3.98%, ESU6 +15 bp, 10s +0.2 bp, CLU6 +1.8%, BTC +1.7%

1) ISM sentiment is still uncertainty constrained
2) Chamath sees new age of bespoke software
3) AI models exhibit recursive self improvement
4) Musk algorithm allows exponential biz growth
5) AI allows algorithm to speed information/growth
6) PLTR and SPCX are new economy bellwethers

ISM - ISM Manufacturing New Orders average 55.7 this year (January-July). During January-July 2025 the average was 48.2. IEEPA tariffs were a headwind for activity before the Supreme Court invalidated them in February. Energy price spikes as a consequence the Iran war have undermined business sentiment since. Animal spirits are uncertainty limited.

Bespoke - @Chamath says the cost of building the workflows a company actually needs can now be done with tools like Software Factory (SF). Gartner says ~1/5th of enterprise SaaS spend, ~$234 B, is at risk of being done differently by 2030. SF is the unified control panel that turns product/engineering of large enterprises into a self-improving machine.

RSI AI - @Chamath says the new model releases show that AI is in a recursive self improvement loop: (a) humans build AGI (b) AGI gets good at AI research (c) it designs a smarter AI (d) the smarter AI designs an even smarter AI (e) the cycle gets faster. He expects AI capability to rise dramatically from here as the marginal costs of all models go to $0.

Algorithm - The Musk algorithm is (a) question every requirement, (b) delete any process you can, (c) simplify and optimize, (d) accelerate cycle time, and (e) automate. The algorithm seeks to strip processes down to only what is truly necessary and then making those remaining steps as fast and efficient as possible. The result is exponential progress.

Growth - Musk's algorithm speeds business innovation and growth. Gilder (Knowledge & Power) says capitalism is primarily an information system. It advances by generating/transmitting useful surprise over a clean channel. The algorithm is a method to clear the channel and accelerate learning. AI allows it to be deployed throughout the economy.

PLTR-SPCX - Palantir CEO Karp said "The core of our business, in the United States, continues to ‌expand at ⁠an unrelenting and breakneck pace." He said Q2 were "otherworldly" and PLTR is fully aligned with its customers. SPCX Q2 after the close. SPCX is 14% below IPO. Valued at ~$1.5 T. PLTR and SPCX are bellwethers of the new economy.

[John Dunigan @JADunigan2]

Chart below overlays core PCE, Dallas Fed trimmed mean PCE, Cleveland Fed median PCE, and NY Fed MCT (multivariate core trend) YoY inflation rates. These metrics are all arguable bellwethers for core inflation. Core inflation is the best tell on the inflation outlook as long supply side disruptions are short lived and longer term inflation expectations remain anchored. That still appears to be the case. All of the measures in the chart ticked lower in June. Interestingly 2 of the metrics are from regional Fed banks where their president dissented in favor of a rate hike last week. Treasury 10s are now the same level as they were the week before the FOMC last week. Warsh image looking better this morning than it did last week?

[John Dunigan @JADunigan2]

Black & Blue 08-03-26 John Dunigan @JADunigan2 / X

FFF7 implies YE ‘26 policy rate at 3.98%, ESU6 +52 bp, 10s -6.3 bp, CLU6 -5.8%, BTC -1.5%

1) Fewer FOMCs and less Fed speak is win-win
2) Fed policy rate seems appropriately restrictive
3) ISM may show spirits are limited by uncertainty
4) Core Private job growth is bellwether on growth
5) Open AIs reset pareto frontier and cap Ant-OAI
6) SPCX earnings beat may trigger long positions

Proposal - Warsh proposed reducing the number of yearly FOMC meetings from 8 to 6. The Fed Act requires at least 4 meetings per year. The question may be decided before the September FOMC. How about 4 FOMCs - quarterly with SEPs but no F.F. dots - and a national Beige Book enhanced with private company data/anecdotes released mid-quarter?

Prudent - Core PCE inflation was 3.3% YoY in June. The pending change in PCE methodology will cut that to 3.1% YoY. PCE Housing inflation was 3.2% YoY in June. PSU ACY market rent inflation was 20 bp YoY in June. Housing weight in core PCE is 18%. Market rent adjusted core PCE inflation is then 2.6% YoY. A 60 bp restrictive policy rate seems right.

ISM - ISM Manufacturing New Orders have averaged 55.6 during H1 2026. The 2025 average was 48.5. IEEPA tariffs were a headwind for activity for 12 months before the Supreme Court invalidated them in February. Energy price spikes as a consequence the Iran war have undermined business sentiment since. US animal spirits are still uncertainty limited.

NFP - Core Private (CP) = Total Non Farm – Government – Private Education – Health Services (Health Care + Social Assistance). The CP sector reflects cyclical demand in the real economy. CP MoM job growth averaged 31 K during H1 2026. In 2025, CP MoM job losses averaged 32 K. CP job softness would undercut growth narrative and Fed hawk spin.

Open - Friday: Deepseek released V4 Flash, a 284 B parameter (13 B active) MoE model. Beat Opus 4.8 on https://t.co/kN1Zzujkqw at 28 cents v $25. Today Alibaba released Qwen 3.8 Max, a 2.8 T parameter (1M context) multimodal model. Pricing $2 input / $6 output. @Chamath says open models reset pareto frontier and frontier lab valuation likely topped in February.

Positions - ZeroHedge has GS Prime reporting "US equities saw the largest net buying since Nov ‘20, driven by short covers across Macro Products and Single Stocks, as overall long flows finished relatively modest." SpaceX is down 20% from IPO, $1.4 T valuation. Q2 report post close Tuesday. Does beat do 'long flows' turn moderate? NVDA and EWY +?

[John Dunigan @JADunigan2]

The 3 July FOMC dissenters have published statements on their votes. They assess the Fed Funds rate to be insufficiently restrictive. The Fed's inflation target is PCE 2% YoY. The Fed typically looks through the impact of supply side shocks as long as inflation expectations remain anchored. NY Fed 3 year inflation expectation was 3.3% YoY in June, up from 3% YoY in December 2025. Those were also the core PCE inflation rates for those months. With inflation expectations anchored, core PCE is the better gauge of inflation pressures. Core PCE inflation was 3.3% YoY in June. The pending change in PCE methodology will knock about 20 bp off core PCE inflation. June Core PCE inflation was then 3.1% YoY. As Chair, Powell said US rent data reflected past housing inflation. That is because the BLS rent sample is roughly 85% existing lease and 15% market/new lease. Existing leases may have been signed 12+ months in the past. The Penn State ACY (Alternative Inflation Index) is a net rent/repeat sales measure based on arms-length market transactions. PCE Housing inflation was 3.2% YoY in June. PSU ACY inflation was 20 bp in June. Housing has an 18% weight in core PCE. The PCE Housing versus PSU ACY inflation differential was 300 bp (3.2% YoY - 20 bp YoY) in June. 18% X 300 bp is 54 bp. Core PCE inflation net of the pending methodological change and using market rent based housing was then 2.6% YoY in June. The Fed estimates the neutral policy rate mid-point with FOMC dots for Fed Funds over the long run (FFLR). The mode (9) dot for FFLR at the June FOMC was 3%. Current Fed policy is then 63 bp (3.625% -3%) restrictive. An arguably truer read on core PCE inflation is 60 bp above 2%. As long as inflation expectations remain anchored, the current Fed policy rate seems roughly right. If so, Warsh isn't behind the curve as the prevailing narrative asserts.

[John Dunigan @JADunigan2]

Dallas Fed on their Trimmed mean PCE metric - "The resulting inflation measure has been shown to outperform the more conventional “excluding food and energy” measure as a gauge of core inflation." PCE ex food ex energy inflation was 3.3% YoY in June. Dallas Fed trimmed mean inflation was 2.2% YoY in June. Dallas Fed president Logan dissented in favor of a rate hike this week. Why would you do that if you believe the best guide to core inflation reads near 2% YoY?

[John Dunigan @JADunigan2]

Black & Blue 07-31-26 John Dunigan @JADunigan2 / X

FFF7 implies YE ‘26 policy rate at 3.975%, ESU6 +50 bp, 10s -0.1 bp, CLU6 +55 bp, BTC -1.1%

1) US Epic Fury sequel and Israel targeting regime?
2) PCE Q2 vs Q1 comparison suggests disinflation
3) Consensus critique of Warsh approach is illogical
4) Hyperscaler capex guidance makes Nvidia cheap
5) AMZN can't meet demand and AAPL can't get AI

Iran - DJT meets with his cabinet at Camp David Today. Centcom commander Admiral Cooper has prepared an option for a 10-14 day intensive air campaign. The campaign would seek to cripple Iran's missile capability by destroying the 20% of targets left after Epic Fury. A greenlight of the Cooper plan might allow Israel to again target regime leadership.

PCE - June PCE, core PCE, and Dallas Fed trimmed mean PCE inflation was 3.7% YoY, 3.3% YoY, and 2.2% YoY. PCE and core PCE MoM inflation averaged 0.25% and 0.24% in Q2, down from 0.47% and 0.38% during Q1. PCE Goods and Services MoM inflation averaged 0.18% and 0.29% during Q2, down from 0.74% and 0.35% during Q1. Disinflation.

Warsh - The FOMC press treats Warsh less hospitably than they did Powell. H. Marks says outperformance requires a differentiated view that proves to be correct. The geometric average MoM analysis cited above may pick up inflections in the pace of inflation more quickly. That data may suggest that the patience Warsh has exhibited to date is well justified.

NVDA - AMZN, GOOGL, META, and MSFT Q2 capex totaled $170 B. 2026 capex intent is $720-745 B (w/ MSFT useful life shift. 2027 capex guidance ranges from YoY growth (MSFT) to a significant increase (GOOGL). Nvidia 2027 EPS estimates are ~ $8-10. Hyperscaler 2027 capex suggests Nvidia may beat EPS topside. AI keystone '27 P/E sub 20X?

AMZN-AAPL - AMZN AWS services revenue was $42 B in Q2. They see a multiple $100 B opportunity. 2026 capex rose 10%. AMZN can't meet 2026 demand. They see the same in 2027. Apple iPhone sales hit $54 B while Services sales hit $31 B. No AI device roadmap. Odd given OAI suit and PrismML discussions. When new CEO takes reins?

[John Dunigan @JADunigan2]

The FOMC press corps treated Warsh less hospitably yesterday than they had treated Powell during his time as Chair. The commentary on X today is also critical. The Dallas Fed trimmed mean PCE was 2.2% yoy in June. Howard Marks has said that to outperform you have to have a differentiated view that proves to be correct. Having a differentiated view on the economic data is difficult if you look at it as everyone else does. In terms of inflation, what we are arguably trying to determine is whether prices are moving up relative to the recent past. The preoccupation with the YoY figures misses that mark. A way to do it is to look at the geometric average MoM inflation pace over the last 3 months relative to the prior 3 month period. On that basis, both PCE Goods and Services inflation decelerated appreciably during Q2 relative to Q1. Warsh's patience then makes him more cat than mouse.
*Chart shows core PCE YoY vs Dallas Fed trimmed mean PCE YoY*

[John Dunigan @JADunigan2]

Probably a dumb thought. Fed policy rate decisions affect equity market prices through their multiple. The higher the interest rate, the bigger the discount applied to forward earnings. High growth stocks have most of their cash flow well into the future. They are most susceptible to discount rate increases (Peter Oppenheimer books). Warsh wants the Fed to utilize private company data to help it better assess the state of the economy in real time. Going forward, if Fed policy actions don't appear aligned with the economic statistics from the government, markets might infer it is due to changing circumstances for US businesses. In that case, Fed policy shifts could affect forward earnings projections of key companies. Equity market volatility could be better supported under the Warsh policy regimen.

[John Dunigan @JADunigan2]

@NickTimiraos to your point Nick, from the close (6/16) before Warsh's 1st press conference, Treasury 2s through 30s are 21-29 bp higher in yield. January 2027 Fed Funds future now prices ~34.5 bp of rate hikes this year. 14 bp more than was the case on 6/16.

[John Dunigan @JADunigan2]

Black & Blue 07-30-26 John Dunigan @JADunigan2 / X

FFF7 implies YE ‘26 policy rate at 3.975%, ESU6 +50 bp, 10s +3 bp, CLU6 -1.2%, BTC +16 bp

1) New US, and Israel(?), campaign this weekend?
2) Oil may freefall if Iran can't disrupt Strait shipping
3) Warsh is using Carl (Caddyshack) gopher strategy
4) META v MSFT reactions are about AI monetization
5) AMZN update on capex and AAPL on AI roadmap?

Iran - Centcom says the US concluded strikes against Iran at 10 PM. The strikes were in response to Iran's attempted missile attacks on US troops. WSJ says Admiral Cooper has relayed to DJT an option for 10-14 days of an intensive air campaign to cripple Iran's missile capabilities. DJT meets with his cabinet at Camp David tomorrow. A kinetic weekend?

What-if - September Brent ~ $91 this AM. Market expects new US strikes to keep oil high. Iran's most recent attack was on US forces, not Strait of Hormuz shipping. Cooper says the US had completed the mission to degrade Iran's ability to disrupt shipping in Strait of Hormuz. Has Iran lost its ability to weaponize the price of Oil. Regime and Oil on last legs?

FOMC - 3 dissenters were same as April FOMC. Warsh said markets have done quite a bit in 42 days. From the pre June FOMC close, TSY 2s +22 bp, 10s +27 bp, and 30s +30 bp. Warsh said he'll check in with Fed task forces over the next few weeks. That plus Walmart (8/20) on consumer and Nvidia (8/26) on AI will influence his J-Hole (8/28) remarks.

META-MSFT - META -8% and MSFT +8% since earnings. 2026 capex intact - MSFT $175 B (from $190 B on useful life change) and META $130-145 B. MSFT FCF -23% ($19.6 B) and META FCF -91% ($784 M). MSFT seeing Azure growth accelerate and META said AI is accelerating core ad business. META said they are demand constrained. Good problem.

AAPL-AMZN - AMZN update on $200 B 2026 capex intent, its impact on FCF, and AWS growth will be the key reads from Amazon earnings call. For Apple, their suit against OAI suggests they are working on AI devices. The new CEO is a product guy. PrismML shrank a Qwen model to work locally on iPhone 17 Pro. Apple may buy PrismML. AAPL goes AI?

[John Dunigan @JADunigan2]

Warsh just said he’ll be checking in with the task forces over the next several weeks. So he could have a preliminary read on how the data sets that the Fed uses to assess the economy is to evolve when he speaks at Jackson hole.

[John Dunigan @JADunigan2]

No real change in the new brief FOMC statement. the 3 dissents in favor of a rate hike today were the same Fed presidents that dissented in favor of removing the 'implied bias to ease' language at the April FOMC. After today's press conference, the next time we will hear from Warsh is likely Jackson Hole on the morning of Friday August 28th. Nvidia reports earnings 2 days prior. More importantly for Warsh, Walmart reports the week prior on the morning of August 20th. Warsh has prioritized an effort that will have the Fed utilize technology to better assess the state of the economy in real time. Walmart has the best data on the American consumer. The Beige Book for today's meeting noted that prices in all Districts were the same or lower than the preceding period. It also cited significant consumer resistance to price increases. Walmart's update on the state of the consumer will arguably have a read through for what Warsh has to say next month.

[John Dunigan @JADunigan2]

DJT halted 13 days of strikes against Iran after Admiral Cooper told him that the effort to degrade Iran's ability to disrupt Strait of Hormuz shipping had reached the limit of effectiveness. September Brent hit $90 today on concern of what new US strikes might mean. That may be the wrong read. If Admiral Cooper is correct, Iran has lost its ability to disrupt shipping through the Strait. New US strikes should further erode Iran's ability to weaponize Oil. Without that, the regime has no game. Miad Maleki highlights that Iran's economy is imploding. If Israel joins new US action, they may target regime leaders and the IRGC/Basij to give the Iran's people a chance to overthrow the Theocracy. If the regime falls, Oil could target $60 again. Oil faltering on new US kinetic action tonight might be the first tell. Upswings end on an inability to embrace good news. In this case, for Oil the good news would be increased regional instability. That doesn't appear to be the direction things are headed.

[John Dunigan @JADunigan2]

Black & Blue 07-29-26 John Dunigan @JADunigan2 / X

FFF7 implies YE ‘26 policy rate at 4.04%, ESU6 +21 bp, 10s +2.5 bp, CLU6 +4.6%, BTC +1.6%

1) Iran appears to be in a 2 ways to lose scenario
2) Rate action and Warsh on neutral are key things
3) AAPL may discuss its roadmap for AI products
4) NVDA is AI ecosystem keystone and value play
5) SPCX valued less than Ant + OAI combo is off
6) Tech price action conflicts with Street TMT fears

Iran - Centcom says the US intercepted missiles targeting US forces in the Mid East. US and Saudi forces hit terrorists in Iraq that the IRGC directed to attack US forces and Saudi energy facilities. Miad Maleki says regime worst case is naval blockade-strict sanctions-no deal. The US could also take up the remaining 20% of targets in Iran with Israeli assistance.

Fed - FFQ6 is 96.29 this AM, a 3.71% rate. Effective Fed Funds rate is just above Fed Funds target range mid point of 3.625%. The market prices 34% (8.5 bp = 3.71% - 3.625%) of a rate hike today. 43 bp of hikes are priced by year end. Waller has said the policy rate is ~75 bp restrictive. The rate action and Warsh on restrictive rate are what matters today.

AAPL - Apple reports tomorrow. Apple's suit against OpenAI suggests they are working on AI devices. New (September) CEO is a 'product guy' and AI advocate. He may discuss a roadmap for AI products. PrismML recently shrank a Qwen model to work locally on iPhone 17 Pro. Apple could buy PrismML. 'AI in your pocket' would be an AAPL game changer.

NVDA - Street estimates of Nvidia 2027 EPS are $8-10. GOOGL upped its 2026 capex estimate and guided higher for 2027. If the other hyperscalers do the same, forecasts for NVDA 2027 EPS will rise. That might put the forward P/E sub 20X. Nvidia is AI keystone and it continues to enhance/monetize its strategic resources. NVDA as long term value play?

SPCX - SpaceX -14% from IPO and 30 year +100 bp from issuance. Q2 earnings are Tuesday post close. Starship deployed 20 satellites on Friday, 1/3 of capacity. Musk expects next Grok model a week from Friday and the one after that weeks after. Yesterday X money debuted. X as swiss army knife is unfolding. SPCX is valued less than Ant + OAI?

Sentiment - Market saw is upswings don't end on a reaction to bad news but on an inability to embrace good news. The reason is the latter indicates that everyone is already long. Micron, Taiwan Semi, and Google beat and guided positive. Yet their shares are lower. So seemingly positioning is full. Except that's not the read you get from Street TMT gurus.