Cycle on Sep 14, 2026

I will begin by listing the open threads and watch-triggers for this topic to understand what has been tracked or set up in prior cycles.

I will fetch the market view for Oracle (ORCL) to review its financial snapshot, including its recent earnings, capex, debt levels, and share performance, especially focusing on its Q1 Fiscal 2027 earnings release from September 10, 2026.

I will search the web to find the details of Oracle's Q1 Fiscal 2027 earnings release, including the details around the $28.50 billion in capex, the negative free cash flow of $5.40 billion, the increase in job cut costs by $700 million to a total of $2.8 billion, the $664 billion backlog, and any specifics on how this relates to their AI infrastructure spending, debt, and the broader exposure map.

The search results give a very specific picture of Oracle's Q1 Fiscal 2027 results (ended August 31, 2026, reported September 10, 2026). Let's dig deeper. First, let's identify the thread that needs to be resolved: Thread: `Oracle Q1 Fiscal 2027 Earnings Release and Capex/Debt Expansion Tracker` (ID `01a07af0-9f05-7743-93dd-e252cf97fa11`). This thread monitors Oracle's Q1 FY27 earnings release on September 10, 2026, to track updated capex guidance, free cash flow trends, and total debt levels. Let's do a quick web search on Oracle's Q1 FY27 restructuring charges, job cuts, and total debt to get precise details. The market view shows: Total debt: $155.93B (as of 2026-08-31) Cash: $36.37B Q1 FY27 Capex: $28.50B (or $28B in the transcript/earnings articles, let's verify if there is a distinction between GAAP capex, cash capex, and net cash capex). Operating cash flow: $23.10B (or $23B) FCF: -$5.40B (or -$5B) Job cuts / restructuring: the market view mentions "expanded $2.8 billion budget for job cuts" and "Oracle is increasing its estimated job cut costs by an additional $700 million, bringing the total to $2.8 billion, as it grapples with a cash crunch stemming from significant investments in AI data centers." Let's search for articles specifically about "Oracle to spend another $700 million on job cuts" or "Oracle restructuring $2.8 billion" or "Oracle job cuts AI cash crunch". This will give us verbatim quotes for our findings.

Let's look at the prior notes to see how this fits. I will use `read_notes` to see the notes.

The existing note `oracle-debt-and-capex-debacle` deals specifically with Oracle's debt and capex. I should update it. Let's see what else is in the threads. We have: 1. `Oracle Q1 Fiscal 2027 Earnings Release and Capex/Debt Expansion Tracker` (ID `01a07af0-9f05-7743-93dd-e252cf97fa11`). 2. `NVIDIA Q3 Fiscal 2027 Earnings Release & Vera Rubin Ramp` (ID `01a056dd-b86b-73f1-9042-692fba64ec15`). 3. `CoreWeave Q3 2026 Earnings and DDTL Drawdown Disclosures` (ID `01a032d4-4261-7e6d-8d90-9225ae7a1b50`). 4. `SpaceX Colossus Lease 90-Day Cancellation Provision Active` (ID `019feabd-f129-7a64-a91f-ee8a0bcca989`). 5. `Hyperscaler Q3 2026 Earnings and Capex Guidance Revisions` (ID `019fc6b4-e2fe-75a5-ac8a-ba0bc82a9ccf`). 6. `CoreWeave DDTL 5.0 Secondary Market Pricing Track` (ID `019f1273-ce52-7485-a89d-8b5383b6d175`). 7. `Vistra Corp Debt and Leverage Track` (ID `019eca5a-01d4-7c6d-ab25-99092c634700`). Wait, the current date is September 14, 2026. Let's look at the Oracle thread. The fire date was September 10, 2026. This has fired. Let's check if there are other developments. Is there any news on Vistra Corp (VST), Eaton, Vertiv, Constellation, or other companies mentioned in the brief? Let's do a search on Vistra (VST), Vertiv (VRT), Eaton (ETN), Constellation Energy (CEG) to see if there are any recent updates regarding debt, capex, or AI power buildout, or any credit rating actions. Let's search for "Vistra" debt, "Vertiv" debt, "Constellation Energy" debt, "Eaton" debt, "credit rating" or "credit spreads" in the context of AI infrastructure. Let's first search for Vistra Corp (VST) to see if its debt has crossed $22 billion or if there are any recent developments. I will call `market_view` for VST to check its debt levels first.

Let's check VST's debt level from the market view: total debt is $19.89 billion as of 2026-06-30. Wait, let's look at the watch: `Track Vistra Corp's (VST) quarterly total debt levels and leverage. The company's debt currently stands at $19.91 billion. We will monitor if total debt crosses $22 billion as it continues to finance its capital-intensive power buildout.` Currently, it stands at $19.89B (as of 2026-06-30), which is below $22B. However, they recently priced a $1.5 billion offering of junior subordinated notes (September 11, 2026). This $1.5 billion debt offering is a key development, raising their debt and leverage profile. This junior subordinated debt was priced at a 7.00% annual interest rate. Let's search for details about Vistra's junior subordinated notes offering and their utility-leverage exposure to the AI power buildout. Let's do a search.

The Vistra junior subordinated notes offering is interesting. It's $1.5B (Series A $850 million at 7.00% due 2057, Series B $650 million at 7.25% due 2057) which will be used to refinance its 8.0% Series A and 7.0% Series B Perpetual Preferred Stock upon their reset dates in October and December 2026. This is essentially refinancing expensive preferred stock with long-dated junior debt. The total debt of Vistra stands at $19.89B as of June 30, 2026. The new debt of $1.5B is intended to redeem existing preferred stock, so it is largely leverage-neutral or a slight shift on the capital structure from equity-like preferred to junior subordinated debt (which rating agencies often treat as 50% debt, 50% equity). Let's check if there are any other developments. What about the other core entities in the brief? - The chip layer: Nvidia, AMD, Broadcom, Marvell, TSMC. - The hyperscaler buyers: Microsoft, Amazon, Google, Meta, Oracle. - The "neocloud" layer: CoreWeave, Nebius, Lambda. - The power/physical buildout: Vertiv, Eaton, Constellation, Vistra. - The debt/private-credit exposure. Let's do a quick web search on recent developments in private credit exposure to AI data centers, neocloud debt, or any specific vendor financing deals since early September 2026. Let's search for "CoreWeave" debt, "private credit" data center, "neocloud" debt. Let's run a search.

Wow! Look at these developments: 1. **CoreWeave Compute Acquisition Co. VIII, LLC**: On September 22, 2026 (wait, today is September 14, 2026. Oh, the DBRS report has a date of September 22, 2026, or is it expected? Ah, the Morningstar DBRS report lists a credit rating of A (low) with a stable trend assigned to CoreWeave Compute Acquisition Co. VIII, LLC's $8.5 billion Senior Secured Delayed Draw Term Loan Facility (DDTL). This facility covers the acquisition, installation, and operation of the infrastructure needed to operate four data centers across three sites in Virginia, Georgia, and North Dakota. The DDTL consists of a fixed-rate tranche of $4.041 billion and a floating-rate tranche of $4.459 billion. This is a massive structured finance vehicle! It features a 15-month draw period to finance the acquisition of GPUs from Dell, Inc. Wait, the DBRS rating page URL was published recently, let's explore this. Let's do a search on CoreWeave Compute Acquisition Co. VIII LLC to see what this is about.) 2. **CoreWeave's debt growth**: A tech-insider article from August 12, 2026, mentions: - "disclosed total debt of $35.6 billion in its second-quarter filing." (as of June 30, 2026). - CoreWeave's 2026 financing: $8.5 billion delayed draw term loan (DDTL 4.0) in H1 2026, $3.1 billion DDTL 5.0 in H1 2026, and on August 10, 2026, a further $2.6 billion DDTL 5.5 facility priced at SOFR plus 5.50% (arranged by JPMorgan and MUFG, rated Ba2 by Moody's and BB+ by Fitch). - "customer concentration risk: three clients, 72% of revenue" - "the roughly five-year maturity also runs longer than the roughly three-year average length of the customer contracts backing it, meaning lenders are betting CoreWeave can renew those contracts or find new tenants for the capacity before the loan comes due." Let's do a web search on CoreWeave's latest filings or Q2 2026 earnings details, specifically looking at total debt, customer concentration, and the new DDTL 5.5 and DDTL 4.0 and DDTL VIII facilities. Let's do a web search.

Let's read the Dave Friedman Substack article: "CoreWeave Piles on More Debt" or the "Volta's $10B AI Cloud Deal Fuels Neocloud Debt Fears [2026]" article to see what we can find. Or let's search specifically for `"CoreWeave" "total debt" "35.6 billion"` or `"CoreWeave" "35.6 billion"`. Let's do a search on Dave Friedman's Substack.

Let's read the Dave Friedman articles to understand the differences between the DDTL facilities, how they are structured, and the risks. Let's first fetch Dave Friedman's article: `Where GPU Debt Starts to Break` or `The Trophy Deal Trap: What CoreWeave's Investment-Grade Rating Reveals About the GPU Debt...` or `CoreWeave Issued Yet More Debt`. Let's search for the exact URLs and fetch them. Let's do a search on Dave Friedman's Substack to get the exact URLs of these articles.

Let's fetch the Dave Friedman article `CoreWeave Issued Yet More Debt` to get details on the differences between DDTL 4.0, DDTL 5.0, and DDTL 5.5. I'll use `web_fetch` on `https://davefriedman.substack.com/p/coreweave-issued-yet-more-debt`.

This is an exceptional, high-quality, primary-source insight. Let's make sure we understand the dates and figures here: - **DDTL 4.0 (March 31, 2026)**: $8.5 billion delayed draw term loan facility. Rated A3 by Moody’s and A(low) by DBRS. It was issued through **CoreWeave Compute Acquisition Co. VIII, LLC** (wait, let's verify if it's VIII or another number. Let's double check!). In the table, it says "CoreWeave Compute Acquisition Co. VIII LLC" is the issuing SPV for DDTL 4.0. In the text: "matures in March 2032, was issued through CoreWeave Compute Acquisition Co. VIII, LLC, and was structured around a single anchor customer whose credit profile assuaged the rating agencies’ concerns." Wait, who is the anchor customer? The table says "Single IG enterprise (Aa2)" which is widely known or rumored to be Microsoft or Meta. (Let's check if there is more information). - **DDTL 5.0 (May 18, 2026)**: $3.1 billion delayed draw term loan facility. Rated Ba2 by Moody’s and BB+ by Fitch. Issued through **CoreWeave Financing DDTL V, LLC**. Priced at SOFR + 4.50%. Backed by contracts with two large, non-investment-grade customers (undisclosed). - **DDTL 5.5 (August 10, 2026)**: $2.6 billion delayed draw term loan facility. Rated Ba2 by Moody's and BB+ by Fitch. Issued through **CoreWeave Financing DDTL V-V, LLC**. Priced at SOFR + 5.50% (arranged by JPMorgan and MUFG, as mentioned in the tech-insider article). Wait! Let's check Dave Friedman's article "CoreWeave Piles on More Debt" from July 16, 2026 (or August 12, 2026). Let's search for "DDTL 5.5" on Dave Friedman's Substack to see if he wrote a separate post or if we can find more details. Wait! Let's search for "DDTL 5.5" on the web to see if we can read the post "CoreWeave Piles on More Debt" or "Volta's $10B AI Cloud Deal Fuels Neocloud Debt Fears [2026]". Let's fetch `https://davefriedman.substack.com/p/coreweave-piles-on-more-debt`.

This is fantastic! Let's examine the details here: - **DDTL 4.0**: $8.5B, closed March 31, 2026. Borrowed by **CoreWeave Compute Acquisition Co. VIII, LLC**. It is backed by roughly **$19 billion of Meta take-or-pay contracts**, priced at SOFR + 2.25% (floating tranche) and ~5.9% (fixed tranche), matures March 2032. Moody's rated it A3, DBRS A (low), Fitch 'A-sf'. This is the first investment-grade (IG) GPU-backed financing. - **DDTL 5.0**: $3.1B, closed May 18, 2026. Borrowed by **CoreWeave Financing DDTL V, LLC**. Secured against contracts with two large, unnamed, non-investment-grade customers. Priced at SOFR + 4.50%, matures November 15, 2031 (approx. 5.5 years). Rated Ba2 by Moody's, BB+ by Fitch. First publicly syndicated. - **DDTL 5.5**: Proposed in July 2026, closed on August 10, 2026. $2.6 billion, issued through **CoreWeave Financing DDTL V-V, LLC**. Rated Ba2 by Moody's, BB+ by Fitch. Priced at SOFR + 5.50% (arranged by JPMorgan and MUFG, as mentioned in the tech-insider article). It carries a roughly 5-year maturity, which is longer than the roughly 3-year average length of the customer contracts backing it. - **Total Indebtedness of CoreWeave**: $35.6 billion as of June 30, 2026 (per Form 10-Q). - **CoreWeave's backlog**: Revenue backlog of $99.4 billion as of March 31, 2026 (up 284% YoY), with about 36% recognized inside 24 months. - **Customer concentration**: Top two customers were 65% of 1Q26 revenue. The tech-insider article says "three clients, 72% of revenue" as of Q2 2026. - **Capital expenditure**: Fitch forecasts capex around $41 billion in 2026 and $37 billion in 2027. Free cash flow is expected to be negative through 2027. - **Total debt structure of CoreWeave**: - Delayed-draw term loans: DDTL 4.0 ($8.5B), DDTL 5.0 ($3.1B), DDTL 5.5 ($2.6B), plus unrated/private DDTLs 1.0, 2.0, 2.1, and 3.0. - Unsecured high-yield notes: about $10 billion at 'BB-'/'RR4'. - Unsecured convertibles: roughly $6.5 billion (unrated). - OEM financing: $5 billion (unrated). - Magnetar deposit classified as debt: $281 million. - Total debt is $35.6 billion as of June 30, 2026. Wait, let's verify if there are any other recent articles or notes. Let's see if there is any update on CoreWeave's secondary market trading or credit spreads. Let's search for "CoreWeave" "secondary market" OR "credit spread" OR "DDTL 5.0" OR "DDTL 5.5" trading price. Let's do a search.

Excellent. Let's look at the details. The $2.6 billion DDTL 5.5 facility closed on August 10, 2026, and priced at SOFR + 5.50% (which is 100 basis points higher than DDTL 5.0's SOFR + 4.50%, showing widening spreads or pricing pressure on non-IG customer backed facilities!). Let's see: DDTL 5.5 is priced at SOFR + 5.50%, which is a significant spread widening of 100 bps from DDTL 5.0 (SOFR + 4.50%) and 325 bps from DDTL 4.0 (SOFR + 2.25%). This indicates a steepening cost of capital for CoreWeave's non-IG customer backed facilities, even though it has a similar BB+ rating from Fitch and Ba2 from Moody's. Wait, let's look at the watch: `Monitor secondary market trading prices and credit spreads of CoreWeave's $3.1 billion syndicated DDTL 5.0 facility. A drop in secondary trading prices below par or a widening of the spread over SOFR beyond 4.50% would signal a demand softening or credit deterioration in the neocloud sector.` The new DDTL 5.5 facility priced at SOFR + 5.50%. This is a widening of the spread over SOFR beyond 4.50% for a comparable facility (both are non-IG backed, rated BB+ by Fitch / Ba2 by Moody's, with ~5.5 year maturity). This represents a clear widening of the marginal borrowing spread in the primary market, which suggests that investors are demanding higher yields for non-IG GPU-backed credit. Let's check if there are any reports of the DDTL 5.0 secondary market trading prices or spreads. Let's do a search.

Let's look at the Electron Economics Substack article: "AI Infra financing: six weeks after 'Who Wears the Risk,' three..." to see if we can find secondary market pricing or details. Let's do a search on Electron Economics Substack.

This is phenomenal. Let's fetch the Electron Economics article: `Everything under the GPU lasts longer than the contract that pays for it`. This is exactly what we need for secondary market pricing and spreads of CoreWeave's syndicated debt! I'll use `web_fetch` on `https://electroneconomics.substack.com/p/everything-under-the-gpu-lasts-longer`.

Ah, the web search budget is spent, but I have already retrieved some very valuable quotes in the preview: - "CoreWeave’s own high yield paper has traded below par above 11 percent." - "CoreWeave’s July term loan marketed at SOFR plus 425 to 450 and closed at SOFR plus 550 at 97, an all-in 10.44 percent, with a 1.35 times debt service coverage test, US$112.5 million of minimum liquidity and full amortization." This refers to the DDTL 5.5 facility! Wait! Let's examine this carefully: - The "July term loan" (which closed in August, DDTL 5.5) was marketed at SOFR + 4.25% to 4.50% but closed at **SOFR + 5.50% at a discount price of 97** (below par!), resulting in an all-in yield of **10.44%**. - It features a **1.35x debt service coverage (DSCR) test**, **$112.5 million of minimum liquidity**, and **full amortization**. - CoreWeave's own high-yield paper (unsecured notes) has traded below par at yields **above 11%**. This is a massive signal of credit tightening! The spread widened significantly during syndication from the telegraphed SOFR + 4.25%-4.50% to SOFR + 5.50%, and it priced at a discount of 97. Let's see if we can resolve the CoreWeave watch: - `CoreWeave DDTL 5.0 Secondary Market Pricing Track`: "Monitor secondary market trading prices and credit spreads of CoreWeave's $3.1 billion syndicated DDTL 5.0 facility. A drop in secondary trading prices below par or a widening of the spread over SOFR beyond 4.50% would signal a demand softening or credit deterioration in the neocloud sector." Wait! The DDTL 5.5 facility, which is a point-release of the DDTL 5.0 facility (same BB+ rating, same template, but closed in August 2026), priced at SOFR + 5.50% at 97 (below par), which represents a widening of 100 bps over DDTL 5.0's SOFR + 4.50% launch price and a pricing discount. This represents a clear signal of credit tightening and spread widening for the neocloud sector. Let's organize our findings. We have two main areas of new, highly concrete evidence from this cycle: 1. **Oracle's Q1 Fiscal 2027 Earnings, Capex Surge, and Job Cut Escalation**: - Oracle reported Q1 FY27 results on September 10, 2026. - Capex surged to **$28.50 billion** (or $28 billion cash capex) for the single quarter ended August 31, 2026, compared to $8.5 billion in the same quarter last year. This led to a negative free cash flow of **-$5.40 billion** (or -$5 billion) for the quarter. - Despite strong revenue growth of **29.6% YoY to $19.34 billion** (OCI revenue accelerated 121% to $7.4 billion, and CPU/GPU revenue within OCI was up 151% to $6.5 billion) and a massive RPO (backlog) of **$664 billion** (up from $638B last quarter), Oracle is facing a severe AI-led cash crunch. - To manage this, Oracle filed a regulatory disclosure on September 11, 2026, increasing its "2026 Restructuring Plan" budget by **$700 million to a total of $2.8 billion**, mainly for severance and job cuts. This follows Oracle shedding **21,000 roles** (13% of its workforce) over the past year. - Oracle's total debt stands at **$155.93 billion**, with a high Debt/EBITDA of **4.6x**. - Co-founder Larry Ellison also canceled a planned $7.5 billion stock sale (10b5-1 plan to sell 50 million shares) on September 12-13, 2026. - This perfectly captures the divergence between the spending narrative (booming revenue and backlog) and the cash flows underneath it (negative FCF, restructuring, cash crunch). - This updates the note: `oracle-debt-and-capex-debacle` (Oracle's High-Yield Leverage and Soaring Capex Fail to Convince Bond and Stock Markets). 2. **CoreWeave's Extreme Leverage, Cash Burn, and Tightening Delayed-Draw Debt Terms**: - CoreWeave's total debt has ballooned to **$35.6 billion** as of June 30, 2026 (per its Q2 2026 Form 10-Q). - CoreWeave's latest financing, the **$2.6 billion DDTL 5.5 facility** (closed August 10, 2026, issued via CoreWeave Financing DDTL V-V, LLC), represents a major credit tightening event. - While DDTL 5.0 (May 18, 2026, $3.1B) priced at SOFR + 4.50%, DDTL 5.5 was marketed at SOFR + 4.25%-4.50% but closed at **SOFR + 5.50% at a discount price of 97** (below par), resulting in an all-in yield of **10.44%**. - This represents a massive **100 bps spread widening** in less than three months for the non-investment-grade customer template (rated BB+ by Fitch / Ba2 by Moody's). - DDTL 5.5 features a **1.35x debt service coverage ratio (DSCR) test** (tighter than DDTL 4.0's 1.15x), a **$112.5 million minimum liquidity covenant**, and **full amortization**. - CoreWeave's unsecured high-yield notes are now trading below par at yields **above 11%**. - There is a major structural risk: the DDTL 5.5 has a ~5-year maturity, which is longer than the ~3-year average length of the customer contracts backing it, meaning lenders are taking substantial contract-renewal and residual-value risk. - Customer concentration is extreme: the top two customers represented **65% of Q1 2026 revenue**, and three clients represented **72% of Q2 2026 revenue**. - This updates the note: `coreweave-liquidity-crisis-and-syndicated-debt` (CoreWeave's Extreme Leverage, Cash Burn, and Tightening Delayed-Draw Debt Terms). Let's look at the Vistra junior subordinated notes offering. - Vistra priced a **$1.5 billion** junior subordinated notes offering on September 11, 2026, consisting of $850 million Series A notes (7.00% interest rate, due 2057) and $650 million Series B notes (7.25% interest rate, due 2057). - The proceeds will be used to redeem Vistra's outstanding 8.0% Series A and 7.0% Series B Fixed-Rate Reset Cumulative Perpetual Preferred Stock following their reset dates in October and December 2026. - This is a refinancing of expensive preferred equity with junior debt, which is leverage-neutral but increases total reported debt. Vistra's total debt stands at **$19.89 billion** as of June 30, 2026, with a Debt/EBITDA of **3.0x**. - This relates to the utility/power buildout layer, showing how utilities are relying on long-dated junior subordinated debt to manage their capital-intensive expansion. - This updates the note: `utility-leverage-and-power-buildout-exposure` (Power Buildout and Utility Exposure to AI Capital Spending Slowdown). Let's check if there is a note for Vistra / Power Buildout utility exposure. Yes! `utility-leverage-and-power-buildout-exposure`. Let's write these three notes. First, let's write the note for Oracle. Slug: `oracle-debt-and-capex-debacle` Title: `Oracle's High-Yield Leverage and Soaring Capex Fail to Convince Bond and Stock Markets` -> Actually, let's keep the slug `oracle-debt-and-capex-debacle` and update the content. Wait, the title of the original note is `Oracle's High-Yield Leverage and Soaring Capex Fail to Convince Bond and Stock Markets`. Let's update the body. Let's write the note for CoreWeave. Slug: `coreweave-liquidity-crisis-and-syndicated-debt` Title: `CoreWeave's Extreme Leverage, Cash Burn, and Tightening Delayed-Draw Debt Terms` Let's write the note for Vistra. Slug: `utility-leverage-and-power-buildout-exposure` Title: `Power Buildout and Utility Exposure to AI Capital Spending Slowdown` Let's prepare the bodies of the notes. Let's make sure we ground them in concrete evidence, use verbatim quotes, name and link the actual things, and include a Sources section. Wait, let's write the Oracle note first. Let's see if we can find the URL for Oracle's Q1 FY27 earnings or filings. We have: - Yahoo Finance article: `https://finance.yahoo.com/markets/stocks/articles/oracle-corp-orcl-q1-2027-090101356.html` - Fortune transcript: `https://fortune.com/company/oracle/earnings/q1-2027/` - Bloomberg/Yahoo article on job cuts: `https://finance.yahoo.com/technology/ai/articles/oracle-spend-additional-700-million-211329828.html` - Seeking Alpha restructuring filing article: `https://seekingalpha.com/news/4642249-oracle-reveals-700m-in-additional-restructuring-costs-tied-to-layoffs` - CNBC article on past job cuts: `https://www.cnbc.com/2026/06/23/oracle-ai-job-cuts-layoffs-21000.html` Let's write the Oracle note `oracle-debt-and-capex-debacle`. Let's draft it. Wait! I must remember: NO YAML frontmatter. The first line must be `# <Title>`. Let's do that.

Now I will write the note for CoreWeave. Slug: `coreweave-liquidity-crisis-and-syndicated-debt` Title: `CoreWeave's Extreme Leverage, Cash Burn, and Tightening Delayed-Draw Debt Terms` Let's double-check the sources we have: - Dave Friedman Substack: `https://davefriedman.substack.com/p/coreweave-piles-on-more-debt` - Dave Friedman Substack: `https://davefriedman.substack.com/p/coreweave-issued-yet-more-debt` - Tech Insider article: `https://tech-insider.org/volta-ai-cloud-deal-coreweave-debt-2026/` - Electron Economics Substack: `https://electroneconomics.substack.com/p/everything-under-the-gpu-lasts-longer` - CoreWeave SEC 8-K filings for DDTL 5.0 and 5.5: `https://content.edgar-online.com/ExternalLink/EDGAR/0001769628-26-000236.html` and `https://www.sec.gov/Archives/edgar/data/1769628/000176962826000357/ex991pr.htm` Let's write this note. I will use the slug `coreweave-liquidity-crisis-and-syndicated-debt` to update the existing note.

Now let's write the note for the utility/power layer. Slug: `utility-leverage-and-power-buildout-exposure` Title: `Power Buildout and Utility Exposure to AI Capital Spending Slowdown` Let's make sure we include Vistra Corp's recent financing. - Vistra priced a **$1.5 billion** registered offering of junior subordinated notes on September 11, 2026. - The offering consists of $850 million Series A junior subordinated notes due 2057 (interest rate of 7.00% per annum) and $650 million Series B junior subordinated notes due 2057 (interest rate of 7.25% per annum). - The proceeds will be used to make distributions to Vistra to fund the redemption of some or all of Vistra's outstanding 8.0% Series A and 7.0% Series B Fixed-Rate Reset Cumulative Perpetual Preferred Stock following their reset dates in October and December 2026. - Vistra's total debt stands at **$19.89 billion** as of June 30, 2026, with a Debt/EBITDA of **3.0x**. - This junior subordinated debt is a hybrid instrument that allows Vistra to refinance expensive preferred stock with long-dated unsecured debt, protecting its corporate credit rating while raising long-term capital for its utility and data center power buildout. - Let's draft this note. I will use the slug `utility-leverage-and-power-buildout-exposure` to update the existing note.

Now let's review our threads and watches to see if we can resolve any. Let's list the threads again or look at what we have: 1. `Oracle Q1 Fiscal 2027 Earnings Release and Capex/Debt Expansion Tracker` (ID `01a07af0-9f05-7743-93dd-e252cf97fa11`). This thread was a watch for Oracle's Q1 FY27 earnings on September 10, 2026. Since we have fully analyzed this and updated the Oracle note, we can resolve this thread! 2. `CoreWeave DDTL 5.0 Secondary Market Pricing Track` (ID `019f1273-ce52-7485-a89d-8b5383b6d175`). Wait, this is an ongoing thread, but we can record an event for it. Let's check what watches we have. Wait, let's call `record_event` on the Oracle watch to show it fired and resolved! The Oracle watch is thread `01a07af0-9f05-7743-93dd-e252cf97fa11`. Let's call `resolve_thread` for the Oracle thread. Wait, let's call `record_event` on the CoreWeave DDTL 5.0 thread or the Vistra thread if applicable, or Oracle. Wait! The tool `record_event` records that a watch you previously registered has fired. Let's see: `01a07af0-9f05-7743-93dd-e252cf97fa11` is of kind `watch` and title `Oracle Q1 Fiscal 2027 Earnings Release and Capex/Debt Expansion Tracker`. Let's record a watch event for the Oracle watch, and then resolve the thread. And let's record an event for CoreWeave DDTL 5.0/5.5 too if it fits, but wait: the CoreWeave DDTL 5.0 watch is an `ongoing` watch. We can record an event on it showing that DDTL 5.5 priced at SOFR + 5.50% at 97, which represents a widening of spreads. This watch is ongoing, so `resolves: false`. Let's do this. First, let's record the event for Oracle.

Now let's resolve the Oracle thread.

Now let's record an event on the CoreWeave DDTL 5.0 watch (thread ID `019f1273-ce52-7485-a89d-8b5383b6d175`). The watch is: `Monitor secondary market trading prices and credit spreads of CoreWeave's $3.1 billion syndicated DDTL 5.0 facility. A drop in secondary trading prices below par or a widening of the spread over SOFR beyond 4.50% would signal a demand softening or credit deterioration in the neocloud sector.` The event is that the newly closed DDTL 5.5 facility (a point release of DDTL 5.0) priced at SOFR + 5.50% at 97 (below par), which is a clear widening of spreads beyond 4.50% and a pricing discount. This is an ongoing watch, so `resolves: false`. Let's record this event.

Wait, let's look at the Vistra watch (thread ID `019eca5a-01d4-7c6d-ab25-99092c634700`): `Track Vistra Corp's (VST) quarterly total debt levels and leverage. The company's debt currently stands at $19.91 billion. We will monitor if total debt crosses $22 billion as it continues to finance its capital-intensive power buildout.` Currently, the total debt stands at $19.89B as of June 30, 2026, which is below $22B. The $1.5 billion offering on September 11, 2026, is intended to redeem existing preferred stock, so it is leverage-neutral and doesn't push them over $22B yet. So we won't fire/resolve this watch, but let's record an ongoing event or just leave it open as it's ongoing. Let's record an event on it to keep the user updated on the $1.5B junior subordinated notes pricing. This is ongoing, so `resolves: false`. Let's do that.

This is excellent. We have recorded events and resolved the Oracle thread. Let's see if we should record any new watches or threads. Wait! Are there any new forward-looking catalysts? - Vistra's preferred stock redemption reset dates are in October and December 2026. This is a dated catalyst. - Oracle's Q2 Fiscal 2027 earnings release (expected in December 2026). - CoreWeave's Q3 2026 earnings release and quarterly filing (currently tracked by watch `01a032d4-4261-7e6d-8d90-9225ae7a1b50` on November 12, 2026). - Let's check if there are any other upcoming catalysts. We have: - `NVIDIA Q3 Fiscal 2027 Earnings Release & Vera Rubin Ramp` on November 25, 2026. - `SpaceX Colossus Lease 90-Day Cancellation Provision Active` after December 31, 2026. - `Hyperscaler Q3 2026 Earnings and Capex Guidance Revisions` on October 28, 2026. Let's register a new watch for Vistra's preferred stock redemptions in October and December 2026, which will confirm if they successfully redeem their 8.0% Series A and 7.0% Series B preferred stock using the proceeds of the $1.5 billion junior subordinated notes. Wait, let's see if this is necessary. It's a useful detail to track the capital structure of the power buildout layer. Let's register a watch for Vistra's preferred stock redemption execution. Watch type: `dated_catalyst` Cardinality: `one_shot` Fire date: `2026-12-31T00:00:00Z` (since the reset dates are in October and December 2026). Let's call `record_watch`.

Let's call `finish_cycle` to complete this research cycle, as we have updated three core notes, resolved the Oracle earnings thread, recorded events for CoreWeave and Vistra, and registered a new watch for Vistra's preferred stock redemptions. This is a complete and successful cycle.

I have successfully completed the research cycle for this topic. Here is a summary of the findings and actions taken: ### 1. Key Findings & Note Updates * **Oracle's Q1 Fiscal 2027 Earnings & Restructuring Escalation** (Updated Note: `[[oracle-debt-and-capex-debacle]]`) * **The Cash Flow Drain:** Oracle reported Q1 FY27 earnings on September 10, 2026. Capital expenditures surged to **$28.50 billion** (up from $8.5 billion in Q1 FY26), draining operating cash flows ($23.10 billion) and pushing free cash flow to **-$5.40 billion** for the single quarter. * **The AI Restructuring Cost:** To manage this AI-led cash crunch, Oracle filed a regulatory disclosure on September 11, 2026, increasing its "2026 Restructuring Plan" budget by **$700 million to a total of $2.8 billion** (severance/job cuts). This follows Oracle shedding **21,000 roles** (13% of its workforce) over the past year. * **Growth vs. Balance Sheet:** While immediate metrics are booming (revenue up 29.6% YoY to $19.34 billion, Cloud Infrastructure up 121% to $7.4 billion, and RPO/backlog at **$664 billion**), Oracle's total debt stands at **$155.93 billion** with a high **4.6x Debt/EBITDA** leverage. Highlighting internal liquidity tensions, Larry Ellison canceled his planned $7.5 billion stock sale on September 12-13, 2026. * *This updates the thesis:* Even for hyperscalers with massive backlogs, the sheer capital intensity of the AI buildout is forcing aggressive cost-cutting in non-AI operations to prevent credit deterioration. * **CoreWeave's Extreme Leverage & Tightening Delayed-Draw Debt Terms** (Updated Note: `[[coreweave-liquidity-crisis-and-syndicated-debt]]`) * **The Debt Mountain:** CoreWeave disclosed a staggering **$35.6 billion in total debt** as of June 30, 2026 (per its Q2 2026 Form 10-Q). * **The Credit Inversion & Tightening:** Lenders are treating GPUs merely as a recovery floor, pricing the off-taker's standalone credit instead. While DDTL 4.0 ($8.5B, backed by Meta) priced at SOFR + 2.25%, the newly closed **$2.6 billion DDTL 5.5 facility** (closed August 10, 2026) priced at **SOFR + 5.50% at a discount of 97** (below par), resulting in an all-in yield of **10.44%**. This represents a **100 bps spread widening** in less than three months for the non-investment-grade template (rated BB+ by Fitch / Ba2 by Moody's). * **Restrictive Covenants & Duration Risk:** DDTL 5.5 carries tighter covenants, including a **1.35x DSCR test** (vs. 1.15x for DDTL 4.0) and a **$112.5 million minimum liquidity requirement**. Lenders are also taking on substantial duration risk: the loan matures in ~5 years, while the underlying customer contracts average only ~3 years. Customer concentration remains extreme, with three clients representing **72% of Q2 2026 revenue**. * **Power Buildout & Utility Exposure** (Updated Note: `[[utility-leverage-and-power-buildout-exposure]]`) * **Vistra's $1.5 Billion Hybrid Refinancing:** On September 11, 2026, Vistra Corp. (VST) priced a **$1.5 billion registered public offering of junior subordinated notes** due 2057 ($850M Series A at 7.00%, $650M Series B at 7.25%). The proceeds will fund the redemption of its outstanding 8.0% Series A and 7.0% Series B Perpetual Preferred Stock upon their reset dates in October and December 2026. * **The Duration Mismatch:** Vistra's total debt stands at **$19.89 billion** as of June 30, 2026 (Debt/EBITDA of 3.0x). While the transaction is leverage-neutral, it highlights the high cost of capital (7.00%-7.25%) utilities must pay. The core risk is that these 20-to-25-year physical power assets are being built and financed on the back of shorter-dated (3-to-5-year) compute contracts. If the AI spending narrative stalls, utilities will be left with massive long-dated debt loads and unhedged generating assets. --- ### 2. Thread Resolutions & Watch Events * **Resolved Thread:** * `Oracle Q1 Fiscal 2027 Earnings Release and Capex/Debt Expansion Tracker` (ID `01a07af0-9f05-7743-93dd-e252cf97fa11`) is now fully resolved with the updated findings in `[[oracle-debt-and-capex-debacle]]`. * **Recorded Watch Events:** * **CoreWeave DDTL 5.0 Secondary Market Pricing Track** (ID `019f1273-ce52-7485-a89d-8b5383b6d175`): Recorded an ongoing event showing that the newly closed DDTL 5.5 priced at SOFR + 5.50% at 97 (below par), representing a 100 bps spread widening over DDTL 5.0's launch price. * **Vistra Corp Debt and Leverage Track** (ID `019eca5a-01d4-7c6d-ab25-99092c634700`): Recorded an ongoing event detailing Vistra's $1.5 billion junior subordinated notes offering at 7.00% to 7.25% to refinance perpetual preferred stock. * **Registered New Watch:** * **Vistra Preferred Stock Redemption Tracker** (ID `01a09efc-171b-7c45-8023-c27c1d4d3c34`): A one-shot watch with a fire date of December 31, 2026, to monitor whether Vistra successfully completes the redemptions of its 8.0% Series A and 7.0% Series B preferred stock following their reset dates in October and December 2026.