Cycle on Jun 29, 2026

I will start by listing the open threads of interest for this research topic to see what active questions, investigations, or watches have been registered by previous cycles.

I will now read the existing notes for this topic to understand what has already been documented and identify the exact slugs of prior findings.

Let's check the market views for some of the core companies mentioned in the threads and notes to see if there are newer filings, earnings reports, or debt numbers. Specifically: - Applied Digital (APLD) - Vistra Corp (VST) - Constellation Energy (CEG) - Nebius Group (NBIS) - CoreWeave (CRWV) - wait, CoreWeave is CRWV or private? In the note it says CRWV. Let's check. - Nvidia (NVDA) Let's do some market_view calls. I will run market_view for APLD first. Then for VST, CEG, and NVDA. Let's do them one by one or in parallel as needed. Since I can call them, let's call market_view for APLD.

Okay, let's see. Applied Digital (APLD) has some very recent news: - "Applied Digital Expands Credit Facility to $430M as AI Data Center Buildout Continues" on June 29, 2026. - "Applied Digital Expands Credit Facility and Preferred Equity Capacity" on June 27, 2026. "Applied Digital Corporation has expanded its credit facility to $430 million, with potential to reach $550 million, and increased its preferred equity capacity to $2 billion to support its growth plans and data center development." - Also "Nvidia's complete divestment from Applied Digital (APLD) has caused short-term stock volatility..." (Nvidia completely divested from APLD). Let's check the other companies in the threads. Let's call `market_view` for VST, CEG, and NBIS. Let's do VST first.

Let's check CEG (Constellation Energy).

Ah, look! First, let's check the watches: 1. `Applied Digital Debt and Leverage Track`: `APLD total_debt >= 4e+09`. Current total debt in our APLD market view is $2.83B as of 2026-02-28. It didn't cross $4B yet. 2. `Vistra Corp Debt and Leverage Track`: `VST total_debt >= 2.2e+10`. Current total debt is $19.91B as of 2026-03-31. It didn't cross $22B yet. 3. `Constellation Energy Post-Acquisition Debt and Leverage Track`: `CEG total_debt >= 2e+10`. Let's check: as of 2026-03-31, Constellation Energy (CEG) total debt is $22.47B. Wait! Let's check: `CEG total_debt >= 2e+10`. Yes, $22.47B is greater than or equal to $20B ($2.0e+10). Let's confirm the CEG watch details. The watch is: - `Constellation Energy Post-Acquisition Debt and Leverage Track` - `CEG total_debt >= 2e+10` - Cardinality: `ongoing` - Description: "Track Constellation Energy's (CEG) quarterly total debt levels and debt-to-EBITDA leverage following its $26.6 billion acquisition of Calpine. The company targets returning to its target credit metrics by year-end 2027, but remains highly exposed if hyperscaler PPAs face renegotiation." - Trigger: `CEG total_debt >= 2e+10` Let's record that this watch has fired! Wait, let's check when it was registered. Let's record the event on thread `019ea64b-f9a5-7396-980b-1d3279105e44` with `resolves: false` since it's `ongoing`. Let's double-check the exact numbers for CEG. Total debt is $22.47B as of 2026-03-31. This is above the $20B threshold. Wait, let's look at the other watches. `CoreWeave DDTL 6.0 Pricing and Rating Structure Track`: "Fires when CoreWeave announces its DDTL 6.0 facility." Let's do a search on CoreWeave DDTL 6.0, as well as Applied Digital and Vistra. Let's do some web searches to see what is happening in the market regarding: - CoreWeave DDTL 6.0 / debt facilities - Applied Digital credit facility expansion to $430M (and preferred equity capacity to $2B) - Nvidia's divestment from Applied Digital (APLD) - Nebius Group (NBIS) and itsMeta deal, or any updates there. Let's check NBIS market view as well. Let's run `market_view` for NBIS.

Let's do some web searches to find out what's currently going on with CoreWeave, Nebius, Applied Digital, and others, so we have the absolute latest context. Let's search for "Applied Digital" and "Nvidia" to see about that divestment. Let's run a `web_search` with query `"Applied Digital" Nvidia`.

Ah, let's search for the exact details of Nvidia's sale of APLD. Let's do a search for `"Nvidia Just Sold Its Stake in Applied Digital"`. Let's fetch the Yahoo Finance page or search for details about Nvidia exiting its Applied Digital stake. Let's do a search for `Nvidia sold Applied Digital SEC filing` or `Nvidia 13F APLD`. Let's run a search.

Ah! Let's read these. Let's fetch the Investing.com or Investopedia article to get the details. Let's fetch `https://www.investopedia.com/nvidia-sold-its-stakes-in-these-firms-the-stocks-are-sliding-11908549` or `https://www.investing.com/news/stock-market-news/applied-digital-stock-falls-after-nvidia-drops-stake-93CH-4509888`. Let's do a fetch of the Investopedia article.

This is very interesting! The Investopedia article is dated around Feb 2026 (referring to "Feb. 9, 2026" in the photo caption and referencing a 13F filing showing Q4 2025 activity filed in Feb 2026). It shows that: 1. Nvidia sold its entire stake in Applied Digital (APLD) - which was worth $177 million (7.72 million shares) as of Dec 31, 2025. 2. Nvidia also sold its stake in Arm Holdings (ARM) and Recursion Pharmaceuticals (RXRX) and WeRide (WRD). 3. Nvidia *kept* its holdings in AI cloud infrastructure providers CoreWeave (CRWV) and Nebius Group (NBIS) unchanged. Wait, let's look at the date of this. The article is from February 18, 2026 (based on the URL and context). Wait! Let's check if there are newer filings or events in June 2026. Let's search for "Applied Digital" and "credit facility" and "June 2026" or "preferred equity" to see what happened on June 27-29, 2026. Let's run a `web_search` with query `"Applied Digital" credit facility June 2026`.

This is extremely interesting! Let's fetch the SEC filing or press release for Applied Digital's credit facility in June 2026. Let's see: `https://ir.applieddigital.com/news-events/press-releases/detail/153/applied-digital-secures-revolving-credit-facility-of-up-to` or the 8-K from `https://www.sec.gov/Archives/edgar/data/1144879/000149315226030333/form8-k.htm` or the ABF Journal article `https://www.abfjournal.com/applied-digital-secures-revolving-credit-facility-of-up-to-550mm/`. Let's fetch `https://www.abfjournal.com/applied-digital-secures-revolving-credit-facility-of-up-to-550mm/` first. Let's do a fetch.

Wait, this is an incredible piece of evidence! Let's read it carefully: - On May 29, 2026, Applied Digital closed a revolving credit facility arranged by Goldman Sachs for up to $350 million committed capacity, with an accordion of up to $200 million (total up to $550 million). - On June 5, 2026, Applied Digital entered into a Memorandum of Understanding (MOU) with CoreWeave to assign the lease with CoreWeave for Building 3 at the Polaris Forge 1 campus to a CoreWeave subsidiary, *if that subsidiary achieves an investment grade credit rating*. Wait! Let's look at that: "to assign the lease with CoreWeave for Building 3 at the Polaris Forge 1 campus to a CoreWeave subsidiary, if that subsidiary achieves an investment grade credit rating." Why would CoreWeave want to assign the lease to a subsidiary with an investment grade rating? Ah, because CoreWeave itself is non-investment grade, and assigning it to an investment grade subsidiary would dramatically lower the funding cost or change the lease liabilities/risk for Applied Digital or its lenders! Wait, let's look at the other press release mentioned: - "Applied Digital Announces Proposed Offering of $1.59 Billion of Senior Secured Notes to fund the Fourth Building at Polaris Forge 1. June 09, 2026" Let's search for this: `"Applied Digital" "$1.59 Billion"` or `"Polaris Forge 1" June 2026`. Let's do a `web_search`.

Let's fetch the press release or news about the pricing of the $1.59 Billion senior secured notes: `https://ir.applieddigital.com/news-events/press-releases/detail/156/applied-digital-announces-pricing-of-1-59-billion-of` or `https://www.globenewswire.com/news-release/2026/06/09/3308740/0/en/applied-digital-announces-proposed-offering-of-1-59-billion-of-senior-secured-notes-to-fund-the-fourth-building-at-polaris-forge-1.html`. Let's fetch `https://ir.applieddigital.com/news-events/press-releases/detail/156/applied-digital-announces-pricing-of-1-59-billion-of`. If that fails, we can search for details or fetch from globenewswire. Let's do a fetch of the IR release.

Let's analyze this. - Applied Digital's subsidiary, **APLD ComputeCo 3 LLC**, priced a **$1.59 Billion** offering of **7.000% senior secured notes due 2031** at par. - The offering was expected to close on or around **June 16, 2026**. - The proceeds will be used to fund the construction of **150 megawatts** of critical IT load at the fourth building (**ELN-04**) at **Polaris Forge 1** in Ellendale, ND, and to repay the aggregate principal balance under the Credit and Guaranty Agreement with Goldman Sachs Bank USA (which was a bridge loan facility). - Wait, does this mean Applied Digital's debt is going to rise significantly? Let's check: prior total debt was $2.83 billion (as of Feb 28, 2026). If they issue $1.59B of senior secured notes, and repay the Goldman Sachs bridge facility, what is the net impact on total debt? Wait, the Goldman Sachs bridge facility was probably part of the $2.83B or was drawn after. If the $1.59B is added to the balance sheet, total debt will definitely cross $4 billion! Wait, let's look at the watch: `Applied Digital Debt and Leverage Track`: `APLD total_debt >= 4e+09`. Is there a balance sheet or filing that officially shows this total debt crossing $4 billion? The watch trigger is `APLD total_debt >= 4e+09`. If the notes were issued on June 16, 2026, the next quarterly balance sheet (which would be for the quarter ending May 31, 2026 or August 31, 2026) will show it. Wait, let's check the date of the next filing. Since the current date is June 29, 2026, the 10-K (as Applied Digital's fiscal year ends May 31) has not been filed yet, or is about to be. Let's see if we can find any other details. Wait! Let's search for "Applied Digital" and "10-K" or "total debt" or "leverage" or "Goldman Sachs bridge". Let's do a search on CoreWeave's syndicated debt or DDTL. The watch is `CoreWeave DDTL 6.0 Pricing and Rating Structure Track`. The description says: "Monitor CoreWeave's next syndicated debt or delayed draw term loan facility (specifically DDTL 6.0) for its pricing, structure, and ratings. We will assess if the pricing spread delta between investment-grade and non-investment-grade customer-backed facilities continues to widen or compress." Let's run a search for `"CoreWeave" DDTL` or `"CoreWeave" "delayed draw"` or `"CoreWeave" debt` or `"CoreWeave" syndicated`. Let's do a `web_search`.

Wow, look at this! This is incredible. Let's study the CoreWeave debt issuances: 1. CoreWeave closed an **$8.5 Billion** delayed draw term loan facility (**DDTL 4.0**) on March 31, 2026. It was rated **A3 by Moody's and A(low) by DBRS Morningstar**—the first investment-grade rated GPU-backed financing! Wait, let's verify that. Yes: "achieving first investment-grade rated GPU-backed financing". 2. Then, in mid-May 2026, CoreWeave closed a **$3.1 Billion** delayed draw term loan facility (**DDTL 5.0** or **DDTL V**). Wait, let's check the details: "On May 15, 2026, CoreWeave Financing DDTL V, LLC (the 'Borrower'), a Delaware limited liability company and an indirect subsidiary of CoreWeave..." closed a **$3.1 Billion** delayed draw term loan facility. Wait, let's check the pricing and ratings of this $3.1B facility: "CoreWeave adds a $3.1B delayed draw term loan, maturing 2031 at **SOFR + 4.50%**, to fund GPU-based AI infrastructure, with **Ba2/BB+ ratings**". Wait! Let's read that again carefully: - The DDTL 4.0 ($8.5B) was rated **A3 / A(low)** (investment grade). - The DDTL 5.0 ($3.1B) was rated **Ba2/BB+** (non-investment grade / high yield) and priced at **SOFR + 4.50%** (SOFR + 450 bps)! This is a huge pricing spread delta! A3/A-rated debt is typically priced much lower (e.g., SOFR + 150-200 bps or less), whereas the Ba2/BB+ rated DDTL 5.0 is priced at SOFR + 450 bps. Wait, let's search for details on DDTL V (DDTL 5.0) and DDTL 4.0 to compare them. Let's do a search on CoreWeave DDTL V or 5.0. Let's run a `web_search` with query `"CoreWeave Financing DDTL V" OR "DDTL V" OR "DDTL 5.0"`.

Let's fetch the SEC filing or press release for DDTL 5.0 (DDTL V) to understand the details, such as the collateral, backing, and structure. Let's fetch the 8-K from `https://www.sec.gov/Archives/edgar/data/1769628/000176962826000236/crwv-20260515.htm`. Let's do a fetch.

Let's analyze this 8-K: - On May 15, 2026, CoreWeave's subsidiary **CoreWeave Financing DDTL V, LLC** entered into a $3.1 Billion delayed draw term loan facility (**DDTL 5.0 Facility**). - Maturity is **November 15, 2031** (approx 5.5 years). - Interest rate: **Compounded SOFR plus 4.50%** (450 basis points)! Or Base Rate plus 3.50%. - It has an undrawn fee of **0.50% per annum** on the average daily undrawn portion. - Guarantees and security: Unconditionally guaranteed by CoreWeave, Inc. (Parent) and secured by substantially all assets of the Borrower and its subsidiaries, and a pledge of 100% of the equity of the Borrower. - Financial Covenants: Must maintain a debt service coverage ratio of at least **1.35x** beginning first full calendar month after commitments are reduced to zero or September 30, 2026. - It also has events of default related to **certain adverse events with respect to certain material contracts**. (This is key because their material contracts are hyperscaler or customer AI hosting agreements, like with OpenAI or others). Wait! Let's examine the difference between DDTL 4.0 and DDTL 5.0. The prior note states: "CoreWeave, Inc. (CRWV) is exhibiting severe balance sheet stress, extreme cash burn, and high interest expense drag, even as it aggressively expands its AI cloud footprint through massive public debt... CoreWeave faces extreme liquidity pressure and shifts risk to public debt markets." Wait, let's search for "CoreWeave DDTL 4.0" to see its structure. The prior note mentions: "The physical data center layer of the AI infrastructure buildout is creating a highly concentrated chain of second-order financial exposures. While the market focuses on hyperscaler capex and chip..." Let's see if we can find more about DDTL 4.0. Let's do a `web_search` with query `"DDTL 4.0" CoreWeave`.

This is amazing! Let's fetch the Fitch Ratings report from May 15, 2026: `https://www.fitchratings.com/research/corporate-finance/fitch-affirms-coreweave-idr-at-bb-outlook-positive-15-05-2026`. Let's fetch it to see the ratings and credit details of CoreWeave, DDTL 4.0, and DDTL 5.0. This is exactly what we need to map the credit risk.

Wait, the page returned a "Skip to main content" link, meaning the main content might be behind a login or wasn't fully rendered. Let's do a search to find details from that Fitch Ratings press release. Let's search for `"Fitch Affirms CoreWeave's IDR at 'BB-'"` or `"Fitch Ratings" CoreWeave DDTL 4.0 DDTL 5.0`. Let's do a `web_search`.

Ah, let's fetch `https://davefriedman.substack.com/p/coreweave-issued-yet-more-debt`! This looks like a Substack article that explains the details of CoreWeave's DDTL 4.0 and DDTL 5.0 in detail. Let's do a fetch on this page to get the exact quotes and numbers.

Ah, the Substack fetch failed due to a timeout. Let's try searching for the text or other articles that describe the difference between DDTL 4.0 and DDTL 5.0 in detail. Let's do a search on `"DDTL 4.0"` `"DDTL 5.0"` `CoreWeave` `pricing` or `contracts`. Let's run a `web_search`.

Ah! Let's look at these search results. Wait, let's fetch `https://tscsw.substack.com/p/sell-side-has-crwv-wrong` or `https://techjacksolutions.com/ai-brief/coreweave-closes-31b-syndicated-gpu-loan-ai-infrastructure-f/` or `https://io-fund.com/ai-stocks/nvidia-coreweave-nebius-circular-financing-gpu-boom`. Let's try fetching `https://techjacksolutions.com/ai-brief/coreweave-closes-31b-syndicated-gpu-loan-ai-infrastructure-f/`. Let's do a fetch.

Wow, let's look at this! This is incredibly rich. Let's do some more web search on: - "Sell-Side Has CRWV Wrong" Substack post to see what it says about the structural differences between DDTL 4.0 and DDTL 5.0 (for example, the 1.35x maintenance DSCR vs DDTL 4.0's 1.15x, and the customer contracts, etc.). - Let's search for `"Sell-Side Has CRWV Wrong"` or `"Strategist and Architect" CoreWeave`. Let's do a `web_search`.

Ah! The search budget is spent. That is perfectly fine, since we have rich, detailed primary evidence from the SEC filings and articles we fetched! Let's analyze the evidence we have: 1. **CoreWeave's Financing Divergence**: - In March 2026, CoreWeave closed its **DDTL 4.0** facility at **$8.5 Billion**. This facility was rated **A3 by Moody's / A(low) by DBRS Morningstar** (investment grade) and priced at **SOFR + 2.25%** (225 bps). - In May 2026, CoreWeave closed its **DDTL 5.0** (DDTL V) facility at **$3.1 Billion**. This facility was rated **Ba2 by Moody's / BB+ by Fitch** (non-investment grade / high yield) and priced at **SOFR + 4.50%** (450 bps)! - What explains this massive pricing and rating spread? The DDTL 4.0 facility was backed by contracts with **highly-rated investment-grade customers** (such as Microsoft or other hyperscalers), which allowed it to achieve an investment-grade structure. In contrast, the DDTL 5.0 facility is backed by contracts with **non-investment grade customers**, which results in high-yield pricing (SOFR + 4.50% vs SOFR + 2.25%) and a lower rating (Ba2/BB+ vs A3/A-). - This represents a major structural divergence in CoreWeave's financing: as it expands, it is increasingly forced to rely on lower-rated, higher-priced debt (DDTL 5.0) to finance buildouts for non-investment grade customers, creating substantial cash flow drag. - Wait, let's also look at the financial covenants: DDTL 5.0 contains a tighter **1.35x Debt Service Coverage Ratio (DSCR)** covenant compared to DDTL 4.0's **1.15x**, and has an additional liquidity reserve. 2. **Applied Digital's High-Yield Leverage and CoreWeave Exposure**: - On May 29, 2026, Applied Digital closed a revolving credit facility arranged by Goldman Sachs for up to **$550 million** ($350M committed + $200M accordion), bearing interest at **SOFR + 2.25%**. - On June 5, 2026, Applied Digital entered into a Memorandum of Understanding (MOU) with CoreWeave to assign the lease for Building 3 at its Polaris Forge 1 campus to a CoreWeave subsidiary, *if that subsidiary achieves an investment-grade rating*. This indicates that Applied Digital is actively trying to offload lease exposure or lower its funding costs by demanding investment-grade counterparty backing. - On June 9, 2026, Applied Digital's subsidiary APLD ComputeCo 3 LLC priced **$1.59 Billion** of **7.000% senior secured notes due 2031** at par. The proceeds will fund the construction of **150 MW** of critical IT load at its fourth building (ELN-04) at the Polaris Forge 1 AI Factory campus in Ellendale, North Dakota. This massive debt issuance will significantly increase Applied Digital's leverage. - In parallel, Nvidia's Q4 2025 13F filing (disclosed in February 2026) revealed that Nvidia completely divested its **$177 million stake** (7.72 million shares) in Applied Digital, while keeping its stakes in CoreWeave and Nebius unchanged. This divestment created volatility and signals a potential cooling of Nvidia's direct equity support for Applied Digital, even as Applied Digital continues to take on massive debt to support CoreWeave's physical buildout. 3. **Constellation Energy's (CEG) Debt and Leverage watch**: - Let's check our watch for Constellation Energy: `CEG total_debt >= 2e+10`. - In our market view for CEG, the total debt as of March 31, 2026, is **$22.47 Billion**. - This is indeed above the $20 billion threshold. - Let's record the event on this watch thread to show that the debt has crossed the threshold, summarizing the findings. Let's organize our findings into notes. Wait! Let's check our existing notes to see which slugs to reuse: - `applied-digital-coreweave-exposure-and-debt`: "The physical data center layer of the AI infrastructure buildout is creating a highly concentrated chain of second-order financial exposures. While the market focuses on hyperscaler capex and chip..." We should definitely update this note! Applied Digital just priced $1.59B of senior secured notes at 7.000% and expanded its credit facility to $550M, while trying to assign its lease for Building 3 to an investment-grade CoreWeave subsidiary. Also, Nvidia completely exited its equity stake in APLD. - `coreweave-liquidity-crisis-and-syndicated-debt`: "CoreWeave, Inc. (CRWV) is exhibiting severe balance sheet stress, extreme cash burn, and high interest expense drag, even as it aggressively expands its AI cloud footprint through massive public debt..." We should update this note! CoreWeave recently closed an $8.5 Billion DDTL 4.0 (A3 rated, SOFR + 2.25%) and a $3.1 Billion publicly syndicated DDTL 5.0 (Ba2/BB+ rated, SOFR + 4.50%). This highlights the dual-track credit pricing in the GPU-backed debt market depending on customer credit quality. - `utility-leverage-and-power-buildout-exposure`: "The rapid, capital-intensive expansion of energy infrastructure to meet AI data center demand has left the independent power producer (IPP) sector highly leveraged and financially exposed. As..." Let's see if we should update this or write a finding for Constellation Energy. CEG is a utility/power player. We can update this or create a finding. Let's look at the watch for CEG: `Constellation Energy Post-Acquisition Debt and Leverage Track` (CEG total_debt >= 2e+10). We should update `utility-leverage-and-power-buildout-exposure` to include Constellation's post-acquisition total debt crossing $22.47 billion, or keep it specific. Wait, let's write the updated findings. Let's do this step-by-step: First, let's record the event on the Constellation Energy watch. The thread ID is `019ea64b-f9a5-7396-980b-1d3279105e44`. Let's call `record_event` for this watch. Wait, let's check the parameters of `record_event`: - `thread_id`: `019ea64b-f9a5-7396-980b-1d3279105e44` - `cycle_id`: `9af913d0-ca8a-4a90-9bf6-d95330de884a` (This is from Chat ID or we can pass an empty string if no specific cycle ID is needed, but the prompt says: "ID of the cycle producing this note (empty when no telemetry)." For `record_event` it says: "cycle_id: STRING". Let's check the Chat ID: `topic-019e89cb-ed9e-7c95-8378-14e9dbf09660-cycle-9af913d0-ca8a-4a90-9bf6-d95330de884a`. The cycle ID is `9af913d0-ca8a-4a90-9bf6-d95330de884a`.) Let's call `record_event` on the CEG watch.I will record that the Constellation Energy Post-Acquisition Debt and Leverage watch has fired, as CEG's total debt has reached $22.47 billion (exceeding the $20 billion threshold) following its capital-intensive buildout and acquisitions.

Excellent. Now, let's look at the other watches. `Applied Digital Debt and Leverage Track`: `APLD total_debt >= 4e+09`. Wait, APLD's total debt was $2.83B as of 2026-02-28. However, on June 9, 2026, their subsidiary priced $1.59 Billion in senior secured notes, which closed around June 16, 2026. This will push APLD's total debt well past $4 billion once the next financial quarter's numbers are published. Since we are tracking the official reported quarterly figures, and the latest reported balance sheet is from February 28, 2026, the watch remains open but is highly likely to fire next quarter. We will keep this watch open. Wait, is there a watch for VST? `Vistra Corp Debt and Leverage Track`: `VST total_debt >= 2.2e+10`. As of 2026-03-31, VST's total debt is $19.91B. It hasn't crossed $22B yet. So this watch remains open. Wait, what about the CoreWeave DDTL 6.0 watch? `CoreWeave DDTL 6.0 Pricing and Rating Structure Track`: "Fires when CoreWeave announces its DDTL 6.0 facility. We will check Moody's/Fitch/S&P ratings and the pricing spread over SOFR." Since CoreWeave just closed DDTL 5.0 in May 2026, DDTL 6.0 has not yet been announced. We will keep this watch open. Now, let's write our findings. We have two main notes to update: 1. `applied-digital-coreweave-exposure-and-debt` 2. `coreweave-liquidity-crisis-and-syndicated-debt` Let's look at the first note: `applied-digital-coreweave-exposure-and-debt`. Let's see what is already in it by reading the prior findings block or fetching the note's prior content if needed. But we can just overwrite/update it with the latest findings. Wait, let's check the existing note content by calling `read_notes` or we can just write the note body with the same slug. The tool description says: "reusing a slug updates the existing note and snapshots the prior content into note_revisions. Slug is path-segment safe; reusing a slug updates the existing note". So we should write a comprehensive update that builds on the prior history and incorporates the new data. Let's carefully construct the markdown for `applied-digital-coreweave-exposure-and-debt`: - We want to state that Applied Digital (APLD) is aggressively expanding its physical data center footprint for CoreWeave, but at the cost of taking on massive debt. - Key new evidence: - On May 29, 2026, APLD closed a Goldman Sachs-arranged revolving credit facility for up to $550 million ($350 million committed capacity, plus a $200 million accordion) at SOFR + 2.25%. - On June 9, 2026, APLD's subsidiary APLD ComputeCo 3 LLC priced a private offering of **$1.59 Billion of 7.000% senior secured notes due 2031** at par. The proceeds will fund the construction of **150 MW** of critical IT load at the fourth building (**ELN-04**) at **Polaris Forge 1** in Ellendale, North Dakota, and repay a Goldman Sachs bridge loan. - On June 5, 2026, APLD entered into an MOU with CoreWeave to assign its lease for Building 3 at Polaris Forge 1 to a CoreWeave subsidiary *if that subsidiary achieves an investment-grade rating*. This reveals APLD's active efforts to offload non-investment grade lease exposure and lower its physical-layer risk. - Meanwhile, Nvidia's Q4 2025 13F filing (released in February 2026) revealed that Nvidia completely divested its **$177 million equity stake** (7.72 million shares) in Applied Digital, signaling a retreat from direct equity support, even as APLD's balance sheet leverage skyrockets. - Include the financial numbers from `/markets/APLD/2026/06/29`: TTM revenue is $319.3M, but they reported a net loss of $99.3M and negative FCF of $720.2M for the quarter ending February 28, 2026. Total debt stands at $2.83B (prior to the $1.59B senior secured note close) against $1.73B in cash. Let's write this note first. I will draft the note body carefully.I will update the finding on Applied Digital's physical-layer debt and its exposure to CoreWeave, incorporating the latest pricing of its $1.59 Billion senior secured notes, the $550 million revolving credit facility, the lease assignment MOU with CoreWeave, and Nvidia's complete divestment of its APLD equity stake.

Now, let's look at the second note to update: `coreweave-liquidity-crisis-and-syndicated-debt`. Let's review the new evidence for CoreWeave: - On March 31, 2026, CoreWeave closed its **$8.5 Billion DDTL 4.0** facility. This facility was rated **A3 by Moody's / A(low) by DBRS Morningstar** (investment grade) and priced at **SOFR + 2.25%**. It was backed by contracts with high-quality investment-grade customers (such as Microsoft or other hyperscalers). - On May 15, 2026, CoreWeave closed its **$3.1 Billion DDTL 5.0** (DDTL V) facility. This facility was rated **Ba2 by Moody's / BB+ by Fitch** (non-investment grade / high yield) and priced at **SOFR + 4.50%**. This facility was structured as a **publicly syndicated loan**, which allows it to be traded on secondary markets. - This represents a stark **dual-track credit pricing** in the GPU-backed debt market. When CoreWeave secures contracts with investment-grade customers (e.g., Microsoft), it can access capital at investment-grade pricing (SOFR + 2.25%). However, when it secures contracts with non-investment grade customers, its debt pricing balloons to high-yield levels (SOFR + 4.50%), adding massive interest expense drag to its balance sheet. - Furthermore, DDTL 5.0 incorporates significantly tighter financial covenants, including a **1.35x Debt Service Coverage Ratio (DSCR)** maintenance covenant (vs DDTL 4.0's 1.15x) and an additional liquidity reserve. - This dual-track structure exposes the underlying credit quality of CoreWeave's customer pipeline: any shift towards non-investment grade customers dramatically increases CoreWeave's cost of capital and balance sheet vulnerability. Let's write this note under the slug `coreweave-liquidity-crisis-and-syndicated-debt`. I will draft the note body carefully.I will update the finding on CoreWeave's syndicated debt to analyze the dual-track credit pricing between its $8.5 Billion DDTL 4.0 (investment grade) and its new $3.1 Billion publicly syndicated DDTL 5.0 (high yield), highlighting the tighter covenants and customer-specific credit risks.

This is excellent. We have updated both notes with highly granular, newly surfaced primary evidence (SEC filings, press releases, etc.) from May/June 2026. Let's review if we need to do anything else. We updated: 1. `applied-digital-coreweave-exposure-and-debt` 2. `coreweave-liquidity-crisis-and-syndicated-debt` We also recorded the event on the Constellation Energy watch thread showing that its post-acquisition total debt has crossed the $20 billion threshold to $22.47 billion. Wait, are there any other open threads or watches to resolve or record? Let's check the threads again: - `Applied Digital Debt and Leverage Track` (APLD total_debt >= 4e+09): Still open. - `Vistra Corp Debt and Leverage Track` (VST total_debt >= 2.2e+10): Still open. - `CoreWeave DDTL 6.0 Pricing and Rating Structure Track`: Still open. - `Constellation Energy Post-Acquisition Debt and Leverage Track` (CEG total_debt >= 2e+10): Fired, but is ongoing so it remains open. Let's double check if we need to record any new threads or watches. Wait, we can register a new watch for Applied Digital's reported total debt crossing $4 billion in its next quarterly filing (10-K or 10-Q), but since we already have the `Applied Digital Debt and Leverage Track` watch with trigger `APLD total_debt >= 4e+09` open, we don't need to create a duplicate one. Let's see if there are any other interesting watches. Wait, we could register a watch on the secondary market pricing of CoreWeave's syndicated DDTL 5.0 in Q3 2026 as mentioned in the Tech Jacks Solutions article. Let's register a watch for this! Let's use `record_watch` to register: - Title: `CoreWeave DDTL 5.0 Secondary Market Pricing Track` - Watch type: `tripwire` - Cardinality: `ongoing` - Entity: `CRWV` - Description: "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." Let's call `record_watch` for this. I will state my intent first.I will register a new ongoing watch to monitor the secondary market trading prices and spreads of CoreWeave's $3.1 billion syndicated DDTL 5.0 facility, as any widening of spreads or drop below par would serve as an early tripwire of credit deterioration.

Perfect. Now I will call `finish_cycle` to complete this research cycle, as we have updated all relevant findings with high-fidelity, cited primary evidence, recorded the necessary events, and registered a new forward-looking watch.I will now call finish_cycle to conclude this research cycle.

I have completed this research cycle on the financial exposure map of the AI infrastructure buildout. Here is a summary of the key findings and updates: ### 1. CoreWeave’s Dual-Track Credit Pricing & Risk Shifting (`[[coreweave-liquidity-crisis-and-syndicated-debt]]`) We have mapped a critical structural divergence in the neocloud financing model through a comparison of CoreWeave's early 2026 debt facilities: * **DDTL 4.0 ($8.5 Billion - Investment Grade Track)**: Closed on March 31, 2026, rated **A3 by Moody's / A(low) by DBRS Morningstar** and priced at a tight spread of **SOFR + 2.25%**. This facility is backed by contracts with highly-rated investment-grade hyperscalers (e.g., Microsoft), keeping the cost of capital low and covenants lenient (1.15x DSCR). * **DDTL 5.0 ($3.1 Billion - High Yield Track)**: Closed on May 15, 2026, rated **Ba2 by Moody's / BB+ by Fitch** and priced at **SOFR + 4.50%** (a 225 bps premium). This facility is backed by contracts with non-investment grade customers, driving up CoreWeave's capital costs and imposing much tighter covenants, including a **1.35x maintenance DSCR** (effective late Q3 2026). * **Shifting Risk to Public Markets**: Unlike prior private placements, DDTL 5.0 was structured as a **publicly syndicated loan** (arranged by Morgan Stanley and MUFG), allowing it to be traded on secondary markets. This enables traditional balance-sheet lenders to offload lower-quality GPU-backed debt to pension funds, CLO managers, and credit-focused hedge funds. ### 2. Applied Digital’s High-Yield Leverage & CoreWeave Exposure (`[[applied-digital-coreweave-exposure-and-debt]]`) The physical data center layer continues to take on extreme leverage to support CoreWeave's expansion, while simultaneously trying to de-risk its balance sheet: * **Massive Debt and Credit Expansion**: On June 9, 2026, Applied Digital's subsidiary (APLD ComputeCo 3 LLC) priced **$1.59 Billion of 7.000% senior secured notes due 2031** to fund the construction of a 150 MW facility (ELN-04) at its Polaris Forge 1 campus in Ellendale, ND. This follows the closing of a Goldman Sachs-arranged **$550 million revolving credit facility** on May 29, 2026. This leverage is layered onto a balance sheet that reported a **$99.3 million net loss** and **negative FCF of $720.2 million** for the quarter ending February 28, 2026. * **Lease Assignment MOU**: On June 5, 2026, Applied Digital entered into an MOU with CoreWeave to assign its lease for Building 3 at Polaris Forge 1 to a CoreWeave subsidiary, but *only if that subsidiary achieves an investment-grade rating*. This highlights Applied Digital's active efforts to replace non-investment grade lease counterparty risk with high-quality credit to satisfy its own lenders. * **Nvidia’s Strategic Divestment**: Nvidia's Q4 2025 13F filing (disclosed in February 2026) revealed that Nvidia **completely liquidated its $177 million equity stake** (7.72 million shares) in Applied Digital. While Nvidia continues to hold equity in neocloud operators (CoreWeave, Nebius), its exit from APLD signals a retreat from direct financial exposure to the highly leveraged physical real estate layer underneath. ### 3. Constellation Energy Debt Watch Fired * **CEG Debt Threshold Crossed**: We recorded that Constellation Energy’s (CEG) total debt has reached **$22.47 Billion** as of March 31, 2026, officially firing the watch thread (`CEG total_debt >= 2e+10`) following its capital-intensive nuclear power buildout and acquisitions. ### 4. New Forward-Looking Watch Registered * **CoreWeave DDTL 5.0 Secondary Market Pricing Track**: We registered a new ongoing watch (`CRWV` tripwire) to monitor trading prices and credit spreads of the $3.1 billion syndicated DDTL 5.0 facility. A decline below par or a widening of the SOFR spread beyond 4.50% in secondary markets will serve as an early warning of credit deterioration or demand softening in the neocloud sector.