CoreWeave after Microsoft spend cut: $76.07 bull, $27.48…

CoreWeave is being priced as if the Microsoft spend cut cancelled a customer, and the documents do not say that. Nasdaq’s official close on Friday 30 October 2026 was $54.87. From that print, 42 sessions through 31 December 2026 put the bull at $76.07, the base at $41.30, and the bear at $27.48. Those three figures are the distribution on the last 90 daily log returns, not a street target. The tape is missing the split between a delivery slip and a cancelled contract. Amy Hood, Microsoft’s chief financial officer, said on the 28 October call that capital expenditures were $45.2 billion because some third-party capacity deliveries expected in the quarter were delayed. She then put calendar-year spend at approximately $170 billion because of those delays, and she said the next quarter would still be over $50 billion. CNBC and Business Insider reported the $170 billion figure as a cut from $175 billion. Hood’s own sentence does not say from $175 billion, and it does not name CoreWeave.

The other missing fact is CoreWeave’s own book, still the quarter ended 30 June 2026. Remaining performance obligation was $103.7 billion. The 11 August release put revenue backlog at approximately $104 billion, and the footnote said that figure does not include more than $25 billion of net new commitments added in early in the third quarter. Against that book the 10-Q shows $35.068 billion of debt, a $626 million quarterly loss, and $35.5 billion of leases not yet commenced. A timing slip can hurt a supplier whose halls were late. It is not, on these pages, a cancelled CoreWeave contract. No CoreWeave-only Microsoft dollar appears in the documents saved here, so none is invented.

What just happened, and why the obvious reading is wrong

Microsoft reported fiscal first-quarter 2027 results on 28 October, for the quarter ended 30 September. Revenue was $92.3 billion, up 19 percent. Operating income was $45.9 billion. GAAP net income was $36.6 billion. Microsoft Cloud revenue was $63.9 billion, up 30 percent. Commercial remaining performance obligation rose 73 percent to $677 billion, and Azure revenue rose 48 percent. The release does not print an Azure dollar.

Hood’s release sentence is about that print, not a cancelled build. “We delivered a strong start to the fiscal year with Microsoft Cloud revenue of $63.9 billion, with growth accelerating to 30% (up 31% in constant currency) year-over-year,” she said. On the call she said: “Capital expenditures were $45.2 billion as some third-party capacity deliveries we expected in Q1 were delayed.” About three quarters of capex was short-lived assets, mostly CPUs and GPUs. Finance leases were $7.4 billion and cash paid for equipment was $39.4 billion. Adding those two does not recover $45.2 billion, and Hood did not present them as the capex total. Free cash flow was $18.6 billion. CNBC put it at $18.63 billion, down 27 percent.

The guide is a third sentence, and it is not the $175 billion comparison. “In Q2, we expect CapEx spend will be over $50 billion. Therefore, for the calendar year, spend should be approximately $170 billion due to the 3P capacity delays noted earlier.” She said second-half capex should accelerate, including deliveries that shifted into that half, and that demand still exceeds supply. Microsoft’s fiscal second half runs through June 2027. CNBC on 29 October described some capacity moving into the first half of 2027. That is CNBC’s calendar reading, not a second Hood quote.

Jordan Novet at CNBC wrote on 29 October that Microsoft reduced the year’s forecast to $170 billion from $175 billion. Ashley Stewart at Business Insider dated the same trim to calendar 2026. A 28 October CNBC recap instead said calendar 2027. This piece follows Hood’s words, “the calendar year,” and does not say Microsoft scrapped the build. The next quarter is still guided above $50 billion.

StreetAccount’s $50.67 billion is $5.47 billion above the print. TechSpot’s Financial Times figure of $51.7 billion is $6.5 billion above it. CNBC said the quarter was up about 30 percent from a year earlier. The saved release does not print that year-ago dollar, so it stays out. The $50 billion line Hood spoke is next quarter, not a reconstructed first-quarter company target.

Who is exposed

Microsoft named no supplier. CNBC wrote that Hood did not provide names, and that Microsoft relies on CoreWeave, IREN and Nebius, among others. That reliance line is CNBC’s. TechSpot, citing the Financial Times, wrote that Microsoft has signed more than $80 billion of contracts with CoreWeave, Nebius and Nscale since 2023. That is a three-company total. It is not a CoreWeave contract value. The same article said an Nscale deal was abandoned and a Nebius New Jersey site is slower. Those lines are about other suppliers.

The 10-Q names Microsoft and OpenAI as other significant customers and gives no Microsoft dollar. OpenAI committed up to about $6.5 billion through 31 May 2031. Meta initially committed up to about $21.0 billion through 20 December 2032. Jane Street committed about $6.0 billion, announced in April 2026. Amazon, Google, Microsoft and Oracle compete and also buy. In the June quarter Customer A was 36 percent of revenue, B 26 percent and C 10 percent, against 71 percent for Customer A a year earlier. The letters may be different customers than before. They are not mapped to Microsoft here. Take-or-pay contracts were 98 percent of that quarter’s revenue.

At 31 July 2026 there were 551,536,602 Class A and Class B shares. Times Friday’s Class A close, that is about $30.26 billion, an illustration only: the count is 31 July, and Class B is marked at the Class A price. At 30 June, assets were $77.070 billion, liabilities $72.046 billion, equity $5.024 billion, and the accumulated deficit $4.009 billion. The next company document is the 10 November call. Morpheu5 Stock Watcher wrote on 31 October that Friday was $54.87, down $5.87 or 9.66 percent, with no CoreWeave news that day. That price matches the Nasdaq row.

What the tape and the filings show

Wednesday 28 October closed at $63.93, up from Tuesday’s $59.57, before the Microsoft report. An after-hours note that evening said CoreWeave and Nebius slid almost 5 percent. The official Thursday close was $60.74, down 4.99 percent. Friday closed at $54.87, down 9.66 percent on the day and 14.17 percent from Wednesday, a $9.06 drop. Friday’s range was $53.48 to $56.90 on 53,477,290 shares. The low did not undercut the 26 October low of $51.24.

The five sessions before Friday ran from $51.24 to $64.39. Friday’s close is inside that range, $9.52 under the high and $3.63 above the low. Friday is not the file’s lowest close, which was $54.235 on 23 October. Over the last 252 sessions the high was $138.36 on 31 October 2025, and Friday is 60.3 percent under it. The Kobeissi Letter listed CoreWeave at minus 60 percent the same day.

Nasdaq official daily closes for CRWV through Friday 30 October 2026. The top panel includes the daily lows so the spring high and the October low stay in frame. FinanceFeeds chart, 31 October 2026.

The 20-day average is $68.77 and the 50-day average is $79.45. Friday is $13.90 under the shorter line and $24.58 under the longer one. The bull at $76.07 is $7.30 above the 20-day average and $3.38 under the 50-day average. The chart runs from March through 30 October. Its high is $138.25 and its low is $51.24.

ItemMicrosoft, quarter ended 30 September 2026CoreWeave, quarter ended 30 June 2026

Revenue$92.3 billion, up 19 percent$2,575 million, against $1,212 million
Profit or lossGAAP net income $36.6 billion, up 32 percentNet loss $626 million, or $1.14. Operating loss $49 million
Spend or interestCapex $45.2 billion. PP&E cash $39.4 billion. Finance leases $7.4 billionInterest expense, net $640 million, 24.9 percent of revenue
Forward bookCommercial RPO $677 billion, up 73 percentRPO $103.7 billion. Backlog about $104 billion
What was saidDeliveries delayed. No supplier named. Reported guide $170 billion from a reported $175 billionDemand accelerating, per Intrator. Microsoft named, with no dollar

June revenue of $2,575 million against $1,212 million is up 112.5 percent. Adjusted EBITDA was $1,510 million. Interest was larger than the $49 million operating loss. Active power rose by nearly 500 megawatts to 1.5 gigawatts, and contracted power was about 3.7 gigawatts.

Mu on the last 90 daily log returns is -0.00676628 and sigma is 0.06283583. N is 42: 20 November sessions, skipping Thanksgiving on 26 November, and 22 December sessions, skipping Christmas on 25 December. Early closes on 27 November and 24 December count. Base is spot times e to the N times mu, $41.30. The bull adds 1.5 sigma times the square root of N, $76.07. The bear subtracts one sigma times that root, $27.48. The bull clears the spot, so a no-drift bull of $101.07 is unused. The 30 July 2026 jump from $60.82 to $73.90, log return 0.194794, stays in the sample.

The year-end pair is not within $1 of the English pairs at $107.76 and $35.73, $90.60 and $30.81, $192 and $47, $303 and $36, or a live $250 bull. The unused ten-session path through 13 November is $69.09, $51.28 and $42.04.

What the companies have said, and what they have declined to say

Satya Nadella, chairman and chief executive, drew a line under customer breadth on the same call. He said he has been waiting for breadth “as opposed to like 4 customers who buy a lot of capacity from you and you can show a few quarters of growth. And that’s not a business as far as I’m concerned.” He named no company. The remark is about Microsoft’s cloud, not a finding that CoreWeave lost a buyer. It does sit next to a June quarter in which three customers were 72 percent of CoreWeave revenue.

Michael Intrator, co-founder, chairman and chief executive, last spoke in the 11 August release. “CoreWeave reached an important inflection point this quarter as our scale began to translate into expanding operating leverage. Customer demand is accelerating, as enterprise adoption broadens and we continue to deepen our technology platform,” he said. That is a June sentence. It is not a reply to Hood. The 27 October release only sets the 10 November call. CoreWeave has not, in any page saved here, confirmed or denied that its own halls were the ones Microsoft was waiting on.

On 29 October Small Cap Snipa wrote that Evercore sees the delay as likely not CoreWeave, citing the 1.5 gigawatts in the August release. The post’s 150 to 200 megawatt estimate is not Hood’s number. Jonathan Neilson is on the participant list. A Reuters line attributed to him did not appear in any page that downloaded, so it is not quoted.

The financing tension

At 30 June the debt lines were $6,235 million, $1,278 million, $25,170 million and $2,385 million. The sum is $35.068 billion. Cash was $5,524 million. Total liquidity was $15,553 million, including $10,014 million of secured commitments. The 10-Q says that liquidity covers obligations due within one year of the report. Leases not yet commenced were $35.5 billion of undiscounted payments, starting from 2026 to 2029, on terms of seven to sixteen years. Operating lease liabilities already booked were $16.319 billion. Deferred revenue was $9.692 billion.

Net interest of $640 million was 24.9 percent of June revenue. Of the $103.7 billion RPO, 41 percent is due in the 24 months ending 30 June 2028, 39 percent in the following two years, and the rest after that. If halls slip, revenue moves right and the coupon does not. That links Hood’s delay to the share without a cancelled contract.

The note on CoreWeave bond yields separated the credit price from the equity. The fleet-rate note asked what a capped dollar per megawatt does to a backlog that still has to be built. Truist upgraded the shares and cut the target in the same note. None of those pages is this 42-session formula.

What happens next

Tuesday 10 November 2026, at 5:00 p.m. Eastern, is the first test. If CoreWeave does not disclose a Microsoft cancellation, and the backlog stays near the June RPO plus the footnoted early-third-quarter commitments, the cancelled-customer reading loses the document that could have confirmed it. The unused path through Friday 13 November is $69.09, $51.28 and $42.04.

The title horizon is 31 December 2026. The $76.07 bull needs the negative drift to fade and the delayed deliveries to arrive in the half Hood named. It retraces part of the drop from $63.93 and still sits under the 50-day average. If the call shows a named customer stepping down, $27.48 is the drift plus one standard deviation, not the older $36 beside a $303 bull.

Bull, base, and bear

The cases are the distribution. An Evercore lean on X, or any other street price, is a separate number. Older FinanceFeeds bulls and bears are the pages this piece replaces. They are not inputs.

CasePriceVersus the $54.87 closeWhat has to be true

Bull$76.07+$21.20, or +38.64 percentThe negative drift fades. Delayed deliveries arrive in the half Hood indicated. The 10 November call does not disclose a cancelled Microsoft contract, and the backlog keeps converting.
Base$41.30-$13.57, or -24.73 percentThe last 90 sessions’ average log return continues for 42 sessions. No extra shock and no extra rescue. The June debt and the uncommenced leases remain the constraint.
Bear$27.48-$27.39, or -49.92 percentThe same drift, plus one sample standard deviation against the stock. A named customer steps down, or funding pushes past the 10-Q’s one-year sufficiency line.

This is not financial advice. The closes used here are Friday’s completed prints from the Nasdaq file saved for this piece. A $54.60 figure posted on Saturday is not the spot.

Frequently asked questions

Where do the sessions through 31 December 2026 put CoreWeave once the Microsoft spend cut is read against the backlog and the debt?

The 42 sessions through 31 December 2026 put the bull at $76.07 and the bear at $27.48 from Friday’s $54.87 Nasdaq close, with the base at $41.30. That pair is the house formula on 90 daily log returns, not a cancelled-customer story. Remaining performance obligation was $103.7 billion against $35.068 billion of debt. The Microsoft documents describe a delivery delay and a reported guide trim, not a CoreWeave cancellation.

Did Microsoft cut spending from $175 billion to $170 billion, or only miss a quarter?

Both facts are in the record, and they are not the same sentence. Capital expenditures were $45.2 billion, below a $50.67 billion StreetAccount estimate. Hood then said calendar-year spend should be approximately $170 billion because third-party deliveries slipped, with the next quarter still over $50 billion. CNBC and Business Insider reported the $170 billion as a cut from $175 billion. Her saved sentence does not say from $175 billion.

Is there a published dollar figure for a CoreWeave contract with Microsoft?

Not in the documents saved for this piece. The 10-Q names Microsoft and OpenAI as other significant customers and prints dollars for OpenAI, Meta and Jane Street, not for Microsoft. TechSpot, citing the Financial Times, grouped CoreWeave with Nebius and Nscale in more than $80 billion of contracts since 2023. That is a three-company total. It is not used here as a CoreWeave line item.

Why is the base under Friday’s close if the bull is above it?

The base is the drift-only price, and the drift is negative. Mu on the last 90 log returns is about -0.0068 a session, so 42 sessions compound to $41.30. The bull adds 1.5 standard deviations and reaches $76.07, which is still under the 50-day average of $79.45. The bear subtracts one standard deviation and reaches $27.48. The bull did not need a no-drift override, because it already clears the spot.

What would change these numbers before year-end?

The 10 November 2026 report is the date that can change the story the formula is being asked to carry. A disclosed cancellation, or a backlog cut, would make the bear the relevant description even if the past 90 sessions stayed put. A clean delivery update would not rewrite mu by itself. It would tell a holder whether the negative drift is the regime to underwrite through 31 December.

How far is $54.87 from the recent trading range?

Friday’s close is under the 20-day average of $68.77 and the 50-day average of $79.45. It is inside the prior five-session range of $51.24 to $64.39. It is 60.3 percent under the $138.36 high on 31 October 2025, the high of the last 252 sessions in the file. It is not the lowest close in the file. That print was $54.235 on 23 October 2026.