
- CoreWeave Q2 2026 Earnings Expectations: Revenue, EPS and Capex
- Can CoreWeave Convert Its AI Backlog Into Revenue and Cash Flow?
- CoreWeave’s $99.4 Billion Backlog: What It Means for Revenue
- CoreWeave Q2 Margins vs EPS: Why Profitability Matters More
- CoreWeave Capex, Debt and Cash Flow: The Funding Gap Explained
- What Changed After CoreWeave’s Previous Earnings?
- How Should Investors Read the Earnings Release?
- CoreWeave Earnings Scenario Comparison
- Our Take on CoreWeave’s Q2 Earnings
CoreWeave does not need to prove that demand for artificial intelligence computing exists when it reports second-quarter earnings today. It needs to prove that it can turn its $99.4 billion backlog into powered, revenue-generating infrastructure fast enough to outrun depreciation, interest costs and enormous capital spending.
CoreWeave will discuss its second-quarter results after the US market closes on August 11, 2026. The earnings call begins at 5:00 PM Eastern Time, which is 2:30 AM IST on August 12 for Indian investors.
Let's break down what Wall Street expects, what changed after the previous earnings, and which numbers can separate a high-quality quarter from a headline-only beat.
We will also build a simple backlog-to-power-to-cash framework that investors can use for this quarter and future CoreWeave earnings.
CoreWeave Q2 2026 Earnings Expectations: Revenue, EPS and Capex
Wall Street expects CoreWeave to deliver another quarter of triple-digit revenue growth. However, estimates for the company’s loss remain unusually wide because depreciation, interest, equipment activation and tax expenses can move sharply between quarters.
| Q2 metric | Company guidance | Analyst benchmark | Q2 2025 actual |
| Revenue | $2.45 billion to $2.60 billion | Around $2.56 billion | $1.21 billion |
| Revenue growth | Not specified | Around 111% | 207% |
| Adjusted operating income | $30 million to $90 million | Around 2.4% to 2.8% margin | $200 million |
| GAAP EPS | No guidance | Loss of roughly $1.22 to $1.42 | Loss of $0.60 |
| Capital expenditure | $7 billion to $9 billion | Midpoint of $8 billion | $2.9 billion |
| Interest expense | $650 million to $730 million | Midpoint of $690 million | $267 million |
Sources: CoreWeave Q1 outlook, CoreWeave Q2 2025 results, Visible Alpha estimates reported by Investopedia, Benzinga Pro estimates
The revenue consensus is close to the upper end of CoreWeave’s own guidance. This leaves little room for a simple revenue beat to surprise investors.
The more important disagreement is over the second half of 2026.
Bank of America expects Q2 revenue growth of around 108% but estimates an adjusted operating margin of 2.4%, below the 2.8% Street expectation. It expects growth to accelerate as more capacity becomes active in the second half. Oppenheimer expects revenue near the high end of guidance, but believes some Q3 capacity additions could be delayed by three to five months and shift into Q4.
There is an even larger disagreement around the backlog. A Cantor Fitzgerald analysis reportedly estimated that CoreWeave’s backlog could reach approximately $131 billion, compared with a broader Wall Street estimate of around $104.4 billion. That is a difference of nearly $27 billion in expectations for the same metric.
Can CoreWeave Convert Its AI Backlog Into Revenue and Cash Flow?
The simplest way to understand CoreWeave is to think of it as a restaurant with a reservation book full for years, but with several dining rooms still under construction.
The backlog represents reservations. Contracted power represents the future dining rooms. Active power represents the tables that can serve customers today. Revenue is the food actually served. Capital expenditure and interest are the cost of building and financing the restaurant.
A full reservation book is valuable, but only if the rooms open on time and each table makes enough money after costs.
This creates what we call the CoreWeave backlog-to-power-to-cash bridge:
| Stage | Latest disclosed position | What investors should ask |
| Customer contracts | $99.4 billion backlog | Is backlog still growing and diversifying? |
| Contracted power | More than 3.5 GW | When will this power become usable? |
| Active power | More than 1 GW | Is CoreWeave on track for over 1.7 GW by year-end? |
| Revenue | $2.45 billion to $2.60 billion Q2 guide | Is capacity being handed to customers on schedule? |
| Adjusted operating income | $30 million to $90 million Q2 guide | Are new deployments becoming profitable? |
| Interest and capex | $690 million and $8 billion midpoints | How much value remains for shareholders? |
CoreWeave explained on its previous earnings call that a new data centre can produce negative contribution margins during the first one to two months. The company begins paying lease, power and depreciation costs before customer revenue starts. By approximately the third month, management says contribution margins normally move towards the mid-20% range.
That makes active power and deployment timing leading indicators. Revenue is the result that appears later.
CoreWeave’s $99.4 Billion Backlog: What It Means for Revenue
CoreWeave reported $99.4 billion of revenue backlog at the end of Q1. Of this, $98.8 billion was classified as remaining performance obligations, or RPO.
RPO is contracted revenue that has not yet been recognised. However, it is not the same as cash in the bank. CoreWeave still needs to deliver the required infrastructure and meet service-availability conditions.
| Backlog calculation | Amount |
| Q1 revenue backlog | $99.4 billion |
| Q1 RPO | $98.8 billion |
| RPO expected within 24 months | 36% |
| Implied 24-month revenue pool | Around $35.6 billion |
| Average annual conversion | Around $17.8 billion |
| FY2026 revenue guidance midpoint | $12.5 billion |
Source: CoreWeave Q1 2026 Form 10-Q
The math suggests that CoreWeave has enough contracted demand to support its revenue ambitions. The constraint is not the size of the order book. It is how quickly CoreWeave can activate power, install GPUs and hand capacity to customers.
Therefore, a growing backlog accompanied by delayed capacity would be a mixed signal. It would increase future opportunity, but also increase the amount of infrastructure CoreWeave must finance and deliver.
A better-quality backlog update would contain three things: meaningful growth, more investment-grade or established enterprise customers, and a clear timeline for converting the contracts into revenue.
CoreWeave Q2 Margins vs EPS: Why Profitability Matters More
CoreWeave reported $2.08 billion of Q1 revenue, up 112% year-on-year. Yet adjusted operating income fell to just $21 million, giving the company a 1% adjusted operating margin. Net interest expense reached $536 million.
Management described Q1 as the bottom of its margin cycle and guided for sequential improvement during the rest of 2026. Q2 is the first test of that statement.
| Q2 midpoint calculation | Result |
| Revenue midpoint | $2.525 billion |
| Adjusted operating income midpoint | $60 million |
| Implied adjusted operating margin | 2.4% |
| Interest expense midpoint | $690 million |
| Interest as a percentage of revenue | 27.3% |
| Capital expenditure midpoint | $8 billion |
| Capex as a multiple of revenue | 3.2 times |
The 2.4% midpoint margin would be an improvement from Q1, but it would still be far below the 16% adjusted operating margin reported in Q2 2025.
More importantly, interest expense at the midpoint would be more than eleven times adjusted operating income. This explains why EBITDA can look strong while net income remains deeply negative. EBITDA excludes interest and depreciation, two of the largest costs in CoreWeave’s business.
For CoreWeave, adjusted EBITDA is a measure of operating scale. It is not a substitute for cash available to shareholders.
CoreWeave Capex, Debt and Cash Flow: The Funding Gap Explained
CoreWeave generated $2.98 billion of operating cash flow in Q1, helped by customer payments and working-capital movements. It spent $7.70 billion on property and equipment.
That produced a simple cash flow funding gap of approximately $4.71 billion before financing activities.
| Q1 cash flow bridge | Amount |
| Operating cash flow | $2.98 billion |
| Property and equipment purchases | $7.70 billion |
| Cash funding gap | $4.71 billion |
| Debt principal at March 31 | $25.15 billion |
| Operating lease liabilities | $10.05 billion |
| Cash and equivalents | $2.24 billion |
CoreWeave also disclosed $40.7 billion of estimated future lease payments for leases that had not yet commenced. These commitments begin between 2026 and 2029 and should not be treated as immediately payable debt. Still, they show how much future infrastructure the company has already committed to securing.
This is why investors should track the cost of capital alongside revenue. When a company must keep raising money, the interest rate and financing structure can materially change the value that ultimately reaches equity shareholders.
What Changed After CoreWeave’s Previous Earnings?
Several important developments took place between the May earnings release and the current report.
| Date | Development | Why it matters |
| May 18 | Closed a $3.1 billion DDTL 5.0 facility at SOFR plus 4.5% | Expanded access to publicly syndicated GPU-backed financing |
| June 12 | Selected for the Nasdaq-100 | Increased index ownership and visibility |
| June 17 | Reported early validation of Nvidia Vera Rubin NVL72 | Supported CoreWeave’s technology-execution case |
| July 30 | Announced a collaboration with Leidos | Opened a path towards US defence and intelligence workloads |
| August 4 | Announced three Indonesian facilities totalling 360 MW | Marked its first Asia-Pacific data-centre expansion |
| August 11 | Closed a $2.6 billion DDTL 5.5 facility at SOFR plus 5.5% | Expanded financing but at a higher spread and with renewal risk |
Sources: CoreWeave DDTL 5.0 announcement, Nasdaq-100 announcement, Vera Rubin update, Leidos collaboration, Indonesia expansion
The new $2.6 billion financing facility deserves special attention. It has an approximate five-year maturity, while the underlying customer contracts average around three years. This means lenders are accepting some risk that CoreWeave will need to renew those contracts or lease the GPUs to new customers.
The facility was priced at SOFR plus 5.5%, compared with SOFR plus 4.5% for the May facility. All else being equal, one additional percentage point on $2.6 billion represents approximately $26 million of extra annual interest if the entire facility is drawn. Since it is a delayed-draw loan, actual interest depends on how much CoreWeave borrows and when.
This does not mean CoreWeave has lost access to capital. The deal was completed and reportedly oversubscribed. But it suggests that financing conditions are not automatically becoming cheaper across every contract type.
How Should Investors Read the Earnings Release?
The following order can help investors avoid getting distracted by whichever number appears in the headline.
| Metric | Stronger signal | Mixed signal | Concerning signal |
| Revenue | Above $2.60 billion | Within $2.45 billion to $2.60 billion | Below $2.45 billion |
| Adjusted operating income | Above $90 million | Within $30 million to $90 million | Below $30 million or a loss |
| Active power | Clear progress towards more than 1.7 GW | Target maintained with limited detail | Delays or reduced year-end target |
| Backlog | Growth with customer diversification | Growth driven by existing large clients | Flat or lower backlog |
| Q3 outlook | Supports the expected H2 acceleration | Revenue shifted from Q3 into Q4 | Another major capacity delay |
| FY2026 guidance | Maintained or increased | Maintained but more back-end weighted | Revenue or margin guidance reduced |
| Financing | Stable costs and contract-linked funding | More debt but matched with capacity | Higher costs without faster activation |
The earnings call may matter more than the press release. Investors should listen for specific comments about powered-shell delivery, labour availability, component shortages, active power, customer concentration and whether the expected Q3 margin inflection remains intact.
CoreWeave Earnings Scenario Comparison
| Scenario | What the numbers might show | Existing investor lens | Potential investor lens |
| High-quality beat | Revenue above guidance, margin above 3.5%, power ramp on schedule and strong outlook | Test whether the improvement is repeatable rather than relying on the stock reaction | Recalculate future margins and funding needs before judging valuation |
| Headline-only beat | Revenue beats, but interest, capex or delays rise faster | Treat the revenue beat cautiously because equity economics have not improved | Avoid using revenue growth alone as evidence that risk has fallen |
| Mixed quarter | Revenue near the high end, modest margin improvement and guidance maintained | Keep the thesis centred on Q3 and Q4 capacity activation | The report may provide more information without materially changing the risk-reward equation |
| Weak execution | Revenue below guidance, capacity delayed, margins weak or guidance cut | Rebuild the model using slower backlog conversion and higher funding costs | Require stronger evidence of execution before relying on long-term backlog |
| Demand warning | Backlog stagnates, major customers slow commitments or prices weaken | Separate an industry demand issue from a temporary construction delay | Reassess whether the original AI infrastructure growth assumption still holds |
A temporary capacity delay and a demand slowdown are not the same problem. A delay pushes revenue into a later quarter. Weak customer demand reduces the economic value of the infrastructure itself. The market reaction may be negative in both cases, but the long-term investment implications are very different.
Our Take on CoreWeave’s Q2 Earnings
The most reasonable pre-earnings base case is revenue near the upper end of CoreWeave’s guidance, accompanied by a modest improvement in adjusted operating margin. That alone would not settle the investment debate.
The decisive information will be whether CoreWeave remains on track to exceed 1.7 GW of active power, whether Q3 still marks the expected acceleration, and whether the company can finance its expansion without interest costs absorbing the benefit of higher revenue.
The bullish version of the story is straightforward: CoreWeave has almost $100 billion of backlog, customers are competing for scarce AI capacity, newer GPU generations strengthen its platform and more activated power creates sharp revenue growth.
The riskier version is equally clear: backlog commitments require enormous spending, the top two customers generated 65% of Q1 revenue, all deployed GPUs come from Nvidia, and the company continues to depend on external capital.
Therefore, this earnings report should not be judged as a contest between an EPS beat and a miss. It is a progress report on whether CoreWeave’s financial engine can catch up with its AI infrastructure engine.