Operator
Good day, and thank you for standing by. Welcome to Boost Run's third quarter 2026 earnings conference call. At this time, all participants are in a listen-only mode. After the speaker's presentation, there will be a question and answer session. To ask a question during the session, you will need to press star 11 on your telephone. You will then hear an automated message advising your hand is raised. To withdraw your question, please press star 11 again. Please be advised that today's conference call is being recorded. I would now like to hand it over to your first speaker today, Jenny Lee, Director of Investor Relations.
Thank you, Operator, and good morning, everyone. Welcome to Boost Run's third quarter, 2026 earnings conference call. Joining me on the call today are our founder and chief executive officer, Andrew Carros, and our chief financial officer, Eric Guckel. Our chief operating officer, Harry Georgikopoulos, is also with us and will be available during the Q&A session. Before we begin, I'd like to remind you that today's discussion will include forward-looking statements within the meaning of the federal securities laws, including statements regarding our expected future financial performance, annual recurring revenue, contracted revenue backlog, the timing of project deployments, data center capacity, hardware procurement, financing and customer prepayments, net cash flow margin, and capital expenditure levels. These statements are based on our current expectations and are subject to risks and uncertainties. Actual results may differ materially from those contemplated by these statements. The factors that could cause results to differ are described in our filings with the SEC, including the risk factors section of our quarterly report on Form 10-Q for the quarter ended September 30, 2026 from our most recent Form 10-Q. We undertake no obligation to update these statements except as required by law. We will also discuss certain non-GAAP financial measures and key operating metrics including ARR, or Annual Recurring Revenue, Run Rate Revenue, Net Cash Flow Margin, and our Total Contract Value, or TCV, Apex Ratio. These measures should not be considered in isolation or as a substitute for the most directly comparable measures prepared in accordance with GAAP and may not be comparable to similarly titled measures used by other companies. A reconciliation of each historical non-GAAP measure to the most directly comparable GAAP measure is included in the appendix in today's investor presentation. A copy of the earnings release and investor presentation release can be found on our IR website at investors.boostrun.com. Please note that this call is being recorded and will be available for replaying via our IR website. And with that, it is my pleasure to turn the call over to our founder and CEO, Andrew Caron. Andrew?
Thank you, Jenny, and good morning, everyone. Thank you for joining us for our third quarter 2026 earnings call. Today, we want to accomplish several things. First, I will review our third quarter performance and what we have achieved since September 30th, 2026. Second, I want to spend some time talking about our business and financial model and why we believe it is the right model for positioning us to create long-term shareholder value. Third, I will talk about how our contracted co-location footprint continues to grow and how our access to equipment and financing continues to improve alongside it. Fourth, I will lay out the funding principles behind our financial model. Lastly, Eric will further elaborate. Into how cash flows in and out of our contracts from start date to finish, and why it's important for shareholders to understand both cash and GAAP accounting in evaluating our business before moving into our Q3 results. For the third quarter, revenue reached $39.9 million, up approximately 496%. Year-over-year, from 6.7 in Q3 2025 and up 27% from the $31.1 million we generated in Q2. We continue to bring new projects live into 2027. Recently, in Q4, 2026, we brought two additional projects live with a combined TCV of approximately $455 million, adding an estimated $9.5 million in monthly reoccurring revenue. Due to a data center delay preventing a project from going live in Q4, our exit ARR would have been approximately $435 million. The reason for this delay has now been resolved. In addition, by the end of the year, we will be receiving approximately $155 million in cash related to completion payments on projects. As a reminder, our $2.7 billion in TCV is either live In current deployment, we're fully contracted and scheduled for deployment in 2027. For further clarity, this means signed contracts scheduled to go live, purchase orders submitted and accepted by OEMs, data center space and power contracted and progressing, non-dilutive financing term sheets with multiple parties. Operating cash flow in the quarter was $65.6 million. We exited the quarter with $157 million in unrestricted cash. Year to date, we have secured approximately $600 million in non-dilutive project-based financing, and we have approximately $482 million of customer prepayments due in connection with upcoming project deployments, providing further non-dilutive funding for our growth. In the corner, we eliminated any warrant overhang, leveraging our ability to exercise the warrants, which resulted in an additional $131.5 million in cash per boost run. Next, I'd like to turn our business and financial model. As we sit here today, our long-term TCB has now surpassed $2.7 billion. Our customer base spans some of the AI ecosystems. Most compute-intensive participants, including Frontier AI Research Labs, foundational model developers, sovereign and enterprise AI, generative AI platforms, AI imprints and application providers, and GPU cloud and marketplace operators. Beyond the AI sector, we also serve enterprise quantitative trading and financial service in life science, reflecting growing demand for large-scale accelerated computing across industries. As of today, we expect to have approximately 26% of our 2.7 TCV live by the end of November 2026. We continue to negotiate additional contract agreements with new customers. Our established projects will be supported by $482 million in prepayments alongside non-dilutive project-level financing with tenors equal to or shorter than customer contract duration. Each project adds positive incremental cash flow, starting with 1-1 on being fully deployed. Our customer agreements continue to execute with longer tenors, and our $1.44 billion purchase agreement with Dell is fully committed and contracted against customer demand. Of the $4 to $5 billion in strategic hardware procurement, we've outlined for 2027 about $330 million in purchase orders. have already been submitted and accepted by our primary OEM partners. Now, I'll turn it to our contracted co-location footprint and how our access to data center capacity continues to expand. We plan to expand from six active U.S. data center sites to nine sites over the next six months in the states of Washington, Minnesota, Illinois, Texas, North Carolina, Georgia, and We expect our deployed co-location power capacity to reach approximately 40 megawatts by the end of 2026, and through our expanded partnerships, we have already secured 253 megawatts of total accessible capacity to support our 2027 and 2028 RAM, and because we lock in long-term co-location leases ranging from 3 to 12 years in advance, We believe we are largely insulated from spot collocation inflation in a tight U.S. data center market. We continue to evaluate and act on strategic capacity on an opportunistic basis. Next, I want to discuss the key funding principles behind our financial model, including how we structure each project and mitigate risk. First, mandatory customer prepayments. We require a substantial upfront cash prepayment at contract signing for 100% of our multi-year customer contracts. Across our portfolio, customer prepayments range between 30 and 40% of total project cost. Second, non-dilutive project-based financing. The remaining total project cost is funded through non-dilutive project-level equipment financing for senior-secured project term loans backed directly by the contracted take-or-pay cash flows of those specific customer deployments. This requires minimal to zero parent-company equity dilution. Third. 100% of project financing is repaid before the customer agreement expires. This is the most critical risk mitigant in our model today. We structure our project financing in equipment leases so that 100% of the project financing is fully repaid before the underlying customer contract expires. In the current financing environment, concerns have risen about debt tenure. is often far exceeding contract length, which then exposes the debt holder to returns based on residual value or contract renewal at better terms. As former institutional risk managers, we designed our financing structure to mitigate tail risk, specifically to mitigate residual value in refinancing risk. First, we are not dependent on GPU residual value. Investors often ask what happens to the resale value of a GPU after years as NVIDIA transitions from Hopper to Blackwell to Vera Rubin. At Boost Run, we underwrite every single project to the initial non-cancelable contract, not to a residual hardware value, because 100% of our project financing is structured to pay down inside the contract term. Our capital is expected to be returned, and our base return on equity is designed to be locked in before the initial contract ends. Every dollar a GPU earns after its project debt is earned on a fully paid-off, debt-free asset. Second, we have no dilutive debt, warrant overhang, or refinancing cliffs. Some industry participants have taken on billions of dollars in high-yield corporate debt with bullet maturities that require constant capital markets refinancing. Our financing is project-specific and paid down every month out of our contracted customer cash flows. Third, we have deliberate portfolio diversification and on-demand margin optimization. We have intentionally constructed a portfolio that is diversified by contracted duration and by client. Currently, no singular customer contract represents more than 23% of our total portfolio. Alongside our three- to five-year anchor contracts with Frontier AI Labs, seven-year AI entities, and enterprise inference leaders, We also allocate a portion of our self-funded capacity to our proprietary boost-run on-demand platform. Today, more than 10% of our revenue comes from short-term on-demand contracts, which commands pricing two to two and a half times higher than longer-term reservations and carry higher margins. In summary, we believe that our funding approach mitigates many of the risk costs associated with some of the more traditional AI infrastructure business models. That allows us to focus on building strong base of long-term customers, securing project-level funding, building and deploying AI clusters efficiently, managing them continuously, and developing tools and software that improve usability, security, and flexibility for our customers. With that, I will turn the call over to our CFO, Eric, to walk through our third-quarter financial results and dive more deeply into how our backlog converts into GAAP revenue and cash flow.
Eric? Thank you, Andrew, and good morning, everyone. Because we're scaling quickly and following a rigorous capital deployment model, I'd like to discuss how our earnings should be interpreted to be a sound measure of our business. After our second quarter transaction, Q3 represented our first clean operating quarter without the overlay of significant adjustments and one-time items. I'll provide some comments on the headline results first and then dig into a bit more detail on operating cash flow and customer prepayments, which are central to how our model works. For the third quarter and nine-month results. Revenue in the third quarter was $39.9 million, up 496% from $6.7 million in Q3 2025, and up 28% sequentially from $31.1 million in Q2. For the nine months, revenue was $82 million, up 325% from $19.3 million, driven by new GPU cluster deployments going live. The net loss in Q3 was $14.1 million, or roughly $0.17 per share, compared with $17.2 million, or $0.58 per share, a year ago, and included a $9.7 million income tax benefit. For the nine months, the net loss was $93.3 million, or $1.78 per share. That figure includes a one-time non-cash income tax provision of $55.7 million recorded in the second quarter when we converted from an LLC to a C-Corp, partially offset by the Q3 benefit. Adjusted EBITDA was $18 million in Q3 and $32.5 million for the nine months. The difference between adjusted EBITDA and net loss. It's largely non-cash, including $34.8 million of depreciation and amortization in Q3. We also present lease-adjusted EBITDA, which adds back co-location lease expense. Co-location lease cost was $9.4 million in Q3 and $24.6 million for the nine months, compared with $3.8 million in the prior nine-month period, reflecting our expansion to additional data center sites. Lease-adjusted EBITDA was $27.4 million in Q3, or 68.6% of revenue, and $57.1 million for the nine months, or 69.7% of revenue. We present this measure to help investors compare us with companies that own their own data centers. The co-location is a recurring cash cost of our business and should be considered alongside our other measures. We completed our warrant redemption on August 20th. Approximately 11.4 million warrants were exercised for cash at $11.50 per share, generating $131.5 million of proceeds for the nine months, $43.4 million in Q2, and $88.2 million in Q3. We now have no warrants outstanding, approximately 84.2 million shares outstanding across Class A and Class B, no additional preferred stock, and no parent-level bank or bond debt. With new projects deploying, total assets crossed the $1 billion mark at $1.05 billion by the end of Q3. In operating cash flow, Net cash provided by operating activities was $65.6 million in the third quarter and $172.6 million for the nine months, compared with $8.6 million in the prior nine-month period. Here is how the quarter builds. Given our large depreciation and amortization, we are reporting a net loss of $14.1 million. Net non-cash items added back $34.5 million, led by depreciation and amortization. Changes in working capital contributed $45.2 million. The largest drivers were $54.4 million net increase in customer deposits and a $39.4 million increase in payables, partly offset by $27.5 million of increase in accounts receivable and $9.5 million of operating lease prepayments. Because working capital drove the majority of Q3 operating cash flow, I would encourage you not to simply annualize it. Operating cash flow also does not include the principal portion for our GPU finance lease payments, which are reported in financing activities and were $109.8 million in Q3 and $212.5 million for the nine months. We look at operating cash flow together with these payments. In customer prepayments, customer prepayments are a defining feature of our model. Customers pay a substantial portion of contract value up front. We record that cash as customer deposits on our balance sheet and recognize it as revenue on a straight-line basis over the contract term. Customer deposits are treated as a liability, not debts, and the cash is already in hand. Total customer deposits were $182.8 million at September 30th, consisting of $45.8 million current and $137 million non-current. That compares with $128.4 million at June 30th and $15.4 million at year-end 2025. The net increase was $54.4 million in Q3 and $167.4 million for the nine months, and those are the amounts reflected in operating cash flow. As Andrew noted, we expect additional prepayments ahead of our upcoming deployments. The timing of those collections will cause operating cash flow to vary from quarter to quarter, but will support project deployments and primarily cover our deployment costs and down payments on financing activities. For liquidity and capital structure, we ended the quarter with $170 million of total cash, up from $133.2 million at June 30th. The total consists of $157 million of unrestricted cash and $13 million of restricted cash supporting our data center standby letter of credit. In Q3, operating cash flow of $65.6 million and $88.2 million of warrant exercise proceeds covered $109.8 million of finance leased principal payments and $6.4 million of investing outflows, and total cash increased by $36.9 million. For the outlook, before I close, a few forward-looking points, keeping in mind the cautions Jenny reviewed at the start of the call. As mentioned by Andrew, had we not had any project shifts, we would have realized approximately $435 million of ARR by year-end 2026. We fully deployed two projects in Q3 and Q4, and we expect multiple deployments in 2027, converting all current TCV to revenues while adding additional new agreements to our diversified portfolio. As projects deployed through Q3 and Q4 reach full billing, we expect revenue to grow significantly in Q4, creating monthly recurring revenue into 2027 through the multi-year customer term. Our $182.8 million of customer deposits will be recognized as revenue over the contract terms, and we expect that balance to grow as upcoming deployments approach full handover to the customer. On funding, we expect to fund upcoming deployments primarily with customer prepayments and non-dilutive project-level financing, structured so that project debt is fully repaid before each underlying customer contract expires. It is noted that finance lease principal payments will rise as our deployed fleet grows, and we plan our liquidity around those payments together with operating cash flow and customer prepayments in a disciplined and structured manner. In our supplemental materials, we have provided a slide that describes a typical project and how that project is structured and financed. At the project level, we continue to maintain solid cash flows for each, structuring contract pricing with prepayments and covering principal and interest, co-location, and overhead with room to spare. We target minimum cash cover that facilitates both business operations and financing. And we look forward to updating you on how our deployed projects and portfolio are tracking against that target as they mature. With that, I'll turn the call back over to Andrew for closing remarks.
I'd like to conclude with a few closing points. First, with over $2.7 billion in TCB, approximately $600 million of non-dilutive project financing closed year-to-date, approximately $482 million of customer prepayments due on upcoming projects. $157 million of unrestricted cash, no warrant overhang, and a clear path to going live on our full $2.7 billion TCB in 2027. We have the team, capacity, OEM supply chain, and capital to deliver long-term value to our shareholders. Second, we have elected to take a different approach with our business and financial model compared to others in our space. We believe our approach meaningfully improves customer outcomes, reduces risk and costs, generates higher returns on invested capital and cash flow margins, and will create additive value for our shareholders. To our customers, our financing partners, NVIDIA, Dell, our other OEM partners, our co-location partners, and the entire Boost Run team, thank you for your relentless execution. I would now like to turn it back over to the operator.
Operator
Thank you. At this time, we will conduct the question and answer session. As a reminder, to ask a question, you will need to press star 11 on your telephone and wait for your name to be announced. To withdraw your question, please press star 11 again. Please stand by while we compile the Q&A roster. Our first question comes from George Sutton of Craig Hallam Capital Group. Your line is now open.
George Sutton
Analyst — Craig Hallam Capital Group (on for Eric Stein)
Thank you. I wondered if you could walk through the project shifts that you discussed up at the top of the call in terms of just give us a good sense of what changed in Q3, what has changed in Q4 from your perspective.
Sure. Thanks, George. Diving directly into your question. So from Q3 project shift, as noted, Eric, we did $39.9 million of revenue. And then in Q3, we pushed one project to Q1, and that was approximately $13 million of monthly reoccurring revenue that got pushed forward. As you can see, approximately 144 to 150 million of ARR right there. And then also, George, we recently just handed off two projects. So you can see when you hit instances like this, where Boost Run were in parabolic growth mode, shifting anything, even by a week, can trigger something to fall in a different quarter. A perfect example, two contracts we handed off, that was over the past five weeks, that didn't fall within the quarter. That's going to be adding an additional $10 million of monthly reoccurring revenue as well.
George Sutton
Analyst — Craig Hallam Capital Group (on for Eric Stein)
So, I wondered if you could just give us a bigger picture on kind of deals that you're working on, what the floor level size you would contemplate in terms of deals, and just maybe a sense of the size of the things that you're looking at. And then you also mentioned term sheets for non-dilutive financing. I wondered if you could give us an update there.
Sure. Let's unpack your questions. The first one on deal sizes. Well, I can give you specifics around this that are signed, take-or-pay contracts, another half dozen, are literally ranging from TCV from $190,000 all the way up to $626,000,000. We have other additional contracts that we're working through to get confirmed. And then from the size of GPUs, the smallest, we're basically dealing right now, George Perk cluster, and say B300s, GB300s, walking into Rubens, is approximately 2,000 GPUs, and then going up to 5,000 plus, which will eventually be going well north of 10,000 plus GPUs per cluster.
George Sutton
Analyst — Craig Hallam Capital Group (on for Eric Stein)
And I'd asked on the term sheet just a broader sense of the financing market, but, you know, you have been able to access these project-level financings. Is that the plan that you see going forward?
Yeah, so in regards to term sheets, current and future, as noted, we have over $600 million of non-dilutive financing already closed. So we're pleased to say we've successfully deployed a number of projects within this business model. And then on a go-forward basis, we're already reviewing term sheets that we have in front of us that are able to be executed. But we have literally a half a dozen projects. And the term sheet financing typically runs in parallel of go-live and build-up. That can be anywhere from four to six weeks in advance. And we are, what I can say, is very pleased with where we're at to continue our non-volutive financing business model.
George Sutton
Analyst — Craig Hallam Capital Group (on for Eric Stein)
Gotcha. Okay. Thanks, guys.
Operator
Thank you. Our next question comes from Ben Somers of U.S. Bancorp BTIG. Your line is now open.
Hey, good morning, and thank you for taking my question. So first, I kind of wanted to ask on the ramp to the potential 253 megawatts of accessible capacity, is there any permitting that needs to be done there, and just kind of how do you think about that side of the equation when you think about scaling into 2027 and 2028?
Yeah, thanks for the question. In regards to permitting, what I can tell you on half a dozen plus take-or-pay contracts that we'll be bringing in production from now throughout 27 and then high probability of many more, we do not have any permitting data center or electric. Everything is powered. So at the current time, we are not engaged in any friction on that. And a lot of that comes from the best way I think I can articulate is on grid power. And we're not going through or entering into things that can be delayed by permitting or zoning or whatnot. So with our process and procedure internally, we check box before we go into something.
Super helpful. And then just as you think about kind of the mix moving forward between term contracts and on-demand, I think you mentioned around like 10% is currently like the spot market. Kind of what do you think is the ideal mix as you continue to, you know, onboard more of the contracts and backlogs?
Yeah, I'm glad you asked that. You know, with the platform and automation and API and accessibility directly to compute through whose run platform. It makes it, I don't want to say simple, but much more accessible for end users to access compute on demand. And then when you look at the distribution of our revenue, we think that will continue to grow, but on the on demand aspect. But I want to point something out that is extremely important. Why is there a lot of on demand shortages? You cannot walk into your bank and say, I would like to finance a Billion dollar transaction because I have this great idea and I don't know what my reoccurring revenue is going to be and I want it on demand. So you see a lot of compute is not really available on a short term. That's one of the reasons. And two, the unpredictability of the future cash flow on that, that could disappear at any given point. So what we're doing is taking a risk management approach towards us. But I also want to point out that When you get into duration of contracts and you get into amortization, we are extremely pleased where Boost Run is on accelerating any pay down of debt. On our first large batch of compute when we started, we're almost paid in fold on that. So you can see how quickly that can be additive revenue.
Super helpful. Thanks for taking my questions.
Operator
Thank you. And our next question is from Nahal Chosky of Northland Capital Markets. Your line is now open.
Yeah, thank you. You already covered this in the first question, but I do want to get more detail. What are the actual sources of the project delays? I guess the first one you saw was a $13 million MRR deal that got pushed from 3Q26 to 1Q27. Can you actually detail the source of that project delay?
Yeah, are you asking for the root cause of that? Yes, please. Yeah, correct. That was one of our Georgia projects, and that was through city council. What we did is consulted with the data center operator, who are very pleased with just their balance sheet perspective. The decision was made to adhere to some of the city requests. This is already post being permitting, and what that did is that caused Work progression, and that was the setback, but very pleased to say that is now a closed case and not even an issue of standing at the current time.
Okay, and then- It provided the needed information. Yep, understood. Understood that's in the past. And why is that delayed to 1Q27 instead of 4Q26?
Can you repeat the question, please?
Given the resolution that you already have in hand, why is that project start date delayed to 1Q27 as opposed to 4Q26?
The scale and magnitude of it. So when you get into via GAAP accounting and handoff and revenue recognition, there's a lot of moving parts. So in this situation, this is in excess of 5, 6, over 7,000 GPU handoffs. Okay. So it's a continuation. That's what I'm saying in simple terms. Handing off from day one to when being in full production cannot be just a sample set snapshot in one week.
I see. The first set of GPUs are to be handed off within 4Q or starting in 1Q27 of that particular project?
We're going to start handing off in Q4 here.
Got it. Great. And then those other two projects, which was just basically, you know, a one- or two-week delay, presumably this is inconsequential, but could you just detail what were the sources of those delays as well?
Sure. On that, as everyone else gives a sound, you really can't power things when rain is occurring. We literally got hit with back-to-back for 17 days. It's actually less than that. And that was in Illinois. But that, again, with powering, there's no permitting issues, no delays. And one thing I think that's important when we're looking at these things, when you have someone like Boost Run who's in parabolic growth mode, and you're bringing on projects and going from Monthly reoccurring revenue from 10 to 20 to 30 to 40 to 50. Full deployment, substantially more of monthly reoccurring revenue. You can clearly see the impact if anything even shifts by two weeks or from December 28th to Jan 1st based on how people are calculating an ARR or quarterly revenue. So, I think that's a great example with, to be specific, $12.7 million of monthly reoccurring revenue. You can see that push. That's approximately $144 to $152 million of ARR. So the key point I want to make is if there's a data center delay or something being shipped two, three weeks late, that is a total different Topic or subject than a data center not being energized, or someone not being able to obtain financing, or they can't get hardware from the OEM. So I want to be very, very clear on those two distinctive differences that we're not incurring the latter that I just mentioned.
Got it. Thank you. I'm going to see the floor and let others ask questions, and I'll get back in the queue. Thank you.
Operator
Thank you. And just as a reminder, if you would like to ask a question, please press star 11 on your telephone and wait for your name to be announced. Our next question comes from Alex Arnold of Odeon Capital. Your line is now open.
Hey, guys. Thanks for taking the question. I guess the only thing I have left is when I look at the environment, it seems like... Scarcity is increasing and prepayments are growing. It's kind of a moving target. Can you translate in terms or give us a little more context around prepayments and how that environment's changing, but also connect the dots as to how much of the prepays on any given contract on average, when you get a prepay, how much of the CapEx, hardware CapEx, is that covering? Yeah, hey, Al.
Yep, thanks for the question. Let's hit that head on. Right now on the go forward projects, half a dozen plus and very pleased to say more coming in the pipeline. Right now that is around 38% off of the CapEx number. So it's very interesting, as you mentioned, scarcity and timing. I think it's important we step back and say, what does that 38% actually mean? In layman terms, is if the hardware costs $100 million to get in full production, we are collecting $38 million of cash upon handing off to the client. We then have a remaining debt structure of project finance of $62 million. And Alex, what you're seeing right now in the market is... What was popular for, you know, approximately $700 billion of Wall Street financing. This actually went into investment-grade contracts. Wall Street is saying that there's a minimum of a six-year life, and we know that due to the fact of hundreds of billions of being financed on six-year terms with as little as 10% down. So you can see that's one structure that currently Boost Run is actually doing the opposite of. What we've elected to do is put down a substantial amount of capital of non-volutive equity passed through that prepayment. And in a specific example on a five-year contract, that $62 million that I mentioned earlier, Against that $38 million down payment, we're going to elect to pay that down in three and a half years for, and that we'll have operating cash flow on that project while we're paying it down. And we will still be generating massive revenue from that contract as the debt is fully retired. And we do that with intent. And Alex, the best way I can explain this is the primary banks, the private credit markets, have really taken note. We're doing something different, and we're not capexing into the data centers. I know that was a very long-winded question, but you kind of hit the key point of what is going on, what are you doing, and why. So, would you step back and sound like Boo Strun. Is placing contracts in front of the debt market, we feel we're placing something that is not being seen a lot of times, and we know that by the feedback.
And how would you characterize access to financing and how has that changed over the past three to six months?
Ironically, I think the traction in our model has gained an astronomical amount of We have people asking, why do we not put less down and can you stretch out your duration? You know, with my historical experience, just dealing with multi-billion dollar credit lines, risk management is seat number one. So we've continued to do that, and what we're seeing is we've had turbsheets put in front of us, and we're evaluating those right now. And that's across multiple projects we're doing. And as noted earlier, the close of those term sheets typically come within four to six weeks or in parallel on handout.
Operator
Thank you. Our next question is from Nihal Choksky of Northern Capital Market. Your line is now open.
Yes, thank you. Andrew, you mentioned that you have some term sheets that are executable now. These are the asset-backed term sheets, correct? Can you disclose what's the value of those term sheets that are indeed executable right now?
We can't disclose that right now, but what I can tell you is we're dealing directly with primary banks and private credit, and we have things in front of us that we can sign.
Okay. And are you willing to look at the delayed draw term loan financing structure?
Yeah, I'm extremely familiar with the DDTLs, especially recent ones in the market. And then there's term loan B. So the way Boost Run decomposes the debt markets is we look at the DDTL structures and the term loan B. And if you just walk through those, the DDTL from a qualifying criteria Of inputs or DSCRs or coverage ratios, what I can say with a great degree of confidence, the way we've structured our contracts, all our ratios and covenants and coverage ratios are blowing through all the common metrics and inputs where decisions are made off of from the private and primary lenders.
And so, this would reflect a shift from leasing then, correct?
Yeah, we still have that optionality on the table and the ability to execute on that. So, would you decompose the stack of, call it the ecosystem of available dollars? To be funded, you start at a primary bank level, and that's essentially a pair of pursuits zone coming in or is it being syndicated. That can be even a direct lending relationship or relationship credit. And then below that sits the private credit that oftentimes the primary banks syndicate directly with the private credit, which is obviously a larger pool. And then a step below the private pool lies the leasing. Call it the common five to ten people who traffic in that. And we're engaged in transacting at all levels of those.
Nahal, this is Eric. I just wanted to add maybe one thing to that. So for every project that we work, we are utilizing a review process to best pair the financing with the project. And in certain cases, combining projects as a portfolio and financing on that level makes far more sense. On others, it makes more sense to go project by project. So as we look forward on this, I would expect that we'll still have a component of lease financing in our portfolio, maybe for several projects a year or something like that, that we might go on a single basis. But, yes, we are looking at different structures at this point and have progressed conversations significantly on a few of them.
Great. It would seem that these delayed draw terminals will be a better use of boost runs time. than the project-by-project lease financing. It really depends.
Yeah, Nahal, it really depends. It depends on cost and flexibility, speed, diligence, cash trap, all sorts of things, right? There are a lot of factors that we consider on this.
Understood. Something that you probably already answered, but I just want to make sure I understand correctly. The 2.7 billion TCV How much megawatts is that going to consume?
We have approximately around 40 right now in production. And then we have another Energize that are filling out additional projects that will be 30. So for the total, it will be 70. But you're going to see variants depending if it's B200, B300, GB300, and Orovera Rubin. So yours, yeah. And then another thing to, yeah, go ahead.
And then using that 1.4x TCV to CapEx ratio on page 7 of your presentation, that implies that there's about 2 billion PP&E that you need to support that 2.7 billion TCV, is that correct? Can I ask the question again, please? Uh, in order to support that 2.7 billion TCV sign, and using the 1.4x TCV to CapEx ratio on page 7 of your presentation, that implies that you need about $2 billion of PP&E to support that 2.7 billion TCV that's been signed so far. Is that correct?
Yeah, based on that number, but what I can tell you is our projects going live and ones that are pending going under contract, we're seeing the TCV to the project cost number go up greater than 1,4. So what happens then is that's going to impact the prepayment deposit amount. That percentage is going to increase. So our duration in contracts, we've been going into four- and five-year contracts, and then we feel we're going to even be stepping into six years. So that 1.4 reference number, we feel is going one way, and that's up. Yep. And that's deemed as a coverage ratio.
Therefore, that $2 billion PP in the estimate is probably high. Right. That's essentially what you're trying to say.
Yeah, you could look at it that way. Right. So you have a total project cost and capex amount and then minus your prepayment is going to be left with what is left for debt structure and or used from equity from the balance sheet or what I refer to as non-doluted financing. Yep. Understood. All right.
Operator
Thank you. At this time, we do have time for an additional question. Please wait a moment while I compile the Q&A roster. We have a question from Alex Arnold of Odeon Capital. Your line is now open.
Hey, guys. Just one quick one. Can you, just based on the last question, can you take that $2.7 billion of TCV and translate that on a fully deployed basis into ARR, roughly?
Yes, that can be done based on the 2.7 fully deployed, then it's a function of when. What we know is this is all deploying between now and into 27 here.
I guess that leaves average duration, so I'm just looking for a rough number of when you fully deploy your current backlog. You know, where are you going to be running on top line?
Yeah, fully, fully deployed across the 2.7 and aggregated with everything just in regard to these contracts. So that's going to put you around a 750 number. Okay. And that's safe to say not including anything outside with a circle on take-or-pay contracts right now. We have what we call ancillary revenue, Alex, which is no longer becoming ancillary. We don't like to talk about our past or reflect on it, but what I can tell you is we have a team that is world class in data storage and CPUs. In prior life, we've made an astronomical money monetizing, utilizing CPUs and storage, and we're seeing that really resonate with our end users. And that revenue base, outside of just GPU revenue, is starting to see substantial upticks, along with our Kubernetes software offering. So our deployment of automation and additional revenue lines continues to grow.
And I guess one other thing based on backlog. I mean, it's very early in terms of you standing up, you know, this compute. How should we think about the upsell, like customers expanding on top of their current contracts? Is that something that you'd expect, and to what magnitude would you expect that over time?
Alex, what I can tell you is from every single project we have in, and when I say every, we're being asked for more compute. Okay. So, I think, one, you have a scarcity issue, but two, that goes back to who are we, what we do, and how, and when you get into cutting no corners, reference architecture, going through the NVIDIA Architecture Review Board, you get a reputation of stability, and that's become a huge, huge issue, and earlier we had a question about drawdown term loans. What I also want to highlight here is the Boosted Run team is introduced to lenders, primary banks, private credit. We're getting feedback. What we're showing them has not been presented before, and that's in a handoff process of our projects that are getting sign-off for remediation and testing, which is going to have an impact on what our cost of borrow is. So when you talk about people expanding or wanting more compute, I think it's a safe assumption that our end users are pleased with performance. And at the current time, that is one of the hurdles we're going over right now is a role for it to expandability on a substantial client that we're in the process of negotiating.
Operator
Thank you. And we do have time for an additional question. If you would like to ask a question, press star 11 on your telephone and wait for your name to be announced. And please stand by a moment while we just wait for any additional questions.
Operator
I'm showing no further questions at this time.
Operator
It does conclude the question and answer session. We thank you for your participation in today's conference, and this does conclude the program. You may now disconnect.