Boost Run Below SPAC Prices

I’ve published 5 YouTube videos on the macro environment so far in October. Check them out to know which frameworks to follow and how to position your portfolio!

🌾 Welcome to StableBread’s Newsletter!

I first wrote up Boost Run (Nasdaq: BRUN), an AI data center company that rents out NVIDIA GPUs to AI startup companies, on April 15, 2026, when it still traded as Willow Lane (WLAC) at $10.70/share, just above its ~$10.60 trust value.

The merger was a success, and the stock peaked at $39.20/share on June 1. By my August 16 update, where I said I’d buy more below $20/share, the stock was at $22.93.

Now BRUN has fallen to $8.89/share, down 77.3% from its peak, even after Boost Run announced its largest disclosed contract on October 6 (worth $525.6M):

Worth mentioning that the three larger listed GPU clouds have fallen from their 2026 highs as well, through the October 16 close:

So Boost Run fell with the group, but further than all three (down 77.3%).

Meanwhile, Boost Run has signed more contracts, delivered more GPUs, and grown its cash balance since my August update.

At $8.89/share, Boost Run’s EV is $989.7M. That’s 2.5x its $400M target for year-end ARR, down from 5.4x at the $22.93 price in my August 16 update.

Below, I break out the $2.6B TCV, discuss the higher rates and Boost Run’s ~3.1x larger float behind the drop, how Boost Run is paying for its GPUs, and the insider selling and RSU grants since my August update. Then I conclude with my valuation (Excel model included) and whether I’m adding.

Cohere agreed on September 30, 2026, to pay Boost Run $525.6M over the agreement’s term for NVIDIA’s GB300 NVL72 racks, each with 72 Blackwell Ultra GPUs.

Part of the $525.6M is a prepayment Boost Run hasn’t disclosed. But if we follow the 22% average Boost Run says its prepayments run, the prepayment would be ~$116M.

Each rack’s term runs 60 months, and the first racks should start in Q2 2027.

Cohere can terminate and get back everything it prepaid if it hasn’t accepted a minimum number of racks by July 15, 2027 (8-K doesn’t specify amount of racks).

Boost Run has disclosed terms for seven contracts, and the two largest are with Cohere and Thinking Machines Lab, the AI startup founded by former OpenAI CTO Mira Murati:

Annualized revenue is TCV divided by the term in years. Unnamed customers are labeled by what they rent, including NVIDIA’s current B300 (Blackwell Ultra) and older H200 GPUs.

Boost Run’s TCV is over $2.6B, so the other $989M+ is in deals without disclosed terms.

The October 6 release calls the $2.6B “total revenue customers have committed to pay,” but the August 14 Q2 deck also counted renewal options and LOIs in TCV, and Boost Run hasn’t said whether the definition changed.

So the conservative read counts only the seven contracts with disclosed terms, worth $1.61B, not the $2.6B headline.

Five of the seven can count toward the $400M year-end ARR target, since the July order and Cohere, $160.7M/year combined, don’t start until 2027.

If Thinking Machines’ GPUs are fully running by December 31 and the June order starts on November 30, the five cover $300.6M (75.1%) of the target, leaving $99.4M for contracts without disclosed terms.

Here’s how much of the $2.6B TCV and the $400M ARR target comes from contracts with disclosed terms:

For comparison, the Q2 deck’s deployment schedule had $0.4B of the then-$1.9B of TCV in production by the end of Q2, with all of it live by Q1 2027:

From there, the deck shows another $0.4B of TCV going into production in Q3 and another $0.4B in Q4, a straight line to $1.2B by year-end that management calls consistent with the $400M+ ARR target.

The deck’s footnote says the Q3 figure is a management estimate “interpolated from disclosed metrics,” so the actual climb will likely be lumpier, with the June order alone starting November 30.

Cohere’s racks start in Q2 2027, right after the deck’s schedule ends, so even without a new contract, Boost Run has signed deployments lined up through mid-2027.

Boost Run’s last ARR figure was $145M as of June 30, from the Q2 call, so the Q3 report (expected in mid-November) will be the first check on H2 2026.

A straight-line climb from $145M (as of June 30) to $400M by December 31 would mean $136.3M of H2 2026 revenue, in line with the $127.9-137.9M needed to hit the $170-180M of 2026 revenue its October 8 prospectus amendment still estimates.

By quarter, that climb works out to $52.2M of revenue in Q3 and $84.1M in Q4, up from $31.1M in Q2:

In other words, the year-end target is mostly contracted, and the risk is getting the GPUs bought, financed, and running by December 31.

Boost Run’s June 1^st peak came within six weeks of its peers’ highs.

When I wrote my June 3 update, Boost Run’s EV was 7.6x its year-end ARR target, measured the same way as today’s 2.5x, and its float was only ~17.5M shares (12.65M public shares plus 10% of insider and sponsor shares that were never locked up).

Neither the GPU cloud stocks’ highs nor Boost Run’s small float lasted.

It doesn’t help that the 10-year Treasury yield has risen from 4.45% on June 1 to 5.08% on October 16. Reuters rightfully tied the October 14 drop in AI infrastructure stocks to rising bets on another Fed rate hike.

Higher rates hurt GPU clouds in two ways, since they (1) raise the cost of the debt that pays for their GPUs and (2) lower what investors will pay today for profits that arrive years later.

From June 1 to October 16, Boost Run fell 77.3%, vs. 37.4-58.7% for CoreWeave, Nebius, and IREN:

But all three kept or raised their 2026 targets:

The problem is that lenders are charging more now. CoreWeave’s 9.625% notes due 2032 yielded 14.74% on October 16, up from 13.19% on October 1, meaning CoreWeave would likely pay ~15%/year to borrow the same way today.

So the selloff came from how investors now see the cost of funding the demand for GPU clouds, not from the demand itself.

Boost Run also had a problem its peers didn’t, since its float more than tripled. In my May 17 post, I expected most insider and sponsor shares to stay locked up until mid-November, based on an S-4 sentence that read as if only 10% could be released early.

But the July 2 S-1 shows the lockup ended for all of those shares once the stock closed at or above $12.00 on 20 of 30 trading days after the closing, which by my count happened on June 8, 2026.

The resale registration for 58.7M shares then went effective on July 10, and after Boost Run called its warrants on July 21, exercises added 11.45M new shares by August 20.

That took Boost Run’s float from ~17.5M shares to all 54.3M Class A shares, ~3.1x as many, between July 10 and August 20, while the stock fell from $28.44 to $19.40:

So I’d blame the drop from $39.20 on two causes that have nothing to do with Boost Run’s core operations: (1) higher rates and borrowing costs, and (2) Boost Run’s float growing ~3.1x, which was expected once its lockup ended and its warrants were called.

Management expects its 2026 GPU deployment to require $1.0-1.2B of gross capital investment.

Boost Run plans to pay for the GPUs with (1) customer prepayments and operating revenue and (2) equipment leases that spread the cost over the lease term and generally end with an option to buy the GPUs outright, so only part of the cost is paid in cash up front.

During H1 2026, lessors (the companies that buy the GPUs and lease them to Boost Run) funded $268.3M of new GPU equipment through finance leases. Boost Run made $39.4M of advance payments to secure those leases and paid $29.6M in cash for equipment it bought outright, so lessors funded 79.5% of its $337.3M of new GPU equipment.

The $268.3M came from 34 separate GPU leases with 30- to 36-month terms, and the June 30 lease schedule implies an interest rate of ~7%.

Its finance leases as of June 30 call for $103.5M of payments in 2027 and $99.8M in 2028, then just $12.1M in 2029, about when its three-year contracts end. So the lease debt runs off with contracted revenue, not with renewals Boost Run still has to win.

Customer deposits, the payments customers make up front on their contracts, were $128.4M as of June 30. New deposits made up $112.9M of H1 2026’s $114.2M of operating cash flow, while Boost Run paid $99.9M on its finance leases over the same six months.

Effectively, customers’ prepayments are already paying for the GPUs they rent.

Founder and CEO Andrew Karos said on the Q2 call that every deal Boost Run closes comes with a prepayment, averaging 22% of TCV. The June and July orders alone carry $93.6M due by their start dates ($37.9M on November 30 and $55.6M in installments through March 2027), before counting Cohere’s prepayment.

Karos also said Boost Run’s TCV-to-capex ratio, its contract value divided by what the GPUs cost, keeps rising to 1.4 and above as contracts get longer. So a three-year contract brings in ~$1.40 for every $1.00 of GPUs, which also has to cover the lease interest, the rent, and the staff.

In other words, a three-year contract mostly covers the GPUs and the cost of running them, so how much Boost Run makes on those GPUs depends on renting them out again after the leases end in 2029, without lease payments, at whatever older B300s fetch by then.

Fixed Commitments

The GPUs also need space and power, which Boost Run rents in other companies’ data centers (colocation).

The four leases signed since April already cover the space and power for the year-end deployments, and Karos said its partnerships now reach 253 MW of power.

Boost Run owes the rent on those sites whether or not the GPUs get financed.

I plotted how long the four leases and Boost Run’s other fixed commitments run next to the contracts with disclosed terms:

In 2027, the commitments Boost Run has already signed call for ~$221M of payments, including $103.5M on its GPU leases, $28.9M of rent on the sites in use as of June 30, ~$69M on the four new leases, and $20M for software.

The $20M is the second of five yearly installments, starting in October 2026, on a $100M software license and support agreement Boost Run signed on April 17.

Boost Run’s $191.8M of pro forma cash and $93.6M of contracted prepayments, $285.4M together, cover the ~$221M of 2027 payments before counting any monthly billings from customers. But the same cash also pays Boost Run’s share of the H2 2026 build, so its monthly billings will have to cover part of the 2027 payments.

H2 2026 Build

Here’s what’s funded and what isn’t:

To reach the $1.0-1.2B of GPU spending management expects for 2026, Boost Run has to deploy $663-863M of GPU servers in H2 2026, 2.3-3.0x its cash and contracted prepayments.

If lessors fund 79.5% of the H2 2026 build, as they did in H1 2026, Boost Run needs $527-686M of new leases, 2.2-2.9x the $238.1M the company owed on finance leases as of June 30. The other $136-176M would be paid in cash.

The $527-686M works out to $264-343M/quarter on average across Q3 and Q4, 1.3-1.6x the $208.2M of leases Boost Run took on in Q2.

Boost Run’s lease debt would rise by less than the $527-686M, since the company keeps paying down its leases while adding new ones. During H1 2026, its finance lease debt rose $207.7M on $337.3M of new GPU equipment, or 61.6%, which for H2 2026 would imply ~$410-530M more debt.

New leases will likely cost more than ~7% after the rise in rates. For context, at the 9.0% IREN agreed to pay on its August GPU financing, the extra interest on $527-686M would be ~$6.5-8.5M/year, 1.6-2.1% of the $400M ARR target.

And Boost Run isn’t slowing down. By June 30, the company had already placed orders well above the $578M first-year minimum in its April 2026 agreement with Dell, which also requires at least $215.5M/year of Dell equipment from April 2027 to April 2031 or a payment for the shortfall.

New Financing

So can Boost Run borrow the $527-686M it needs for H2 2026 without selling stock?

To begin, CFO Erik Guckel said on the Q2 call that Boost Run is working on new ways to finance the build:

The October 8 prospectus amendment names asset-backed securitizations (bonds backed by its GPUs and customer contracts) and rated parent-level debt as new forms of financing Boost Run expects to access. But I found no new facility, lender, or term sheet in its filings or press releases from August 14 to October 16.

Karos acknowledged on the Q2 call that there’s “a big gap in the market right now” in financing for customers without investment-grade credit. Boost Run’s named customers are private AI startups without investment-grade ratings, so its contracts are harder to borrow against.

For comparison, Nebius’s first secured loan on July 17, $775M at 2.50 points over SOFR (a short-term benchmark rate), was backed by GPUs running under a contract with an investment-grade customer.

Thinking Machines alone accounts for $157.2M/year, 52.3% of what the disclosed contracts should bring in by year-end, and was still in talks to raise $5-6B as of early September. I found no announcement that it closed the round through October 16, so lenders financing its GPUs are also betting that the round closes.

Karos also mentioned NVIDIA’s AI Compute Partnership (AICP), a program NVIDIA launched in July to backstop GPU clouds’ clusters with its investment-grade credit, for a share of their revenue. However, NVIDIA paused some of its AICP deals over antitrust concerns, the Wall Street Journal reported on August 27.

NVIDIA said the program is still in place, but The Information, a tech news site, reported on October 5 that NVIDIA is rethinking its revenue-sharing deals with GPU clouds.

Still, lessors funded 79.5% of the H1 2026 build for GPUs already rented out on contracts worth ~1.4x their cost, so I believe Boost Run most likely borrows the $527-686M without selling stock, which is my base case below.

The Q3 report (mid-November) will show how much Boost Run leased in Q3. If lessors funded 70% of the H2 2026 build instead of 79.5%, Boost Run would need $199-259M of its own cash for the build, more than its $191.8M of pro forma cash.

Covering the ~$7-67M gap and keeping three months of the ~$221M of 2027 payments in cash would take a $62-122M stock sale. At $8.89/share, the sale would mean 7.0-13.8M new shares, 8.4-16.4% of the 83.8M outstanding.

So the stock at $8.89 depends less on demand or the book Boost Run has signed than on whether it can finance the next ~$1B of GPUs without selling shares at beaten-down prices.

Karos hasn’t reported a single sale since the listing. He takes a $1.00/year salary, owns 44.6% of the shares, and controls 87% of the vote, so his pay comes almost entirely through the stock.

His stake is all 29.5M high-vote Class B shares plus 7.9M Class A earnout shares, which the merger agreement paid him once the stock held above $12.50, $15.00, and $17.50.

Since my August update, the only open-market sales came from director Sean Goodrich, who holds his stake through Goodrich ILMJS LLC, an investor on Willow Lane’s side of the merger.

Goodrich’s sales came after a forced warrant exercise.

Once BRUN closed at or above $18.00 on 20 of 30 trading days, Boost Run could redeem its warrants for $0.01 each, so after it called them on July 21, Goodrich paid $12.7M on August 19 to exercise his 1,101,986 warrants.

He then sold 198,200 shares at $20.15 on August 27 and 500,000 at $14.00 on September 14 for $11.0M total, 86.8% of what the exercise cost him, and still holds 4,437,921 shares (8.2% of the Class A).

So I don’t read the sales as a verdict on the business, since he sold about what the exercise cost him when he could have sold far more.

Meanwhile, three executives have more stock coming. The compensation committee granted them 4.59M time-based restricted stock units (RSUs) on September 29, 5.5% of the shares outstanding.

Each RSU turns into a share if the executive is still there on its vesting date:

At the $16.65 grant-date price, the 4.59M RSUs are worth $76.4M, 42.5-45.0% of the $170-180M 2026 revenue estimate.

Guckel’s grant matches, almost share for share, the 2,890,761 RSUs, performance shares, and options his May 8 employment agreement made him eligible for.

Giving up 5.5% of the shares is expensive, but the grants give the CTO who runs all of Boost Run’s GPU deployments and the CFO who has to finance the GPUs a reason to stay through mid-2029.

Controls

Boost Run still reports three material weaknesses, the most serious kind of gap in a company’s financial reporting controls.

They’re in (1) IT general controls, (2) segregation of duties (keeping one person from handling a transaction end to end), and (3) the number of qualified accounting staff.

Boost Run’s filing record hasn’t been clean either. Both 10-Qs came late, after notices of late filing on May 26 and August 14, and the original Q1 10-Q, filed June 1, called its disclosure controls effective until an amended Q1 10-Q on August 18 said they weren’t.

I’m comfortable with the weaknesses for now, since (1) the S-4 disclosed all three before the merger, (2) the amendment didn’t change any reported numbers, and (3) Boost Run has hired a chief accounting officer, a COO, and a CIO to fix the gaps.

But lenders, and any agency rating the parent-level debt Boost Run wants, will read the same filings, so a third late 10-Q would push me to reconsider.

Here’s the share count and EV build in my model:

Pro forma cash includes the $134.7M of cash Guckel reported as of August 12, plus $57.1M of later warrant exercises, and leaves out the $13.0M of restricted cash that backs a data center letter of credit.

The 88.8M fully diluted shares are the 83.8M outstanding plus the 4.59M RSUs granted on September 29 and the 336,000 shares director B. Luke Weil earned as a consultant, which Boost Run hasn’t issued yet.

Let’s now compare Boost Run to the three larger listed GPU clouds to see where it stands: