What does it do?
Boost Run is a hypergrowth business that earns money by leasing high-performance GPU and CPU compute power to companies training large-scale artificial intelligence models. The company builds and operates specialized data centers where customers pay recurring fees to access clusters of NVIDIA chips through a managed cloud platform. Money flows through long-term contracts where clients commit to a specific amount of compute capacity, typically over several years. Boost Run handles the complex orchestration, cooling, and power management, allowing enterprise customers to scale their AI projects without the multi-billion dollar cost of building their own physical infrastructure.
Where does revenue come from?
Almost all revenue comes from high-performance computing services where customers pay for dedicated GPU and CPU cluster access. The company also earns fees from managed storage and Kubernetes orchestration, which helps developers organize their AI workloads across thousands of computing nodes. While the company is currently focused on the United States, its six active data center sites serve a global base of enterprise AI developers and research institutions.
Who are its customers?
Boost Run serves enterprise-level AI developers and high-performance computing labs that require massive, specialized compute clusters to train their models. The company currently manages a total contract value of $1.9 billion, which includes over $1 billion in new contracts signed during the second quarter of 2026 alone. Because the business is in a rapid scaling phase, it focuses on a smaller number of large-scale enterprise contracts rather than individual users. These customers sign long-term agreements for dedicated power and hardware, with the company currently managing 253 megawatts of total accessible power across its data center partnerships.
What gives it staying power?
Boost Run relies on its NVIDIA Preferred Cloud Partner status and technical validation on the Blackwell architecture to stay ahead of rivals. This certification creates high switching costs because AI developers build their software pipelines specifically for Boost Run's optimized Blackwell clusters.
Where is it headed?
The company is making a massive strategic bet on expanding its physical GPU capacity through a $1.44 billion hardware purchase agreement with Dell. This expansion is designed to convert its nearly $2 billion backlog into realized revenue. Management is also aggressively securing an additional $4 billion to $5 billion in compute hardware to meet what they describe as "unconstrained" demand.
Revenue has accelerated dramatically to $31.1 million in the most recent quarter, representing a 270% increase compared to the $8.4 million reported a year ago. This growth is driven by the rapid energizing of new GPU clusters that were sold out long before the hardware was even installed.
The business is currently burning cash as it spends billions on hardware, with a massive $1.44 billion purchase agreement recently committed to Dell. While free cash flow is negative today, the company targets a sustainable 15% to 20% net cash flow margin once these hardware clusters are fully deployed and generating revenue.
The balance sheet was recently strengthened by a $130 million cash infusion from warrant redemptions, leaving the company with a simplified capital structure and a solid liquidity cushion. This cash, combined with the $114.1 million in net proceeds from its public debut, provides the funding needed to secure long-term leases for data center power.
Boost Run is a high-growth infrastructure play with a massive contract backlog that should drive profitability as it moves from the construction phase to the utilization phase.
Boost Run does not pay a dividend and does not buy back shares, as it currently reinvests every dollar of capital into purchasing AI hardware and expanding its data center footprint. The company's priority is scaling to meet its $1.9 billion contract backlog, which leaves no excess cash for distributions to owners. Because it recently went public via a merger, the share count is expanding to fund growth rather than shrinking through buybacks. Investors should hold this stock for its potential for massive future earnings growth rather than for current income.
Contract signing is outpacing hardware deployment, with over $1 billion in new contracts added in the last quarter alone. This provides high visibility into future revenue because demand is currently so strong that capacity is often sold out before the chips are even delivered.
The single biggest risk is the execution of the hardware build-out, specifically the company's ability to procure $4 billion to $5 billion in new compute hardware. Any supply chain delays from NVIDIA or Dell would prevent the company from converting its signed contracts into actual revenue.
The specialized AI cloud infrastructure market is a subset of the $230 billion infrastructure software market and is growing at over 30% annually as companies shift from general cloud to GPU-heavy clusters. This industry currently holds its prices well because demand for the latest AI chips far exceeds the available supply in data centers. Boost Run stands as a specialized challenger to the giant hyperscalers, offering deeper technical optimization for AI training that makes it a primary choice for high-end research and enterprise labs.
Competitive dynamics are brutally capital-intensive, as winning requires billions of dollars in upfront spending to secure the latest chips. Barriers to entry are high because NVIDIA only grants "Preferred" status to a handful of providers who can prove they meet strict technical standards. This scarcity of hardware creates a rationally structured market where price is currently less important than guaranteed access to compute power.
CoreWeave and Nebius are the most direct threats, as they also specialize in GPU-only clouds and have secured massive funding to lock up data center power. The giant hyperscalers like Microsoft and Amazon represent a different threat, as they could eventually commoditize the market if chip supply ever catches up to demand. The most dangerous threat comes from CoreWeave, which has a larger existing footprint and similar early access to NVIDIA's latest Blackwell architecture.
Boost Run is aggressively gaining share in the specialized niche, as evidenced by its contract backlog doubling in a single quarter.
Boost Run's protection comes from its technical validation and specialized configuration for the NVIDIA Blackwell architecture. This intangible asset allows it to offer performance that general-purpose clouds cannot easily replicate without expensive custom retooling. The company achieved NVIDIA Exemplar Cloud status, proving its clusters perform within 5% of NVIDIA's own reference targets.
Gross margins of 53.9% show that the company can price its specialized compute at a premium, as customers are willing to pay for performance and guaranteed availability. This is consistent with a narrow moat that is currently being built through hardware scarcity and technical expertise. A well-funded rival could eventually build the same data centers, which makes this protection dependent on staying first in line for the latest hardware.
The moat is strengthening as the backlog grows. Signed contracts for $1.9 billion create massive switching costs because customers have already integrated their development pipelines with Boost Run's specific Blackwell orchestration.
Signed over $1 billion in new contracts in a single quarter.
Simplified capital structure by redeeming warrants, raising $130 million.
Founder-led with significant stake; successfully steered company through SPAC merger.
Capital Allocation Track Record
Management has demonstrated exceptional strategic judgment by securing a $1.9 billion contract backlog before most of its hardware was even energized. Founder and CEO Andrew Karos has successfully navigated the company through its transition to the public markets while raising the capital necessary to secure a massive $1.44 billion hardware pipeline. This ability to match aggressive capital spending with signed customer commitments shows a disciplined approach to risk in a high-stakes industry.
The primary governance risk is the founder's central role in driving the technical vision and the critical relationship with NVIDIA. While the executive bench includes experienced operations and finance leaders, the company's "Preferred Partner" status likely relies heavily on the founder's leadership. There is a clear dual-class structure that gives the founder control, which is common in early-stage tech but means shareholders must trust his long-term expansion strategy.
We expect revenue to grow from $0.2B in FY2026 to $1.8B in FY2031 (~60% CAGR), with EPS growing from $-1.13 to $4.83. Demand for specialized AI compute clusters is outstripping supply, allowing the company to take a small but growing share of the massive high-performance computing market. The massive upfront costs of building data centers are spread across more customers as the network reaches higher utilization. EPS grows faster than revenue because the company moves from heavy initial losses to consistent profitability as it scales. Operating margin expected to reach ~20% by FY2031.
Converting $1.9 billion backlog into realized recurring revenue. Converting the current massive TCV into realized revenue would push ARR well past the $400 million target and stabilize cash flows.
Expanding compute hardware procurement by $5 billion. Securing additional Blackwell architecture hardware allows the company to capture "unconstrained" demand that is currently outstripping its supply.
Achieving 15% to 20% sustainable net cash flow margin. Moving from the heavy investment phase to operational maturity will reveal the high operating leverage inherent in the GPU-leasing model.
Supply chain delays for NVIDIA Blackwell chips. Any delay in receiving the $1.44 billion Dell order would prevent the company from meeting its contract obligations and slowing revenue growth.
Large contract cancellations as AI demand shifts. If enterprise customers pivot away from large model training, the company could be left with expensive hardware leases and no revenue.
Commoditization of GPU compute as hyperscalers increase supply. If Microsoft or Amazon flood the market with GPU capacity, Boost Run's ability to charge premium leasing rates will disappear.
Below is our estimate of current and future fair value, with detailed reasoning and assumptions. Fair value is a judgment, not a fact, and other analysts will likely land on different numbers. Use it as one data point in your research, and apply your own discretion in any investing decision.
We value Boost Run based on what it will earn in 2031 and discount that future value back to today. This method captures the long-term value of the company’s $1.9 billion contract backlog, which a simple multiple of next year's earnings would ignore. Because the company is currently in a heavy building phase, its true earning power will not be fully visible for several years.
Applying a 25x multiple to the 2031 profit of $4.83 per share gives a future price of $121, which equals $71 today after accounting for a 10% annual discount. We chose a 25x multiple because it aligns with mature technology infrastructure giants like Oracle (28x) and Microsoft (31x) while staying conservative for a newer player. These profit estimates come directly from the deterministic model, which tracks the expected ramp-up of the company's massive contract backlog over the next five years.
Valuing the company on next year's sales instead, at a 10x multiple, we get $68—which is very close to our $71 fair value. This second method takes the 2027 revenue estimate of $551 million and divides it by the 81 million shares outstanding. Since high-growth AI infrastructure rivals trade between 8x and 15x sales during their expansion phases, a 10x multiple is a reasonable and conservative middle ground. The close agreement between the profit-based and sales-based methods gives us higher confidence in the $70 price range.
The biggest risk is the massive amount of money Boost Run must spend to build out the data centers needed to fulfill its $1.9 billion in contracts. If the company cannot secure affordable financing or faces significant construction delays, the fair value would likely drop toward the $15 bear case. Watch the cash balance and debt-to-equity ratio closely to see if the company is running out of runway before the contract cash starts flowing in.
Bear case ($15): Deployment delays or chip supply issues cause FY2026 exit ARR to fall below $300 million; or Net Cash Flow Margin fails to reach the 15% target due to rising electricity or financing costs.
Bull case ($125): Accelerated data center expansion pushes FY2026 exit ARR above $500 million; or The company wins an additional large-scale enterprise contract exceeding $500 million in total value.
Clearthesis wrote this report from 40 sources, including SEC filings, analyst estimates, industry research, and recent news.
How did you like this thesis?
Your feedback helps us make reports better for you
© 2026 Clearthesis.ai · Report generated on September 11, 2026
This is an AI-generated analysis for informational purposes only and does not constitute financial advice. Data and analysis may not reflect recent developments if viewed significantly after the generation date. Always conduct your own due diligence before making any investment decisions.