Jonah Lupton
CEO & CIO of FirstWave Capital, Long-Biased Growth
Customer prepayments and delayed-draw debt could finance Boost Run's prospective $4-5B hardware procurement and change its TCV and ARR outlook. Management mentioned the procurement, and an analyst spoken to confirmed the delayed-draw term loan. Customers pay 22% upfront on 3-5 year deals, while 40-45% of hardware cost converts into ARR; on that basis, TCV and ARR could look very different soon. x.comÂ
Boost Run's $525.6 million five-year deal with a leading sovereign AI company is small next to the bigger contract that could or should come, which could carry the stock far higher within two to three months. The agreement, for Nvidia GB300 NVL72 compute starting in early Q2 2027, brings its total contract value to over $2.6 billion. x.com x.comÂ
Boost Run's delivery ahead of schedule and ability to capture higher current compute pricing strengthen its upside case. A release this week shows faster delivery, while management says it will not dilute with equity; guidance could imply $2B ARR exiting 2027. Unlike CoreWeave, it is not locked into stale contract terms. Short interest is 26%, even as the shares are down 75% from their highs and fundamentals improve. x.comÂ
Boost Run is a high-risk, asymmetric AI-infrastructure bet with real contracts and Nvidia tailwinds, worth a small position if it hits $375 million-plus ARR and proves its platform scales. It has a $940 million contracted backlog, just won a $472 million deal for 5,000 B300 GPUs for Thinking Machines Lab, and deploys GPUs faster than hyperscalers. But at 60-80x sales on $27 million of revenue, loss-making, with high debt and a delayed 10-Q, it can easily drop 50% on a miss or AI cooldown. x.comÂ
Boost Run looks like a leasing company rather than a neocloud, left to compete on price forever. It owns little: GPUs are leased, data centers are rented colocation, and its 125 MW is mostly others' capacity. It sells undifferentiated bare metal, the layer even CoreWeave is leaving, and about 21% of FY25 revenue came from pointing idle GPUs at crypto networks amid a GPU shortage. Its customers are leftovers, and a $1.44B Dell commitment dwarfs about $90M raised, after a bridge loan default. … Show more
AI agents' productivity over human labor should lift GPU demand and rates for neoclouds. The bottom looks in for CoreWeave, Nebius, Boost Run, Sharon AI and IREN. Gavin Baker says token prices can fall as revenue per GPU rises; paraphrasing Elon, even $38 per GPU-hour can pay off. CoreWeave's five-year fixed contracts may lag current rates; Boost Run has less capacity locked in. If SpaceX takes more growth, slower cash burn could rerate neoclouds, though smaller customers may pay less. x.com … Show more
Nebius's already-raised cash and expected customer prepayments should buffer its 2027 plans against a short-term debt-market freeze. Rising GPU rental prices could put new deals at $35m+/MW annually, up from last quarter's $20m-$25m, reducing estimated payback from 22 months to around a year. If Nebius deploys 1 GW next year at a one-year payback, equity value creation could be tremendous. x.comÂ
Boost Run's lock-up expires within days, letting insiders and the CEO sell their shares, which could cause massive selling pressure. That is a big deal and a huge risk against the earlier bull case for the newly listed AI infrastructure provider, built on Nvidia Preferred Cloud Service Provider status, a $1.44B Dell purchase agreement and about $471.7M of take-or-pay contracts for 5,000 Nvidia B300 GPUs. x.com x.comÂ
Getting compute six months early can be worth $20B in high-margin revenue, dwarfing the cost of a 50bp rise in rates. On a $50B, 1GW data center, that rise adds $250M in interest, or 1.25% of the cited revenue; most centers are smaller. That leaves a massive buying opportunity in CoreWeave, Nebius, Oracle and Boost Run despite higher yields. x.comÂ
Boost Run's contracted utilization answers the depreciation bear case for neoclouds. GPUs are depreciated over 4-6 years, so GAAP margins can look thin while free cash flow is positive; the real risk is chips going obsolete faster than assumed. Thinking Machines' take-or-pay deal is non-cancelable and pays for the full term, locking utilization through the depreciation window. The de-SPAC lockup is an overhang, but insiders with control and earnouts rarely dump, so dips are a chance to add. … Show more
Boost Run faces a near-term lock-up overhang. If the stock holds above $12 for 20 of 30 trading days, an early-release trigger around June 5-9 lets about 24M shares, roughly 3.3x the float, become eligible to sell. A similar post-SPAC release hit CoreWeave 15-22% over 30 days, and typical recent de-SPACs fell 25-40% over 60 days. This is a mechanical event, not a thesis change: Boost Run guides ARR up 1,233% to $400M+ by FY26 year-end and is FCF positive. Remaining shares unlock November 8. … Show more
Oracle's plan to fund its AI buildout from cash flow makes last week's financing fears look completely overdone. It expects capex to peak in FY27-28, free cash flow to turn positive in H2 FY28 and top $30B in FY29, funding growth from then. Oracle plans at most $20B in further capital raises through FY30, excluding refinancing, after an already completed $20B equity raise. Vendor financing, customer prepayments and customer-supplied hardware should reduce upfront cash needs. x.com x.comÂ
