A new reality is here. Even trillionaires are struggling to rent enough AI compute. Shopify CEO Tobi Lütke converts his wine cellar into a data center. Some plan to deploy in space, some in the ocean. Some, like Safeera CEO Panin, want to take a computer into your backyard, paying you for mining intelligence.
Even Trillioners Can’t Find GPUs
Elon Musk says he cannot find the cloud compute he needs. On SpaceX’s October 20 earnings call, he compared the experience to Monty Python’s cheese shop: requests, but no available capacity. His team sought hardware ranging from GB300s to older H100S, yet existing providers simply don’t have supply.
Hyperscalers Struggle To Deploy What’s Promised
On the other hand, CoreWeave reported a $104 billion revenue backlog as of June 30. That figure is subject to delivery and service requirements, with 1.5 GW of active power against 3.7 GW contracted. The company added more than $25 billion in net new commitments early in the next quarter. On September 17, it announced three-to-six-month contracts priced at approximately $40 million in annualized revenue per megawatt. That is before equipment, electricity, financing, and service costs.
On the All-In Podcast, Ben Horowitz cited a conversation with Amazon CEO Andy Jassy. Jassy had told him the shortage of powered, cooled compute extended beyond 2028. Jassy’s 2025 shareholder letter separately stated that AWS added 3.9 GW that year and still had unserved demand.
“Intelligence Factory” In A Wine Cellar
NVIDIA CEO Jensen Huang compared NVIDIA infrastructure to “fine wine” in an August CNBC interview. He said software improvements increase installed systems’ productivity, and some cloud operators reported infrastructure worth more than its original cost. Shopify co-founder Tobi Lütke might take Jensen’s words close to heart. He has literally converted his home’s wine cellar into a “rack cellar,” packed with GPUs. Safeera CEO Vlad Panin seemed to pick up on the idea that the kitchen team is developing a gigantic egg-shaped data center he wants to put inside every American house. He calls it a Domesticated Intelligence factory. The “egg” combines GPUs, networking, and liquid cooling with a capacity marketplace. Panin’s vision clearly resembles rooftop solar: install productive equipment on suitable property, use some output, and sell available capacity. During TechWeek earlier this October, Safeera showed a noiseless 16-B300 demonstration in the Los Altos neighborhood, in the SF Bay Area. Both Vlad Panin and NVIDIA CEO Jensen Huang reside there.
The Math On A “Smart Egg”
As for the numbers, Safeera’s financial model values an 8-GPU pod at $825,000. Of that, $132,000 is paid upfront, and the balance is financed over six years at 7%. It projects $40K in monthly customer cash flow. That implies roughly 20 months to recover the initial payment, with financing payments continuing. It also implies about a 50% first-year cash-on-cash return, not including tax benefits from deductions. According to the Big Beautiful Bill, those might be as high as 100% of hardware cost plus operational expenses.
Hosts could sell it to the Safeera marketplace or reserve capacity for their own confidential document analysis, financial records, research, and other applications, keeping that work on-site. Deployment requires quite significant due diligence to ensure such enormous compute usage will comply with US law. First commercial deliveries have already started. Yet given Elon Musk’s feedback on GPU scarcity in general, good luck receiving your order earlier than 6 months.
An Intelligence-Factory Unit In Every Backyard
Tech CEOs like Lütke (Shopify) and others are deploying compute at home. We may soon live in a world where every American has an intelligence-mining unit in their backyard. At least, that’s how some of these companies view our tomorrow. Panin, for example, believes the world will only be safe when intelligence is distributed equally between people rather than corporations, and literally named his venture SafeEra. Yet how safe such a concentration of domestic intelligence is remains to be seen.
A seasoned technology executive with a proven record of developing and executing innovative strategies to scale high-growth SaaS platforms and enterprise solutions. As a hands-on CTO and systems architect, he combines technical excellence with visionary leadership to drive organizational success.
