By Kokab Rahman, Founder & CEO, Radeya Global
On 25 October 2026, Elon Musk stated on X that SpaceX had closed approximately $40 billion in debt financing primarily to purchase Nvidia GPUs, with deliveries expected in the first and second quarters of 2027. The financing is largely callable debt structured to remain off SpaceX’s balance sheet. Musk attributed the ability to complete the deal to the company’s investment-grade rating and revenue from Starlink and its launch business. He simultaneously warned that debt markets have “almost entirely shut off access to money for any company that doesn’t have substantial revenue independent of AI data centers,” predicting widespread failures and consolidation among more than 200 neocloud providers.
This announcement arrives days after SpaceX’s third-quarter results, in which its Super Intelligence segment (formerly reported as AI) generated $7.56 billion in revenue—surpassing Starlink and the launch business combined—and posted its first quarterly operating profit of $431 million. Most of that revenue came from leasing compute capacity.
The scale of the debt raise underscores the capital intensity of frontier AI infrastructure. It also highlights a growing bifurcation: companies with diversified, investment-grade cash flows can still access large-scale financing, while pure-play AI compute providers face tightening credit.
The Payoff Gap Outside Big Tech
At nearly the same moment, Apollo Global Management chief economist Torsten Slok published a note concluding that AI has so far delivered “zero financial payoff to corporate America outside of tech.” Examining the S&P 500 excluding the Magnificent Seven (Alphabet, Amazon, Apple, Meta, Microsoft, Nvidia, and Tesla), he found revenue per worker and profit margins essentially flat, with no visible lift in utilization-adjusted total factor productivity. Mainstream companies, he wrote, are beginning to treat token and model spending as a cost center that pressures margins rather than a source of growth. Spending remains highly concentrated: the top tier of firms accounts for the overwhelming majority of inference and neocloud outlays.
For executives, the implication is straightforward. Large infrastructure commitments are being underwritten by a narrow set of technology platforms and their customers. Broader enterprise adoption has not yet produced the measurable operating leverage many board presentations assume. Companies that treat AI primarily as a productivity tool rather than a revenue or margin driver should re-examine expected payback periods and governance thresholds before expanding spend.
The Opportunity Divide
The same week, a senior official at the UN’s International Telecommunication Union warned of an emerging AI “opportunity divide.” Africa and Latin America together hold only about 3.3 percent of global data-center capacity, roughly 2.2 billion people remain offline, and generative-AI use among working-age adults stands at approximately 29 percent in developed economies versus 16 percent in developing ones. Productivity gains concentrated in higher-income countries risk widening existing socioeconomic gaps.
Parallel efforts are underway. The African Development Bank and UNDP are advancing an AI 10 Billion Initiative aiming to mobilize $1 billion annually for a decade, beginning in five countries and expanding toward 20 by the end of 2027, with an estimated contribution to 25 million jobs by 2035. Latin American analyses similarly emphasize that ownership of strategy, compute, and skills—not merely access—will determine whether the region captures meaningful gains.
Implications for Executives, Businesses, and Job Seekers
Executives and businesses. Access to capital for AI infrastructure is increasingly reserved for entities that already generate substantial non-AI revenue or carry investment-grade ratings. Pure-play compute providers face higher financing costs and consolidation risk. On the demand side, most companies outside the technology sector have not yet seen AI translate into higher revenue per employee or wider margins. Prioritize use cases with clear, measurable operating impact; treat agentic and token-based spending as a governed cost category until evidence of return appears; and monitor credit conditions for suppliers of compute.
Job seekers and professionals. Demand remains strong for roles that embed AI into real business processes—forward-deployed engineering, implementation, and domain-specific application—particularly in markets where firms must demonstrate value beyond experimentation. At the same time, aggregate tech layoffs in 2026 have already approached 196,000 globally as companies reallocate resources toward AI infrastructure. Skills that improve output per employee and can be demonstrated with measurable results will continue to differentiate candidates. In emerging markets, local-language applications in agriculture, small-business finance, and primary healthcare are areas where new roles are being designed around the initiatives now being funded.
The capital being raised for GPUs is real and large. The operating returns outside a handful of technology companies are not yet visible at the same scale. That gap—between the financing available to investment-grade platforms and the measured results inside most enterprises, and between high-capacity regions and the rest of the world—will shape hiring, investment, and competitive positioning through 2027.
ArtificialIntelligence, AI, BusinessStrategy, JobMarket, TechNews, SpaceX, EnterpriseAI, CareerDevelopment, GlobalBusiness, ComputeInfrastructure, radeyaglobal
