Nvidia's Hidden $40 Billion Business Beyond Big Tech Is Growing Twice as Fast

Nvidia reported $89.0B of Data Center revenue in fiscal Q2, with direct hyperscaler sales up 13% sequentially to $49.0B. The newer ACIE bucket reached $40.0B (+25% sequential, +138% YoY) and is expected to drive most of next quarter’s Data Center growth before hyperscale re-accelerates as Vera Rubin supply improves. However, diversification may carry financing risk: days sales outstanding rose to 60 from 45 as extended payment terms suggest buyers are increasingly dependent on outside capital, with guarantee obligations capped at $108.5B and plans to mobilize $500B+ of third-party capital.

Analysis

The key shift is not demand breadth; it is who is financing demand. When buyers move from hyperscaler capex to capital-markets-dependent counterparties, NVDA starts to behave more like a credit-intermediated business: higher receivables, more counterparty support, and more sensitivity to funding conditions than the headline growth rate implies. Near term that can extend the multiple because the story looks less concentrated, but the quality of incremental revenue is more cyclical than the market is likely pricing.

That creates a cleaner winner set among BLK, BX, KKR, BN, and GS than among chip names. These firms can capture origination, placement, and asset-management fees tied to AI infrastructure while outsourcing much of the hardware risk; the second-order issue is that they may become the toll booth for a crowded, capital-hungry trade. If private credit spreads widen or one of the financed buyers stumbles, the same channels that accelerate order growth can shut quickly and freeze follow-on demand.

Catalyst window is 1-3 months around the next financing update, receivable trends, and any spread wobble in AI-linked private credit. The thesis is falsified if DSO stops worsening or guarantee exposure begins to shrink relative to revenue; over 6-18 months, this looks less like de-risking and more like a financed demand cycle that can compress NVDA’s premium multiple if funding costs rise faster than AI ROI.

Key Decisions for Investors

• Relative-value idea: long a basket of BN/BX/KKR/BLK (and GS as the syndication/placement leg) vs. short NVDA on a 1-3 month horizon; the spread works if AI buildout remains capital-market-funded and NVDA’s working-capital burden keeps rising. Cut the short if NVDA’s next update shows DSO stabilizing or guarantee disclosures flattening.
• For existing NVDA exposure, add a 3-6 month put spread or collar into strength rather than chasing upside here; the risk is a financing hiccup that hits the multiple before it hits revenue. Falsifier: another quarter of growth with no further DSO deterioration.
• Overweight BLK/BX/KKR/BN on pullbacks as the cleaner monetization of AI infrastructure capital formation; these names can compound fee income without taking the same inventory and receivable risk as NVDA. Reduce if AI credit spreads widen materially or if deal pipelines shift from funded to delayed.
• Watch item, not a trade: any move in NVDA receivables or guarantee obligations above the recent trend line is an early warning that 'diversification' is actually leverage to external financing conditions.