AI demand remains strong, while financing and capex strain persist; China adds competition but no new shock.

No dashboard signal or composite changed. New evidence chiefly tests financing and infrastructure assumptions: NVIDIA disclosed large contingent support for an OpenAI-linked Ohio campus; Oracle and AI-cloud providers continue to tap debt, equity, customer prepayments and project finance; demand evidence remains firm in NVIDIA, HPE and contracted AI-cloud capacity. In China, Alibaba's cloud growth and profitability improved, while its equity raise and ByteDance's reported offshore loan underscore that competitive progress is also capital intensive. This is evidence, not a rescore.

• Capex and funding are increasingly linked to long-dated leases, guarantees, special-purpose vehicles and private/project credit, increasing sensitivity to utilisation, counterparty performance and refinancing.
• NVIDIA's disclosed $105 billion capped guarantee exposure for a 4.25GW OpenAI-linked Ohio site illustrates the scale of contingent commitments now accompanying AI infrastructure.
• Oracle reported $43 billion of fiscal-2026 debt financing and expects roughly $40 billion more debt and equity financing in fiscal 2027, directly testing whether AI-cloud expansion remains cash-funded.
• China's capable open-weight models and lower-cost supply could compress premium proprietary-model pricing and shift value from model providers toward infrastructure, applications or local ecosystems.

• NVIDIA reported Data Center revenue up 117% year over year to $89.0 billion and guided to $108.0 billion of total revenue, supporting the semiconductor-demand signal.
• HPE reported Cloud & AI revenue growth of 25.4%, Data Center Networking growth of 112.2%, higher free cash flow and improved margins, providing broader infrastructure-demand evidence.
• IREN and Hut 8 reported contracted capacity, customer prepayments and investment-grade-backed financing; these structures reduce near-term funding needs where counterparties perform.
• Alibaba reported 45% AI Cloud and Compute Services revenue growth and 133% adjusted EBITA growth, showing that Chinese AI-cloud demand is not solely speculative supply expansion.

• Whether AI-cloud revenue and free cash flow begin to catch up with capex, rather than only hardware and contracted-capacity indicators remaining strong.
• Terms, spreads, guarantees and customer concentration in new AI infrastructure financings, especially for non-investment-grade counterparties.
• Evidence that power, delivery or chip supply constraints defer commissioned capacity or weaken utilisation.
• Whether cheaper Chinese open-weight models increase enterprise adoption enough to expand inference demand, or instead chiefly compress model-provider economics.