CITIC Securities: US AI Sector Remains Resilient Against High Rates, but Financing Shock Phase Has Arrived
Alina Collins
CITIC Securities argues US AI stocks remain resilient under high rates, but the pressure has shifted from valuation to financing — hyperscalers are burning cash near the limit of their own cash flow, and the cost of borrowing to keep expanding is now being priced in.
Why hasn't high interest hurt AI stocks?
CITIC Securities notes the AI rally was born in a high-rate environment — asset performance and real-economy investment data since 2023 both show sustained resilience.
This means → the market is trading the "infrastructure build-out of a tech revolution" thesis: high expectations for long-term returns outweigh the drag from higher discount rates.
The firm draws a parallel to the dot-com era: from 1999 to early 2000, the Nasdaq kept climbing even as rates rose; high rates suppressed other sectors, funneling capital into "new economy" assets and creating a scarcity premium.
In plain terms = when AI is the only visible growth story in the market, money crowds in — high rates actually eliminate the competition.
What does the "two-phase shock" framework mean?
CITIC proposes a two-phase model of rate impact: Phase 1 is a valuation shock (higher rates compress the present value of future cash flows); Phase 2 is a financing shock (rates directly raise the cost of borrowing to expand).
The firm judges we have entered the early stage of Phase 2 — AI infrastructure expansion is shifting from internal cash-flow funding to external financing dependence.
This means → interest rates are no longer just a variable in a valuation formula; they have become a real-world constraint — "how much interest do we pay each month."
What three signals confirm financing pressure has arrived?
Signal 1: Hyperscalers' (the mega cloud companies — Microsoft, Google, Amazon) capital expenditure-to-operating-cash-flow ratio is approaching 100%. In plain terms = they are spending nearly everything they earn; their own money is running out.
Signal 2: Recent media reports show hyperscalers have begun raising external financing, confirming an internal funding gap.
Signal 3: After the June 5 non-farm payrolls data triggered a short-term sell-off, the market's focus shifted from "how high are rates" to "can AI capex still find cheap money under high rates."
How far are we from dot-com-level danger?
CITIC ran a quantitative comparison against the dot-com peak: cash-flow pressure stands at 46% of the 2000 peak, leverage at 14%, financing-cost pressure at 68%, cumulative debt scale at just 7%, and market sentiment toward AI capex at 50%.
This means → the financing shock has begun, but overall pressure is still at a significant distance from "bubble-burst" territory; the macro rate impact can still be absorbed.
In plain terms = we are at the "starting to borrow" stage, not the "can't repay" stage — but the direction has changed, and it needs watching.
What risks should investors watch next?
CITIC flags three risks: US inflation and rates rising beyond expectations, hyperscaler external financing costs spiking, and AI companies failing to deliver on earnings.
This reflects the firm's core verification point: whether AI capex can continue to secure external financing will determine if this rally extends — or replays the dot-com path.
In plain terms = the key question now is not "does AI have a future" but "where does the expansion money come from, how much does it cost, and how long can it last."
Content is for reference only, not financial advice.