After NVIDIA's Q2 Earnings: Bulls and Bears Clash Over Gross Margin and Financing Risks
nashnova research
Nvidia's Q2 revenue hit $96.22 billion, doubling year-over-year, with EPS beating estimates by 6% — but analysts are now split on falling gross margins and the company's customer-financing model, meaning the earnings beat itself is no longer the story; earnings *quality* is.
How strong was the quarter?
Revenue reached $96.22 billion, up 106% year-over-year; non-GAAP EPS came in at $2.22, topping the $2.09 Street consensus by roughly 6%.
Management guided to about 70% growth for fiscal 2028. This means → the company is betting AI compute demand keeps accelerating over the next two years, not slowing.
In plain terms = the scorecard isn't in dispute — the fight is over how long this pace can last.
Why do bulls see $425?
Nvidia trades at roughly 23.5× earnings, below the Nasdaq average of 28.4×; its PEG ratio — price-to-earnings divided by earnings growth, a gauge of how much you pay per unit of growth — is just 0.52. This means → relative to profit growth, the stock is arguably cheap.
Analyst Michael McGrath rates the stock a strong buy with a $425 target; even his conservative $300 scenario implies only 21.4× earnings.
Vera Rubin GPU — Nvidia's next-generation graphics-processor platform — entered volume ramp in the second half, with full-year free-cash-flow potential exceeding $150 billion. That is the core basis for bulls arguing the multiple should re-rate to 25×.
What three gaps are bears watching?
Falling gross margins: rising memory costs are squeezing profitability — Q3 margins are expected at 74%, dropping to 71–72% by Q4. In plain terms = the more Nvidia sells, the thinner each dollar of profit becomes.
Cash-flow vs. profit gap: Q2 free cash flow was $21.34 billion, while GAAP net income reached $59.69 billion — book profit nearly triple actual cash in hand. This reflects a real question about earnings quality.
Customer-financing risk: Nvidia sells chips to clients while also holding their equity and guaranteeing their debt. Analyst "Bears of Wall Street" rates the stock a sell, arguing "if demand were truly as strong as management claims, these arrangements would be unnecessary."
Ecosystem building — or backstopping your own customers?
Bull analyst Rob Tanner frames the financing as "ecosystem building," citing AI compute demand from Meta, Microsoft, and Amazon spreading well beyond frontier labs — evidence of downstream breadth.
Bears counter that CEO Jensen Huang recently pushed AI clients to secure an additional $500 billion in outside financing to keep them funded, while the company appears not to have fully vetted clients' ability to repay large contracts.
This means → the same action carries opposite readings — the decisive variable is not the financing itself but whether downstream AI customers eventually earn enough to service their debt.
Customer financing: ecosystem investment or systemic risk?
BULL
Ecosystem logic
AI compute demand from Meta, Microsoft, Amazon is spilling over — financing accelerates ecosystem growth.
Still undervalued
PEG at just 0.52; P/E below Nasdaq average — growth is not fully priced.
BEAR
Seller and creditor
Selling chips while guaranteeing client debt — unnecessary if demand is truly strong.
Substitution risk
OpenAI's in-house chip Jalapeño poses inference-compute competition to Blackwell.
Margin pressure
Q4 gross margin may fall to 71–72%, steadily eroding earnings quality.
In plain terms = bulls see growth that hasn't been priced yet; bears see growth whose cash may never come back — both sides have facts, and the verdict hinges on whether downstream AI customers actually earn a return over the next year.
Why are neutral analysts sitting this out?
Analyst Vladimir Dimitrov holds a neutral rating: current earnings multiples reduce downside risk but are not enough to justify an above-market return expectation.
Analyst Oliver Rodzianko is bullish on fundamentals but finds sentiment and technicals unfavorable, choosing to wait for a better entry.
Seeking Alpha's quant system rates Nvidia a strong buy — ranked 10th out of 4,269 stocks and 4th among 70 semiconductor names. This reflects a gap between pure-data models and human analysts on sentiment calls.
What to watch next?
Near term: whether Q3 gross margin holds at 74%, and whether margins stabilize after the Q4 trough as bulls expect.
Medium term: whether credit risk from customer-financing arrangements surfaces in financial disclosures — the key window for validating the bear thesis.
In plain terms = nobody doubts the growth; the real suspense is how much of the profit turns into cash and whether the money clients borrowed can actually be repaid.
市场有风险，内容仅供研究参考，不构成投资建议。