but Raises the Financing Question
$NVIDIA(NVDA)$’s fiscal-second-quarter results were extraordinary even by its own recent standards. The company reported on August 26 for the quarter ended July 26: revenue reached $96.22 billion, up 18% sequentially and 106% year over year; Data Center revenue rose 117% to $89.0 billion; and non-GAAP EPS increased 120% to $2.22. Non-GAAP gross margin held at 75.0%. Nvidia’s official release provides the results and product milestones.
The bullish thesis is that demand is broadening while the product cycle is accelerating. Vera Rubin is moving into full production across major cloud providers, Groq 3 LPX is in production, and Nvidia expects third-quarter revenue of $108 billion, plus or minus 2%, without assuming any China Data Center compute revenue. That guide implies another large sequential increase despite component constraints. Nvidia also returned approximately $26 billion through repurchases and dividends during the quarter and retained $99 billion under its repurchase authorisation.
The platform advantage is larger than the GPU alone. CUDA, networking, rack-scale systems, CPUs and software give customers a common architecture, raising switching costs. If inference, agentic AI and sovereign deployments expand simultaneously, Nvidia can grow even as hyperscalers develop custom accelerators because the total compute market is expanding faster than substitution.
The bearish question is who ultimately finances the demand. Large cloud companies and AI laboratories are committing enormous sums before end-customer revenue has proved equally durable. Nvidia is increasingly helping organise infrastructure capital, and some customers or counterparties depend on Nvidia-linked investment. This does not make reported revenue circular, but it increases the need to distinguish creditworthy demand from capacity built on optimistic utilisation assumptions.
Margins are the second warning. Management guided third-quarter gross margin to 74.0%, below the just-reported 75.0%, while memory and component costs remain elevated. Export controls leave China as upside optionality but also a continuing geopolitical risk. Customer concentration, custom silicon and the possibility of an AI-capex slowdown remain the principal threats.
NVDA Daily Chart
NVDA fell 1.6% in the August 26 regular session to $209.66 after trading from $209.23 to $213.60 on 179.9 million shares, about 28% above its 65-day average. It initially weakened after the release, reversed during the call and finished after-hours at approximately $219.53, up 4.7% from the regular close. By about 5:25 a.m. Eastern on August 27, it had accelerated to approximately $224.52, up 7.1%, on 5.4 million premarket shares.
That move has already cleared the article’s original $220 trigger, but a premarket print is not equivalent to a confirmed breakout. If the regular session accepts prices above $220, that zone should become first support, with $209–$214 as the gap-fill area and $200 as deeper support. Resistance is now approximately $230, followed by the 52-week high near $236.54. An opening spike that falls back below $220 would be a failed-breakout warning rather than confirmation.
Because the premarket gap has already reached roughly $225, the cleaner setup is to wait for regular-session confirmation rather than assume $220 has become support. If NVDA closes above $220 and later holds $215–$220 on a retest, an illustrative 30–45-day bull put spread with the short put near 0.10–0.15 live delta below the gap-fill zone and the long put $10 lower could express a constructive view with defined loss. Exact strikes should come from the live chain. A regular close below $209 alongside falling estimates invalidates the earnings-gap setup. Maximum loss is the spread width minus credit.
Personally, I sold some puts before the earnings. [Sly]
The evidence leans moderately bullish. Revenue, Data Center demand, product breadth and guidance all accelerated, but financing quality and the coming margin dip prevent maximum conviction. The view would be invalidated by $108 billion guidance being reduced, major customers cutting AI capex, gross margin falling materially below 74% or the shares losing $200 while estimates decline. This is personal opinion for education and is not financial advice; it is not an instruction to enter any trade.
Comments