NVDA
NVIDIA Corporation
Dominant GPU and accelerated-computing platform for AI training/inference, gaming, and auto.
$115.08
-0.19 (-0.16%) session
Demo quote — not a live market feed
Market cap
$2.9T
P/E
48.5
Div yield
0.03%
Volume
7.7M
Price history
+8.49% over 1Y
Illustrative demo series
At a glance
Primary R&D focus
Horizon score
Strong forward setup
Unmatched platform position, but valuation and concentration risk demand perfect execution.
Forward brief
What matters next — not a recap of last quarter
NVIDIA’s forward question is not ‘did they beat last quarter’ — they usually do — it is whether Data Center demand compounds through the next architecture cycle while software/systems deepen the moat. R&D is almost entirely core platform defense. Next earnings: sequential Data Center, gross margin, and any China/export language. Those lines decide if the future is still accelerating or entering a digestion phase.
Next earnings
Q2 FY2027 · Aug 27, 2026
Expected EPS
$0.92
Expected revenue
$44.8B
Report date
Aug 27, 2026
Market prices perfection; any data-center sequential miss or China export update will reprice the multiple.
What to watch on the call
- Data Center revenue sequential growth and backlog commentary
- Blackwell / next architecture ramp vs supply constraints
- Gross margin — is networking mix holding margins elevated?
- China export compliance and alternative SKU contribution
Questions that actually matter
- Is customer capex still accelerating or are hyperscalers digesting prior buys?
- How durable is CUDA moat if custom ASICs (Google, Amazon, Microsoft) take share of inference?
- When does networking (NVLink/Spectrum) become a first-class earnings story?
R&D deep dive
Not just “they invest in R&D” — in what, for when, and is it core?
Annual R&D / tech spend
~$10B+
Spend trend
Rising
% of revenue
~10%
R&D is laser-focused on staying the default AI compute platform — silicon, systems, and CUDA software. Automotive/Omniverse is adjacent diversification. Almost none of the spend is ‘science for science’s sake’; it is defense of an extraordinary franchise under competitive and export pressure.
Next-gen GPU architectures (Blackwell successors)
Higher FLOPS/watt, larger HBM packages, improved transformer engines for training and inference.
Strategic bet: Stay 12–18 months ahead of AMD and custom silicon on training economics.
Full-stack systems (DGX, GB200 NVL racks, networking)
Selling entire AI factories — GPUs + NVLink + Spectrum-X + software — not discrete chips only.
Strategic bet: Raise switching costs; make ‘NVIDIA rack’ the unit of purchase.
CUDA / software & inference frameworks
CUDA ecosystem, TensorRT, NIM microservices, enterprise AI software layers.
Strategic bet: Software stickiness is the real moat if silicon competitors close the gap.
Automotive / robotics / Omniverse simulation
Drive platforms, Isaac robotics, industrial digital twins.
Strategic bet: Diversify beyond hyperscaler GPU cycles into physical AI.
Strategic bets
Where management is placing multi-year chips
AI factory as product
Shift from chip vendor to systems vendor so customers buy multi-year platform roadmaps.
Horizon · 1–3 years
Software monetization of CUDA install base
Increase software attach so margins and stickiness survive silicon competition.
Horizon · 2–4 years
Upcoming catalysts
Aug 27, 2026
Q2 earnings
Data Center print + next-arch supply color
Next GPU architecture keynotes / sample timelines
Multiples re-rate on roadmap confidence
Forward risks
Customer concentration & capex digestion
A handful of hyperscalers drive the majority of Data Center; pauses show up immediately.
Export controls
China SKU restrictions can erase a material revenue stream with little warning.
Custom ASIC competition
Inference workload migration to in-house chips could slow GPU TAM growth longer-term.