Key areas to watch
AI infrastructure demand durability
Signals from hyperscalers (Meta, Microsoft, Alphabet, Amazon, etc.) will indicate multi-year capital expenditure (capex) commitments to GPU clusters versus any pivot to custom application-specific integrated circuits (ASICs) or internal accelerators. Breadth of demand updates from enterprises, sovereign AI projects, vertical workloads, and non-Big 5 customers will show whether the AI boom is broadening.
Product cycle and roadmap execution
Look for fresh details on the Blackwell ramp (production scale and contributions), early Blackwell Ultra performance, and power and efficiency wins. Rubin architecture visibility is crucial for the second half (H2) calendar year (CY) 2026 ramp, ensuring smooth transitions without order pauses, while preserving pricing power and mid-70s margins.
Supply, lead times, and capacity planning
Watch current lead times for flagship accelerators, foundry/packaging ramp progress (e.g., TSMC/CoWoS). Assess if supply constraints still cap revenue upside or if easing bottlenecks shift the story toward normalised (but strong) growth.
Regulatory and export-control impacts
Latest updates on US export restrictions (high-end products to certain regions, especially China) should be watched, noting that no meaningful data centre revenue is assumed.
Competitive landscape
Management may discuss rivals (Advanced Micro Devices, Intel, hyperscaler custom silicon) and how NVIDIA’s CUDA/software ecosystem continues to secure workloads. Framing competition as expanding the AI pie (more adoption overall) rather than share erosion will ease multi-year growth concerns.
Software, services, and recurring revenue
Expect updates on AI Enterprise/software monetisation, subscription-like streams, partnerships (e.g., OpenAI/Anthropic co-optimisation), and shifts toward a durable platform business for higher long-term margins and valuation.
Capital returns and balance sheet
Details on share repurchase updates (ongoing firepower as volatility buffer), dividend policy, and allocation priorities (R&D, capacity, potential mergers and acquisitions (M&A)) will be in focus.
Guidance for Q1 FY 2027
This could be the real mover; anything exceeding $75 billion – $76 billion (some bulls see higher) would signal sustained high growth.
Circular deals: fuelling growth or bubble risk?
One of the most debated aspects of NVIDIA’s recent surge is the role of circular AI deals – where cloud providers, model labs, and AI start-ups buy, sell, and rent GPU capacity to each other in a tight loop.
On the surface, this ecosystem appears self-reinforcing. Hyperscalers commit billions to NVIDIA GPUs, then create AI services sold to model builders, enterprises, and start-ups. These start-ups often raise capital to buy more GPU time, feeding demand back into the system. AI funding drives GPU orders, higher GPU orders validate high AI valuations, and those valuations in turn justify more funding.
Since Q3 FY 2026, however, the picture has become more nuanced. Higher funding costs and increased AI business model scrutiny have become a natural stress test, creating clearer separation between well-capitalised, scale players with tangible revenue prospects and smaller, speculative projects. NVIDIA has also highlighted a broader demand base: not just cloud giants and headline-grabbing start-ups, but also enterprises, sovereign AI initiatives, and industry-specific deployments in areas such as healthcare, automotive, and financial services.
The key question for Q4 FY 2026 is whether this demand mix is shifting from a capital markets-driven scramble for GPUs to a more durable, usage-driven infrastructure build-out. Evidence of multi-year contracts, diversified customer cohorts, and rising utilisation of deployed capacity would support the view that today’s elevated spending is funding real, long-lived AI workloads rather than a transient bubble. If growth still appears focused on aggressive AI players and resale-driven cloud offerings, concerns about circularity – and risk of a sharper correction if sentiment turns – will remain a concern for investors.
Is NVIDIA a buy or a sell?
NVIDIA has a TipRanks Smart Score of ‘7 neutral’ and is rated as a ‘strong buy’ by analysts with 37 ‘buy’, 1 ‘hold’, and 1 ‘sell’ recommendations as of 17 February 2026.