Image source: © 2025 Krish Capital Pty.Ltd

NVIDIA Corp. (NASDAQ: NVDA) remains one of the most closely watched equities in the U.S. market. As of the latest trade, NVDA shares trade at USD 167.02, down USD 4.75 (-0.03%) from the previous close. The session opened at USD 167.89, hit an intraday high of USD 171.48, and touched a low of USD 164.08, with an active trading volume of over 224 million shares.

Across major market data providers, 12-month analyst price targets generally cluster around the USD 210–215 range, implying mid-teens upside from current levels. The most bullish targets run toward USD 250, while the most cautious hover near USD 100, highlighting the wide dispersion of expectations and the uncertainty around AI infrastructure demand, competitive threats, and macro risks like power constraints.

Where Wall Street Stands

Sell-side consensus remains tilted bullish, though analysts acknowledge a wide band of possible outcomes. The median target sits near USD 210, while high estimates from firms such as Loop Capital approach USD 250, citing an ongoing “AI golden wave” that could extend well into 2026. More conservative voices hold back from revising targets upward until there is greater visibility on Blackwell supply, data center power availability, and hyperscaler capital expenditure trends.

In practice, these targets reflect assumptions about three major factors:

  • How quickly Blackwell ramps and converts backlog into revenue.
  • Whether hyperscalers divert more spending to custom accelerators such as TPUs or Trainium.
  • The trajectory of global AI capex into 2026 and beyond.

Results and Guidance

NVIDIA’s most recent earnings report, for Q2 FY2026, delivered USD 46.7 billion in revenue, up 6% sequentially and 56% year over year. Data Center revenue stood out at USD 41.1 billion, also up 56% year over year. Management guided Q3 revenue to approximately USD 54 billion ±2%, suggesting another step-up in growth as Blackwell ramps further into production.

Gross margins remain at exceptional levels—over 73% on a non-GAAP basis—thanks to high-value AI accelerators and network products. NVIDIA also executed a 10-for-1 stock split earlier in the year, alongside a small but symbolic dividend increase, which has made shares more accessible for retail investors and potentially boosted liquidity.

Product and Roadmap

The Blackwell architecture (B200/GB200) is the backbone of NVIDIA’s near-term growth story. Independent industry trackers estimate that Blackwell will represent more than 80% of NVIDIA’s high-end GPU shipments by 2025 as GB200-based HGX systems and rack-scale deployments ramp across hyperscalers.

Beyond Blackwell, NVIDIA has unveiled its Rubin architecture, with systems expected to enter volume in the second half of 2026. Rubin will advance NVLink topologies and large-scale NVL systems, strengthening NVIDIA’s full-stack platform positioning. Timely delivery and customer adoption of Rubin will be critical for sustaining growth beyond the Blackwell cycle.

Competitive Dynamics

Competition remains NVIDIA’s biggest risk factor and potential price mover. AMD has publicly detailed its MI350 and MI355X accelerators, with the latter featuring dense rack configurations and HBM3e memory intended to close the performance-per-dollar gap with NVIDIA in training and inference workloads.

Hyperscalers are also stepping up their in-house silicon efforts. AWS Trainium2 instances claim 30–40% better price-performance than certain GPU fleets for appropriate workloads, while Google’s TPU v5p is being scaled for production across its data centers. Reports suggest Broadcom has secured a USD 10 billion ASIC order from a major customer, widely believed to be OpenAI, underscoring a growing interest in custom accelerators for specific AI tasks.

Despite this, NVIDIA’s CUDA software ecosystem, NVLink interconnect, and networking solutions create significant switching costs. The company remains deeply entrenched in the training market, though investors must monitor whether incremental AI capex continues to flow primarily toward NVIDIA or starts to fragment more meaningfully.

AI Capex Backdrop

The global AI infrastructure build-out remains the most significant external driver of NVIDIA’s top line. Multiple trackers peg 2025 hyperscaler capex around USD 300 billion or more, with debate centering on whether 2026 maintains a similar slope or begins to flatten. Sustained capex growth supports NVIDIA’s bull case, while any deceleration would likely compress valuation multiples, even if revenue continues to grow.

Policy and Export Controls

NVIDIA’s revenue mix has become less reliant on China, but export restrictions still introduce volatility. Earlier in 2025, the company guided to an USD 8 billion headwind from lost China sales of H20-class products, yet still posted guidance above Street expectations, suggesting resilient demand elsewhere. Investors should price in occasional air pockets tied to policy changes when modeling downside cases.

Base, Bull, and Bear Scenarios

Base Case:
Blackwell ramps successfully across hyperscalers and sovereign AI projects, with supply gradually improving and margins holding in the 73% range. Q3 revenue lands near the guided USD 54 billion mark, and buybacks continue to absorb dilution. Under these conditions, NVDA tracks consensus estimates around USD 210–215 over the next 12 months, consistent with the “central tendency” of analyst targets.

Bull Case:
Blackwell demand exceeds available supply through mid-2026, hyperscaler capex plans are revised higher, and Rubin previews provide strong visibility into the next cycle. Under these conditions, NVDA could rally toward USD 230–250, near the Street’s high end, with potential overshoots on strong quarterly beats or macro tailwinds.

Bear Case:
Rapid adoption of custom silicon by one or more megacap customers, combined with power-grid delays or stricter export controls, causes guide-downs and margin compression. If AI capex growth decelerates meaningfully in 2026, NVDA could retest the USD 120–160 range, which currently represents the low end of sell-side targets.

Multi-Year Contour

Looking out to 2027, NVIDIA’s shipments will be dominated by Blackwell through 2026, with Rubin scaling into high-volume deployments by late 2026 and 2027. Networking (InfiniBand, Spectrum-X Ethernet) and NVLink continue to deepen NVIDIA’s integration at the rack level, reinforcing its transition from a component supplier to a full-stack platform provider.

AMD’s MI400 series and more mature custom accelerators could start to capture meaningful share by this time, particularly in inference workloads. Investors will need to assign probabilities to these competitive shifts when modeling NVIDIA’s longer-term earnings power.

Key Variables to Watch

  • AI Capex Trajectory: Will 2026 sustain the hot pace of 2025, or roll over?
  • Blackwell Supply Chain: Are HBM3e and advanced packaging capacities scaling in line with NVIDIA’s needs?
  • Custom Silicon Adoption: How much workload share do AWS, Google, and others shift toward their own silicon?
  • Policy Developments: Will export restrictions tighten or relax, impacting near-term revenue?
  • Roadmap Execution: Is Rubin on time and well-received, sustaining visibility into the next architectural cycle?

Bottom Line

NVDA’s price forecast over the next 12 months depends heavily on the pace and durability of the AI data center build-out. The Q2 results and Q3 guide suggest continued acceleration, supporting the Street’s current consensus in the low USD 200s, with room for upside if demand and supply stay aligned. Conversely, any significant shift toward custom silicon or a pause in AI capex could test the lower end of current valuation ranges.