Influential analyst Dan Ives believes the market is significantly underestimating NVIDIA's earnings power, particularly its exposure to physical AI applications and the massive Blackwell GPU rollout happening across major cloud providers. As NVIDIA prepares to report second-quarter fiscal results on August 26, 2026, Ives argues that consensus analyst models sit 15 to 20 percent below what the company will actually deliver over the next 12 to 18 months.

The timing of Ives' call matters because prediction markets are already pricing in optimism. Traders on Polymarket assign a 96.6 percent probability that NVIDIA beats consensus earnings per share expectations next week, and an 84.5 percent probability of a new all-time high by year-end. With 48 analysts rating the stock a Buy, 10 Strong Buy, and only one Sell, the question becomes what Ives actually sees that the broader market has missed.

What Is Physical AI, and Why Does It Matter for NVIDIA?

Physical AI refers to artificial intelligence systems deployed in robots, autonomous vehicles, and other machines operating in the real world, rather than in data centers. Ives emphasized this as the piece of NVIDIA's business that receives the least credit from investors. "NVIDIA right now, you're almost giving them minimal credit for physical AI. And what's really going to be the next evolution of AI," he stated.

Ives

The market has reason to overlook this segment. NVIDIA disclosed that physical AI revenue exceeded $9 billion over the last 12 months, which is small relative to the company's $75 billion in data center revenue last quarter. However, the strategic argument is compelling: as AI inference moves out of centralized data centers and into machines operating in the world, the addressable market expands dramatically. This could extend NVIDIA's growth story well beyond the current hyperscaler capital expenditure cycle.

The weakness in this thesis is timing. Robotics and autonomous systems have historically taken longer to commercialize than advocates projected, making this more of a long-term duration bet than a near-term earnings catalyst.

What Must NVIDIA Deliver in Wednesday's Earnings Report?

For Ives' underestimation thesis to hold, NVIDIA's earnings call needs to demonstrate several concrete developments. The company must provide a Q3 revenue guide that clearly steps up on Blackwell and Rubin GPU shipments, confirm that Vera Rubin production shipments begin in the third quarter, and offer enough color on sovereign and ACIE (Artificial Intelligence Compute Infrastructure Expansion) demand to support the $1 trillion in Blackwell and Rubin revenue visibility through calendar 2027 that management outlined in the Q1 release.

Evidence of massive Blackwell adoption is already visible. Colette Kress, NVIDIA's Chief Financial Officer, told analysts that Amazon Web Services (AWS) plans to deploy more than 1 million Blackwell and Rubin GPUs starting this year, and Microsoft's Fairwater data center is live ahead of schedule, powered by hundreds of thousands of Blackwell GPUs. This scale of deployment across multiple hyperscalers suggests demand far exceeds what consensus models may have anticipated.

How to Evaluate NVIDIA's Earnings Guidance and Growth Claims

• Data Center Revenue Trajectory: Watch for year-over-year growth rates in data center revenue. Last quarter showed 92 percent year-over-year growth with sovereign revenue up more than 80 percent, deployed across nearly 40 countries. Any deceleration or acceleration from this pace will signal whether hyperscaler capex is accelerating or stabilizing.
• Blackwell and Rubin Production Confirmation: Management must provide specific shipment timelines and production ramp details for Vera Rubin, the next-generation GPU. Delays or production constraints could indicate supply chain challenges that analysts haven't fully priced in.
• Physical AI Revenue Disclosure: Look for updated physical AI revenue figures and forward guidance. If this segment is growing faster than the $9 billion annual run rate, it validates Ives' thesis that the market is underweighting this emerging opportunity.
• China Revenue Restart Potential: Last quarter's Q2 guide of $91 billion explicitly excluded China data center compute revenue. Any restart of H200 GPU shipments to China would land as pure upside to current quarter results, not as validation of a broader re-rating.

Ives' skepticism runs counter to the prevailing market sentiment, which is already quite bullish. The stock closed at $216.85, up 23.79 percent over the past year and 16.41 percent year to date. Analyst consensus targets average $302.83, suggesting roughly 40 percent upside from current levels. For a call about underestimation to resonate, it must identify something concrete that the sell-side has genuinely missed, not merely project that analyst conservatism will persist.

The strongest version of Ives' argument centers on the scale and speed of hyperscaler Blackwell adoption. AWS and Microsoft's deployment plans, combined with sovereign AI initiatives across 40 countries, suggest the addressable market for NVIDIA's GPUs is expanding faster than traditional models anticipated. The weaker version relies on physical AI, which requires patience and carries execution risk.

What remains clear is that NVIDIA's earnings report will provide concrete data on whether the company's guidance and production ramps support the $1 trillion Blackwell revenue visibility through 2027 that management has staked its reputation on. If that number holds, Ives' thesis gains credibility. If it doesn't, the market's current pricing may already reflect the company's true earning power.