Nvidia just reported Q2 fiscal 2027 revenue of $96.2 billion, a 106% jump from the year-ago period, and guided Q3 to $108 billion. Jensen Huang then flew to Scotland, stood beside King Charles III, and told reporters he expects Nvidia to ship twice as many chips next year as it does this year. The market’s response: a forward earnings multiple that has been cut in half from its 2025 level, now sitting under 17 times projected profits. The Philadelphia Semiconductor Index, by comparison, carries a multiple of roughly 20 times forward earnings. The world’s dominant AI chip franchise is trading at a discount to its own sector.

That is the anomaly active traders need to resolve before positioning in either direction.

Why the Multiple Compressed

TCW senior portfolio manager Eli Horton put the skepticism plainly, telling Bloomberg that the stock has “de-rated pretty significantly,” suggesting the market expects earnings power to prove less durable than consensus models imply. The concern is sequenced around margins. Gross margin is expected to bottom in Q4 fiscal 2027 at 71% to 72%, CFO Colette Kress said on the August 26 earnings call, with memory prices the primary culprit. Nvidia guided Q3 gross margin at 74%, plus or minus 50 basis points. The margin trough lands before the volume doubling does. Shipping twice as many chips at 71% gross margin is a different financial profile than shipping at 75%, even if the top-line number is extraordinary.

Nvidia is also one of the weaker performers in the SOX this year. Names like Micron, AMD, and Intel have outpaced the company that built the AI infrastructure cycle everyone is still talking about.

The Bull Case Rests on Vera Rubin Economics

Huang’s volume pledge is supply-constrained, not demand-constrained. The CFO framed the roughly 70% fiscal 2028 revenue growth guidance explicitly as a supply-limited view, with actual customer demand implying something closer to doubling. That separation matters. Non-hyperscaler customers, including sovereigns, NeoClouds, and enterprises, already represent roughly half of Nvidia’s data center business.

The platform economics underpin the bull case. Grace Blackwell carries a roughly $25 billion revenue opportunity per gigawatt of compute deployed. Vera Rubin, which began production shipments in fiscal Q3 and commenced initial revenue shipments in August, steps that figure up to about $40 billion per gigawatt. Nvidia has also pointed to Vera Rubin delivering 30 times higher throughput per megawatt and 35 times lower token cost relative to Grace Blackwell Ultra. At under 17 times forward earnings with roughly 70% revenue growth guided for fiscal 2028, a company growing at that rate priced at that multiple is, by any classical screen, an industrial cyclical valuation on a secular growth asset.

The Bear Case Has Structural Legs

The competing argument is that Broadcom’s trajectory tells you exactly what hyperscalers plan to do with their next dollar of AI capex. Broadcom’s AI semiconductor revenue hit $16.7 billion in its most recent quarter, up 221% year-over-year, with Q4 guidance of $21.7 billion, up 236%. CEO Hock Tan has also described a path to more than $100 billion in AI semiconductor revenue in fiscal 2027. CUDA-dependent GPU infrastructure is not disappearing, but the trajectory of capital flowing toward custom ASICs is no longer theoretical. Google’s TPU momentum is contractually committed multiple generations forward.

The revenue growth China exclusion compounds the uncertainty. Nvidia has said it is assuming no data center compute revenue from China in its outlook, removing a segment that once represented meaningful upside.

Technical Framework

NVDA has pulled back roughly 6% from its May 2026 all-time high of $235.74. Volume on the post-earnings surge showed institutional participation, but the stock has struggled to hold moves above its 50-day moving average decisively. Key levels to monitor: the August post-earnings consolidation base near $210 as primary support, with the all-time high at $235.74 as the first meaningful resistance. A sustained hold above the 20-week moving average on rising volume would shift the near-term structure from recovery to resumption.

Scenario Modeling

Bull Case: Vera Rubin shipments ramp faster than feared through Q4 fiscal 2027, gross margin stabilizes above 73% rather than troughing to 71%, and non-hyperscaler demand remains the incremental swing factor. NVDA recaptures the $235 all-time high and extends toward analyst targets in the $276 to $350 range.

Base Case: Margin pressure lands as guided, volume growth meets the roughly 70% revenue forecast for fiscal 2028, and the stock re-rates modestly toward 20 times forward earnings as earnings revisions continue upward. Price range: $230 to $250 over a 12-month horizon.

Bear Case: Hyperscaler ASIC displacement accelerates, China revenue remains structurally excluded from the outlook, and gross margin fails to recover from the trough. At 14 to 15 times forward earnings on downward revisions, the stock revisits the $185 to $195 range.

Stocks to Watch

NVDA: The core positioning question. Supply-constrained roughly 70% revenue growth guidance at under 17 times forward earnings creates asymmetric risk-reward if margins stabilize above trough.

AVGO: The ASIC displacement trade. Broadcom’s client list includes Alphabet and Meta. With $100 billion-plus in AI semiconductor revenue discussed for fiscal 2027, it is the clearest structural alternative.

TSM: TSMC manufactures Nvidia’s Vera Rubin chips and sits at the bottleneck of every scenario. Hyperscaler capex projected at nearly $800 billion in 2026 and $1.3 trillion in 2027 flows through its fabs regardless of who wins the GPU-versus-ASIC debate.

AMD: The second-mover positioned to capture share if Nvidia’s pricing power softens during the margin trough.

ARM: Participates in Nvidia’s Vera CPU architecture, hyperscaler processor designs, and competing silicon simultaneously. The one name that wins share regardless of which platform architecture dominates the next cycle.

Strategic Conclusion

The Nvidia valuation compression is not noise. It reflects a genuine institutional debate about earnings durability, margin trajectory, and the pace of ASIC displacement. What the multiple does not reflect, at least not obviously, is a 106% year-over-year revenue quarter, a $40 billion per gigawatt platform entering production shipments, and a volume commitment from a CEO who now has, as he told Scotland reporters, significantly more supply chain visibility than a year ago. Disciplined traders focus on what changes the scenario, not what confirms the consensus. Right now, the Q4 fiscal 2027 gross margin print and the cadence of Vera Rubin shipments are the two data points that will determine which scenario is live.