A record Data Center print and a dividend that jumped from a penny to a quarter met a Wall Street target that is not the only cautionary datapoint — the same SEC filing that fed the rally also shows two customers controlling 39% of revenue.
A record revenue print, a near-doubled Data Center business, and a dividend hike from a penny to a quarter — Nvidia's Q1 FY27 earnings release delivered the kind of numbers that usually end the conversation. It didn't. The same window has produced a Bank of America price target reported at $350 and a resurfaced concentration cut — "the 54% problem" — that the bulls have not been able to wish away. The print is real. The question is whether it closes the gap.
For the quarter ended April 26, 2026, Nvidia reported $81.6 billion in revenue, up 20% sequentially and 85% year over year, per the company's Q1 FY27 release. Data Center revenue reached $75.2 billion, up 21% quarter over quarter and 92% year over year. GAAP gross margin held at 74.9%; non-GAAP at 75.0%. GAAP EPS was $2.39; non-GAAP $1.87.
Management guided Q2 FY27 revenue to $91.0 billion, plus or minus 2%, and explicitly excluded any Data Center compute revenue from China — a caveat that matters for any model that pencils China back in. The capital-return math is also concrete: $20 billion returned to shareholders in the quarter, an $80 billion buyback authorized on May 18, 2026, and a quarterly dividend raised from $0.01 to $0.25, payable June 26 to holders of record June 4.
That is the company line. The Wall Street counter-narrative is the reason the print has not closed the multiple.
The $350 figure traces to a Bank of America price-target hike reported via Yahoo Finance / 24/7 Wall St. — a target that, in the source-article framing, is supposed to sit comfortably above the stock. The "54% problem" originates in the same SEC-backed concentration discussion that Investing.com's "Nvidia Trap" analysis uses to argue the print cannot be read in isolation. The precise 54% line could not be independently verified in this research turn and is reported here as the source-title framing rather than as a confirmed concentration cut; what the SEC Form 10-Q for the period ended April 26, 2026, summarized by StockTitan, does confirm is the two-customer concentration at the heart of the bear case.
Per Investing.com's reading of the SEC filing, two customers — most likely Microsoft and Amazon — accounted for 39% of total revenue (roughly 23% and 16%). Both, the analysis notes, are simultaneously designing competing AI silicon. If half of those customers' AI workloads shift to in-house accelerators over a three-year horizon, the piece estimates a revenue-at-risk shift north of $10 billion. Deep Research Global's Nvidia SWOT reaches the same risk from a different direction: customer concentration, TSMC manufacturing dependency, China restrictions, and the custom-silicon threat all sit at the top of the risk register, alongside the AI capex sustainability question.
There is a third leg to the bear case, and it has nothing to do with GPUs. UncoverAlpha's "The Forgotten Chip" argues that agentic AI is shifting the compute mix back to CPUs: a Georgia Tech and Intel paper cited in the piece finds that tool processing accounts for 50–90% of total latency in agentic workloads, and an Intel VP quoted in the analysis argues that multi-agent orchestration runs on CPUs. CNBC reports Nvidia has forecast a $200 billion CPU data-center market that includes China — a number that reframes the long-run upside story beyond accelerators, and that pairs uncomfortably with the guidance line that strips China Data Center compute out of the next quarter.
None of this negates the print. But it is what the multiple is being asked to digest at the same time: a 92% Data Center growth rate, a 39% two-customer concentration, a $200 billion CPU TAM that management itself has called out, and a $350 Wall Street target that the market is choosing not to chase. The constructive read is that this is a calibration event — for holders weighing valuation, for buyers pricing in the capex cycle, and for AI builders deciding how much of their roadmap to put on one supplier. What the next quarter has to show is not another record. It is whether the gap closes.