Nvidia Beats, Stock Dumps—BofA Says Buy the Dip

Nvidia just delivered the biggest revenue quarter in its history. The stock fell anyway. That's become a pattern—the chipmaker has declined after three of its last four earnings calls, even as the numbers keep getting bigger.

Bank of America isn't fazed. Lead analyst Vivek Arya and his team reiterated their buy rating yesterday, named Nvidia a top pick, and raised their price target from $320 to $350—implying 56.6% upside from the current price of $223.47.

The investment note’s top line read: "Beat/raise speaks volumes, ignore noise, buy top pick."

Before unpacking why, a quick decoder for readers who don't follow Wall Street jargon daily. A "beat" means a company earned more than analysts predicted. A "raise" means its guidance—its own forecast for next quarter—also came in above expectations. When both happen at once, it's typically very, very, good news. The fact that the stock corrected is the "noise" Bank of America is telling investors to ignore.

The engine behind all of it: data centers—the giant warehouses of servers that power AI models, cloud computing, and pretty much everything else on the internet.

Earnings per share—the profit attributable to each share of stock—came in at $1.87 on an adjusted basis, beating the $1.73 analysts expected. Gross margin, the slice of revenue left over after production costs, held at 75%. Free cash flow for the quarter—meaning the actual cash the company generated after all expenses—reached $48.6 billion. Jensen Huang's comment on the earnings call: "Agentic AI has arrived, doing productive work, generating real value and scaling rapidly."

Of course, agentic AI is the miracle everyone on Wall Street is talking about right now.

The core argument isn't about one quarter. It's about the size of the market Nvidia is selling into, and how fast that market is growing.

BofA previously estimated the total AI market at $1.7 trillion. They now expect it to grow 4x, reaching over $3 trillion by 2030. Within that, they model Nvidia holding roughly 78% of the AI accelerator market—the chips specifically built to run AI workloads. That's basically a near-monopoly in the fastest-growing technology market BofA says it has ever tracked.

There's also a newer opportunity the bank recently upgraded. Nvidia is moving into agentic CPU chips: processors designed for AI agents, the kind of software that can autonomously complete complex tasks without human input. BofA raised its estimate for that market from $125 billion to $200 billion, and says Nvidia already has $20 billion in demand locked in for the second half of this fiscal year.

BofA lists six formal risks. Two deserve attention from anyone holding or considering Nvidia stock.

The first is the stock's own gravitational pull. Nvidia now represents 8.3% of the entire S&P 500 index—the benchmark that tracks America's 500 largest public companies. About 78% of active fund managers already own it. When that many people already hold a stock, there's simply a smaller pool of potential new buyers to push the price higher.

There's also a recurring criticism that Nvidia's investments in companies like OpenAI and Anthropic—as a vendor selling them chips—amount to circular spending.

BofA raised its earnings-per-share estimates by 9% for fiscal 2027 (to $9.09) and 15% for fiscal 2028 (to $13.27). Put another way: Nvidia earned $4.55 per share last fiscal year. BofA expects that to roughly double to $9.09 this year, then hit $13.27 the year after. Earnings per share growing at 43% annually is rare for any company, let alone one already worth $5.5 trillion.

At its current price, Nvidia trades at 19.7 times its estimated 2027 earnings. A metric that adjusts the P/E for growth rate, where lower is better—sits at 0.5x against a Mag-7 average of 3.9x.

Free cash flow is projected to go from $96.7 billion last fiscal year to $186.8 billion in 2027 and $282 billion in 2028. The company also raised its quarterly dividend 25-fold, from $0.01 to $0.25 per share, and announced an $80 billion additional share buyback authorization—bringing total repurchase capacity to roughly $120 billion.

The $350 price target is based on 26x estimated 2027 earnings, within Nvidia's historical range of 25x to 56x. The next concrete date on the calendar: CEO Jensen Huang's keynote at Computex on June 1, where BofA expects him to lay out Nvidia's agentic AI roadmap and CPU strategy in detail.