Nvidia's $96 Billion Quarter and a 70% Growth Forecast That Stunned Wall Street

For most companies, a quarter with revenue north of $90 billion would be a career-defining event. For Nvidia, the three months ending July 2026 were merely the latest step in the steepest revenue ramp in the history of the technology industry — and the guidance that came with it was the real news. On August 26, 2026, Nvidia reported second-quarter fiscal 2027 revenue of $96.2 billion, up 106% year over year, and told Wall Street that next fiscal year’s revenue should grow by roughly 70%, a figure that blasted past even the most optimistic consensus estimates.

The numbers behind the beat

The headline results were strong across the board. Revenue of $96.22 billion rose 18% sequentially and more than doubled from the $46.7 billion a year earlier, beating the analyst consensus. Adjusted earnings came in at $2.22 per share against an expected $2.10, according to LSEG data compiled by CNBC. Data Center revenue — the engine of the entire business — reached $89.0 billion, up 117% year over year, meaning the AI infrastructure business now represents better than 92% of the company’s total sales. Gross margin held around 75%, a level most chipmakers can only dream about, even as Nvidia absorbs the costs of ramping its newest system generations.

Guidance was just as forceful. Nvidia forecast third-quarter revenue of $108.0 billion, plus or minus 2%, comfortably above the average analyst estimate of $104.19 billion. That would represent another double-digit sequential jump and keep the company on a trajectory that analysts say could carry it past the $396 billion consensus for full fiscal 2027 — which already ends in January 2027 — toward something closer to $673 billion for fiscal 2028 if the 70% growth outlook holds.

The 70% bombshell

The single most consequential line of the earnings call was not any of the quarterly figures but the preliminary outlook for fiscal 2028. Wall Street had been modeling roughly $570 billion in fiscal 2028 revenue, about 44% growth from fiscal 2027 — itself a heroic assumption for a company already this size. Nvidia’s chief financial officer told investors to expect approximately 70% growth instead, and management noted that demand would have been even stronger were it not for persistent supply constraints. In other words, the ceiling on Nvidia’s growth right now is not customer appetite; it is how fast the supply chain can ship chips, memory, and server racks.

On the call, Chief Executive Jensen Huang compressed the entire bull case into four words: “compute is revenue.” The argument is that in an agentic-AI economy, companies that secure compute capacity can directly convert it into revenue-producing products — coding agents, customer-service automation, physical AI and robotics workloads — which in turn makes AI infrastructure a productive investment rather than speculative capex. Huang has been building toward this framing for quarters, but presenting it against a backdrop of 106% growth gave it a weight that skeptics could not easily dismiss.

Wall Street’s reaction: an $8.7% surge and a new pecking order

Markets needed less than a day to render their verdict. On August 27, Nvidia shares jumped 8.7% — their steepest one-day gain since April 2025 — adding roughly $442 billion in market value in a single session, according to WSJ live-market coverage. The rally lifted the entire semiconductor complex and the Nasdaq with it, as investors once again embraced the AI trade that has powered the US market for the past three years. By the close, Nvidia stood as the most valuable public company, its market capitalization around $3.3 trillion and climbing, and CNBC analysis suggested the 70% growth path would put it on track to become the tech industry’s No. 2 company by revenue, behind only Apple.

It is worth remembering how unusual this pattern is. Nvidia has beaten earnings estimates five straight quarters, yet the stock had actually declined on earnings day four times in a row going into this report, with investors habitually asking whether the law of large numbers had finally caught up. The 70% fiscal 2028 forecast answered that question — for now — by shifting the debate from whether growth is decelerating to how long a supply-constrained company can keep compounding at these rates.

The AWS deal: 2 million more GPUs for the agentic era

Running alongside the earnings news was a major partnership expansion that underscores where Nvidia sees the next wave of demand. Amazon Web Services and Nvidia announced plans to deploy an additional 2 million Nvidia GPUs — spanning Blackwell Ultra, Rubin, and Rubin Ultra generations — across AWS’s global infrastructure in 2027 and 2028. The capacity is explicitly aimed at what the companies call agentic AI and physical AI workloads, and it comes bundled with deeper collaboration on CPUs, networking, and robotics platforms, plus plans for US government “AI factories” with federal workloads in mind.

The timing is deliberate. Hyperscaler capital expenditure has been the subject of intense scrutiny throughout 2026, with bears arguing that cloud providers cannot keep spending at these levels without proven returns. A multi-year, multi-million-GPU commitment from AWS — the largest cloud operator — functions as a public vote of confidence in exactly the thesis Huang was selling on the call: that leading models, and increasingly AI agents and robotics systems, require compute budgets measured in gigawatts, not server racks.

What to watch next

Three things will determine whether the 70% forecast holds. First, gross margin trajectory: Q3 guidance implies margins stabilizing in the mid-70s, and any slippage as Rubin ramps would pressure the narrative. Second, supply: management repeatedly framed growth as supply-limited, so incremental capacity from packaging, memory, and system integration partners effectively translates into revenue. Third, the financing environment: Nvidia itself has joined the AI borrowing wave, raising $25 billion in high-grade bonds in June 2026 — its first such offering since 2021 — and working with Wall Street asset managers on vehicles that could mobilize more than $500 billion for AI infrastructure by treating it as an investable asset class.

The larger lesson of this quarter extends beyond one company. Nvidia has now demonstrated that AI infrastructure demand can keep accelerating even at a run-rate approaching $400 billion a year, and that the world’s largest cloud providers are willing to lock in multi-year capacity commitments to secure it. The debate over an “AI bubble” is not over — debt-funded buildouts and circular vendor-financing arrangements still draw scrutiny — but as of August 2026, the burden of proof has shifted decisively to the skeptics. As Huang put it on the call, compute is revenue, and for now the market believes him.