NVIDIA는 AI 인프라 수요가 2027 회계연도까지 지속되어 2028 회계연도로 이어질 것이라고 보고했다.

NVIDIA’s Q2 FY 2027 earnings demonstrated sustained demand for AI infrastructure, driven by data center expansion, Blackwell Ultra adoption, and Vera Rubin production. The results also highlighted extended demand visibility, anchored by firm supply commitments, partner deployments, and projected FY 2028 revenue growth.

NVIDIA (NASDAQ: NVDA) reported Q2 FY 2027 revenue of $96.22 billion, a 106% year-over-year increase that surpassed Wall Street consensus of $92.38 billion. Data Center revenue reached $89.02 billion, up 117% YoY and 18% sequentially. Within the Data Center segment, Hyperscale contributed $48.71 billion, while AI Clouds, Industrial, and Enterprise accounted for $40.31 billion. Edge Computing revenue totaled $7.20 billion, up 27% YoY. Adjusted operating income came in at $63.96 billion (+124% YoY), exceeding consensus estimates of $61.19 billion, with an adjusted operating margin of 66.5% compared to 61.1% in Q2 FY 2026. Adjusted earnings per share rose to $2.22 from $1.01 a year earlier.

“The AI infrastructure build-out is at full steam,” said Jensen Huang, founder, president, and chief executive officer of NVIDIA. “Vera Rubin, now in full production, was built to power exactly this moment.”

NVIDIA’s Q2 FY 2027 results confirm that AI infrastructure demand remains supply-constrained rather than demand-constrained. The quarter revolved around three interconnected themes: surging compute requirements, platform expansion beyond discrete GPUs, and innovative financing structures engineered to fund large-scale AI factory deployments. Vera Rubin’s transition into full production is particularly significant, as it provides customers and partners with a defined roadmap extending past Blackwell Ultra. Concurrently, NVIDIA is deepening its footprint in CPUs, networking, storage processing, and reference infrastructure, complicating traditional valuation models that treat the company solely as a semiconductor vendor.

Vera Rubin’s entry into mass production stands out as the quarter’s most critical strategic development. The platform is already operational across key partners including CoreWeave, Google Cloud, Microsoft Azure, Oracle Cloud Infrastructure, and Nebius. This diverse customer base underscores NVIDIA’s simultaneous pursuit of hyperscalers, neocloud providers, and specialized infrastructure firms. Integration of Spectrum-6 switch systems, Vera CPUs, and Vera BlueField-4 STX security capabilities further broadens the platform’s scope beyond pure accelerator demand. By delivering a unified AI factory architecture rather than fragmented compute, networking, and security components, Vera Rubin solidifies NVIDIA’s role as the reference architecture provider for the upcoming wave of AI infrastructure investment.

Blackwell Ultra continues to fuel hyperscale momentum, with hyperscale revenue more than doubling year-over-year and climbing 13% sequentially. This trajectory directly addresses investor concerns regarding potential slowdowns in enterprise AI capital expenditure. The favorable product mix from Blackwell Ultra also bolstered Q2 FY 2027 gross margins, even as memory expenses exert downward pressure. Several customers have received notifications of AI server price hikes exceeding 15%, primarily reflecting elevated memory costs. NVIDIA’s pricing leverage remains crucial, enabling the company to safeguard profitability while fulfilling urgent capacity requirements. Strong Blackwell Ultra adoption effectively neutralizes near-term anxieties about decelerating AI infrastructure spending.

To secure long-term supply, NVIDIA expanded its procurement commitments to $279 billion from $119 billion in the prior quarter, predominantly allocated to memory inventory. This scale demonstrates a deliberate strategy to lock in essential components for multi-year demand cycles. On the capital front, strategic financing alliances with Apollo, BlackRock, Blackstone, Brookfield, Goldman Sachs, and KKR target the mobilization of over $500 billion in third-party infrastructure capital, pending definitive agreements. Additionally, a partnership with SB Energy at the PORTS-Pike Technology Campus in Ohio secures land, power, and shell capacity for future compute deployment, with guarantee obligations capped at $105 billion and initial operations slated for FY 2029. Collectively, these initiatives address power, capital, memory, and deployment bottlenecks simultaneously, constructing the financial and logistical foundation required to translate AI factory demand into recognized revenue.

Looking ahead, NVIDIA guided Q3 FY 2027 revenue to $108 billion (±2%), comfortably above Wall Street’s $105.15 billion expectation. The forecast excludes Data Center compute revenue from China, offering a clearer view of organic demand in other regions. GAAP and non-GAAP gross margins are projected at 74% (±50 basis points), with non-GAAP operating expenses estimated at roughly $9 billion. NVIDIA anticipates gross margins contracting to 71%–72% in Q4 FY 2027 as elevated memory costs fully impact the cost structure, followed by a recovery to 72%–73% in FY 2028 as pricing mechanisms adjust. Full-year FY 2028 revenue growth is expected to approximate 70%.

The forward outlook positions NVIDIA to enter its next product cycle with robust demand visibility extending well beyond the Blackwell Ultra ramp. Vera Rubin manufacturing, $279 billion in procurement commitments, and expanded access to institutional infrastructure financing establish a durable platform for capacity scaling, even as rising memory costs introduce short-term margin compression. Excluding assumed Chinese Data Center compute revenue further elevates the importance of execution across hyperscalers, neoclouds, enterprise clients, and sovereign AI initiatives. Should NVIDIA successfully convert Vera Rubin output and secured supply chains into active deployments while leveraging pricing to counterbalance component inflation, AI infrastructure growth is likely to sustain elevated levels through FY 2028 despite the company’s expanding scale.

For detailed financial statements and supplementary disclosures, refer to the official NVIDIA Q2 FY 2027 earnings press release.