US ▾
NVDA
Earnings call · Jul 2026 (Q2 FY27)

NVIDIA Q2 FY27 earnings call NVDA

Aug 26, 2026 Source

Executive summary

NVIDIA Q2 FY27 — Record Revenue and Strong AI Demand Drive Growth Amid Supply Constraints

NVIDIA delivered another quarter of record revenue, driven by surging AI demand across hyperscalers, neo clouds, and enterprises, despite significant supply constraints. The company is expanding its full-stack AI factory platform, including new VeraRubin and Vera CPU products, and is strategically investing in Frontier AI labs and regional neo clouds to capture a broader market. Gross margins are expected to temporarily decline due to memory pricing but are projected to recover in FY28.

Highlights

5

  • Total revenue of $96 billion, more than doubled year-over-year, accelerating for the fourth consecutive quarter.

  • Data Center revenue increased 18% quarter-over-quarter to $89 billion, with strong contributions from Hyperscale and ACIE segments.

  • ACIE revenue grew 25% sequentially and 138% year-over-year to $40 billion, driven by Neo cloud capacity additions.

  • Networking business achieved record revenue, growing 18% sequentially, with Spectrum-X Ethernet up 2.6x year-over-year.

  • Returned a record $26 billion to shareholders in Q2 FY27, comprising $20 billion in share repurchases and $6 billion in dividends.

Concerns

5

  • Expected fiscal year 2028 revenue growth of approximately 70% is supply-constrained, despite much higher demand.

  • GAAP and non-GAAP gross margins are expected to decline to 74% (plus or minus 50 basis points) in Q3 FY27 due to extreme memory pricing.

  • Gross margins are projected to bottom in Q4 FY27 at 71% to 72% before settling at 72% to 73% in fiscal year 2028.

  • Less than 1% of total data center revenue in Q2 FY27 came from China, with no China data center compute revenue in the forward outlook due to geopolitical uncertainty.

  • Inventory increased to $32 billion and Days of Sales Outstanding (DSO) rose to 60 days, reflecting preparation for VeraRubin launch and extended payment terms.

Guidance & targets

CategoryTargetConfidence
Total revenue
$108 billion, plus and minus 2%
high materiality
High
Fiscal Year 2028 revenue growth
approximately 70% year-over-year
high materiality
High
GAAP and non-GAAP gross margins
74% plus or minus 50 basis points
high materiality
High
GAAP and non-GAAP gross margins
71% to 72% range
high materiality
High
GAAP and non-GAAP gross margins
72% to 73%
high materiality
High
GAAP operating expenses
approximately $9.2 billion
medium materiality
High
Non-GAAP operating expenses
approximately $9.0 billion
medium materiality
High
Full-year FY27 OpEx growth
low 50s
medium materiality
High
Full-year FY27 GAAP and non-GAAP tax rate
between 16% and 18%
low materiality
High
CPU revenue growth
more than double
high materiality
High
VeraRubin share of Data Center revenue
about 20%
medium materiality
High

Segment performance

SegmentRevenueYoYQoQMargin
Data Center
Strong contributions from both Hyperscale and ACIE segments.
$89 billion—18%—
Data Center - Hyperscale
Driven by sustained strength in Blackwell, reinforcing that more compute drives more revenue.
$49 billion—13%—
Data Center - ACIE
Growth driven by Neo cloud capacity additions to meet rising demand from enterprises, AI start-ups, sovereigns, and hyperscalers.
$40 billion138%25%—
Networking
Achieved another record quarter, with Spectrum-X Ethernet helping NVIDIA become the largest and fastest-growing network company.
Spectrum-X Ethernet growth: 2.6x YoY
——18%—

Orderbook & backlog

Cloud industry backlog >$2 trillion Q2 FY27
AWS GPU deployment 2 million GPUs Q2 FY27

additional deployment

Starting this quarter through Q2 FY29. Includes Vera CPUs, some integrated with Ruben, other standalone.

OpenAI compute commitment 12 gigawatts Q2 FY27

Existing and planned commitments for NVIDIA AI infrastructure through 2030.

Frontier AI lab credit enhancement 2 gigawatts Q2 FY27

For compute, complementing substantial NVIDIA compute capacity secured independently.

Product announcements

ProductTypeDetails
VeraRubinlaunch
Vera CPUlaunch
Grok 3 LPXlaunch

Deals & partnerships

AWS Expansion of existing partnership for GPU deployment and AI stack adoption starting Q2 FY27 through Q2 FY29

AWS is deploying an additional 2 million GPUs, along with Vera CPUs. AWS will serve NVIDIA Nematron family of open models on Amazon Bedrock and SageMaker, and adopt Omniverse, Cosmos, Isaac, and Jason for warehouse robots.

Apollo, BlackRock, Blackstone, Brookfield, Goldman Sachs, KKR Partnerships to establish financing platforms for AI infrastructure over $500 billion

These partnerships will enable AI labs to build and assess AI infrastructure funded by long-term institutional capital at attractive rates.

SoftBank Energy Partnership to exclusively host NVIDIA compute at Portsmouth Campus

Secured land, power, and shell capacity at their Portsmouth Campus. The initial deployment will support 4.25 gigawatts of AI factory capacity for OpenAI.

Nora Japan Partnership to build an NVIDIA DSX AI factory

Will create open models to power AI agents, digital twins, robotics, and physical AI applications for Japan's national AI company.

LG and Hyundai Motor Group Partnering to build and scale AI

South Korean companies partnering with NVIDIA for AI initiatives.

Care, Mevius, NScale, Firebird, Pasaba technologies, GMI Cloud, Yoda, Anesa, Hermes, YTL AI cloud Partnerships with regional Neo Clouds to build AI infrastructure

These Neo Clouds are emerging globally, pairing local land, power, and operating expertise with NVIDIA's platform. NVIDIA's model provides a minimum revenue guarantee to give lenders confidence to underwrite projects.

Capital programs

Portsmouth Campus AI factory initial deployment

Benefit:4.25 gigawatts of AI factory capacity

Secured land, power, and shell capacity through partnership with SoftBank Energy to exclusively host NVIDIA compute. Initial deployment will be utilized by OpenAI. The site could support multiple upgrade cycles over 20 years.

Risks & headwinds

Supply constraints at least through the end of fiscal year '28

Expected FY28 revenue growth of 70% is supply-constrained, despite demand being much higher.

Mitigation:Working with the entire supply chain to increase capacity; greater visibility upstream and downstream for planning.

Extreme memory pricing Q3 FY27 through FY28

Q3 FY27 GAAP and non-GAAP gross margins expected to be 74% (±50 bps), bottoming at 71-72% in Q4 FY27, before settling at 72-73% in FY28.

Mitigation:Executed price increases taking effect in Q1 FY28; long-standing deep relationships with all 3 major memory suppliers to increase capacity.

Geopolitical uncertainty (China) Q2 FY27 and forward outlook

<1% of total data center revenue from China in Q2 FY27; no China data center compute revenue in forward outlook.

Mitigation:Operating in accordance with U.S. government licenses; Hopper shipments to China are dilutive to corporate gross margins.

Inventory increase and extended payment terms Q2 FY27

Inventory increased to $32 billion; Days of Sales Outstanding (DSO) increased to 60 days.

Mitigation:Inventory build is in preparation for VeraRubin launch; extended payment terms are for large purchases by certain investment-grade customers to be shipped over multiple quarters.

What to watch in Q3 FY27

VeraRubin contribution to Data Center revenue

Q4 FY27 and into FY28
Current about 20% (Q3 FY27 guide)
Target Increase in Q4 FY27 and beyond

Why it matters

VeraRubin is expected to be the fastest product ramp and a key driver of future growth, with its share indicating successful adoption and supply ramp.

We see verarubin accounting for about 20% of data center revenue in Q3. Looking ahead, our preliminary expectation is for fiscal year 2018 revenue to grow approximately 70% year-over-year, although we will work to close the supply-demand gap we expect supply to remain a bottleneck, at least through the end of fiscal year '28.

Q&A highlights

What gives confidence for the 70% FY28 growth guidance, what is the gap between demand and supply, and what are the key constraints?

Jensen Huang explained that demand is driven by agentic AI, the full-stack AI factory platform, and significant growth in non-hyperscaler segments (ACIE). He stated that demand is much greater than 70%, but supply allows for confident delivery of 70%. The entire supply chain is challenged, requiring extensive planning for land, power, and cooling, but NVIDIA has greater visibility upstream and downstream.

“even though our demand is much greater than 70%, our supply allows us to confidently deliver 70%.”

asked by Joseph Moore · answered by Jen-Hsun Huang

2 min read 6 chapters

Detailed narrative

AI Demand and Infrastructure Buildout

The surge in AI demand is driving a global infrastructure buildout, supported by a diverse set of customers including hyperscalers, AI labs, AI natives, enterprises, and sovereign entities. The cloud industry backlog now exceeds $2 trillion, with CapEx by the top 5 hyperscalers projected to reach nearly $800 billion in 2026 and $1.3 trillion in 2027. This demand is fueled by the increasing complexity and utility of AI agents, which require significantly more compute than human interaction.

Full-Stack AI Factory Platform Expansion

NVIDIA's revenue opportunity per gigawatt has significantly expanded across product generations, from approximately $18 billion with Hopper to $25 billion with Blackwell, and now to $40 billion with VeraRubin. The VeraRubin platform integrates VeraCPU, Ruben GPU, NVLink, InfiniBand, Ethernet, and GrokLPU, enabling extreme co-design for ex-factor performance gains. This full-stack approach allows customers to utilize the infrastructure across the entire AI lifecycle, from data creation to agentic inference.

Strategic Investments and Partnerships in AI Infrastructure

NVIDIA is making strategic investments in Frontier AI labs, committing nearly $50 billion, and partnering with leading capital providers (Apollo, BlackRock, Blackstone, Brookfield, Goldman Sachs, KKR) to establish financing platforms that will raise over $500 billion for AI infrastructure. A key partnership with SoftBank Energy will host NVIDIA compute at its Portsmouth Campus, supporting 4.25 gigawatts of AI factory capacity for OpenAI, which has committed to 12 gigawatts of NVIDIA compute through 2030.

Growth in Non-Hyperscaler Segments (ACIE)

The ACIE segment, encompassing sovereign AI, regional neo clouds, enterprise edge, and air gap data centers, is a major growth driver, increasing 35% sequentially and tripling year-over-year in Q2. This segment is expected to represent roughly half of NVIDIA's data center business. NVIDIA is facilitating the growth of neo clouds globally through partnerships and a revenue-sharing model, where NVIDIA provides a take-or-pay commitment on capacity in exchange for a share of rental revenue.

Memory Pricing and Gross Margin Outlook

Extreme memory pricing, driven by the AI build-out, is expected to impact gross margins. Q3 FY27 GAAP and non-GAAP gross margins are projected to be 74% (±50 bps), declining from 75% in Q2. Margins are anticipated to bottom in Q4 FY27 at 71% to 72% before recovering to 72% to 73% in fiscal year 2028, as executed price increases take effect in Q1 FY28. Management emphasized transparency regarding this temporary headwind.

CPU and Networking Business Expansion

NVIDIA's Grace CPU has achieved over $5 billion in trailing 12-month revenue. The next-generation Vera CPU is now in full production, with shipments underway to lead partners including AWS, and is expected to more than double CPU revenue in FY28. The networking business also reported a record quarter, with revenue growing 18% sequentially, and Spectrum-X Ethernet experiencing 2.6x year-over-year growth, contributing to NVIDIA's position as a leading network company.

AI-generated summary of the company's earnings call. Not investment advice.