Hội nghị công bố kết quả kinh doanh Q3 FY2027 của NVIDIA (NVDA): Doanh thu đạt 116 tỷ USD, nâng triển vọng FY2028
NVIDIA báo cáo tổng doanh thu quý 3 năm tài chính 2027 đạt 116 tỷ USD, tăng hơn gấp đôi so với cùng kỳ năm trước. Doanh thu mảng Trung tâm Dữ liệu đạt 108 tỷ USD. Dòng sản phẩm Vera Rubin ghi nhận doanh thu gần 20 tỷ USD trong quý đầu thương mại. Công ty nâng triển vọng tăng trưởng doanh thu năm tài chính 2028 lên khoảng 75%. Colette Kress giữ chức CEO tạm quyền thay Jensen Huang. NVIDIA hoàn trả kỷ lục 28 tỷ USD cho cổ đông và tăng hạn mức mua lại cổ phiếu thêm 150 tỷ USD.
Điểm nhấn chính
- NVIDIA báo cáo doanh thu kỷ lục quý 3 năm tài chính 2027 đạt 116 tỷ USD, cao hơn gấp đôi so với cùng kỳ năm trước và tăng 20% so với quý trước. Doanh thu mảng Trung tâm Dữ liệu tăng 21% so với quý trước lên 108 tỷ USD.
- Dòng sản phẩm Vera Rubin đã tạo ra gần 20 tỷ USD ngay trong quý đầu tiên xuất xưởng thương mại, trở thành sản phẩm tăng tốc sản lượng nhanh nhất của NVIDIA. Ban lãnh đạo cho biết nhu cầu tiếp tục vượt quá nguồn cung.
- NVIDIA đã nâng triển vọng tăng trưởng doanh thu năm tài chính 2028 lên khoảng 75% so với cùng kỳ năm trước, từ mức 70%, nhờ nhu cầu mạnh mẽ hơn và công suất nguồn cung được bổ sung.
- Doanh thu quý 4 năm tài chính 2027 dự kiến đạt 130 tỷ USD, cộng hoặc trừ 3%. Quý này sẽ có thêm một tuần, trong khi doanh số sản phẩm Trung tâm Dữ liệu Hopper được cấp phép sang Trung Quốc không được đưa vào dự báo.
- Colette Kress được bổ nhiệm làm CEO tạm quyền sau khi Jensen Huang từ chức Chủ tịch, Tổng Giám đốc (CEO) và thành viên Hội đồng Quản trị. Hội đồng Quản trị đã chỉ định Heidrick & Struggles dẫn dắt quá trình tìm kiếm CEO chính thức.
- NVIDIA đã hoàn trả mức kỷ lục 28 tỷ USD cho các cổ đông trong quý 3 và tăng hạn mức ủy quyền mua lại cổ phiếu thêm 150 tỷ USD, nâng số tiền còn lại sẵn có vào cuối quý lên 227 tỷ USD.
Dữ liệu tài chính quan trọng
| Chỉ số | Kết quả quý 3 năm tài chính 2027 | Mức thay đổi hoặc bối cảnh |
|---|---|---|
| Tổng doanh thu | 116 tỷ USD | Tăng hơn gấp đôi so với cùng kỳ năm trước; tăng 20% so với quý trước |
| Doanh thu mảng Trung tâm Dữ liệu | 108 tỷ USD | Tăng 21% so với quý trước |
| Doanh thu từ nhóm khách hàng Hyperscale | 60 tỷ USD | Tăng 23% so với quý trước |
| Doanh thu mảng Điện toán đám mây AI, Công nghiệp & Doanh nghiệp | 47 tỷ USD | Tăng 18% so với quý trước; gần gấp ba lần so với cùng kỳ năm trước |
| Doanh thu mảng Điện toán biên | 8,2 tỷ USD | Tăng 14% so với quý trước và 42% so với cùng kỳ năm trước |
| Doanh thu dòng Vera Rubin | Gần 20 tỷ USD | Quý đầu tiên xuất xưởng thương mại |
| Doanh thu mảng Thiết bị mạng | Hơn 21 tỷ USD | Tiếp tục mở rộng hệ sinh thái sản phẩm chủ lực |
| Biên lợi nhuận gộp GAAP | 74,8% | Cao hơn mức dự báo nhờ cơ cấu sản phẩm |
| Biên lợi nhuận gộp Non-GAAP | 74,9% | Cao hơn mức dự báo nhờ cơ cấu sản phẩm |
| Hoàn trả vốn cho cổ đông | 28 tỷ USD | 22 tỷ USD từ mua lại cổ phiếu và 6 tỷ USD từ cổ tức |
Kết quả kinh doanh và hoạt động
Dòng sản phẩm Vera Rubin cùng nhu cầu duy trì mạnh mẽ đối với Grace Blackwell là động lực thúc đẩy tăng trưởng mảng Trung tâm Dữ liệu. NVIDIA cho biết các hệ thống Rubin đã được đưa vào vận hành tại mọi nhà cung cấp dịch vụ quy mô siêu lớn (hyperscaler) lớn và các neocloud hàng đầu chỉ trong vòng ba tháng kể từ đợt giao hàng đầu tiên. Các đối tác hyperscale đã đưa Rubin vào vận hành trong khoảng một tháng, bằng một nửa thời gian cần thiết so với Grace Blackwell.
Ban lãnh đạo ghi nhận nhu cầu trên diện rộng từ các hyperscaler, các phòng thí nghiệm AI tiên phong, các công ty thuần AI, các doanh nghiệp và chính phủ. Các phòng thí nghiệm tiên phong lớn dự kiến sẽ đóng góp khoảng 15% doanh thu mảng Trung tâm Dữ liệu trong năm tài chính 2027. Khối lượng đơn hàng tồn đọng của ngành điện toán đám mây đạt 2,5 nghìn tỷ USD, so với mức 2 nghìn tỷ USD của ba tháng trước đó.
NVIDIA cho biết hạ tầng tính toán của hãng vẫn hoạt động hết công suất tại các hyperscaler và neocloud. Giá token giảm đang thúc đẩy mức độ sử dụng cao hơn, trong khi giá thuê GPU đã tăng lên. Các GPU Hopper đã được ký lại hợp đồng nhiều năm với mức giá lên tới gấp đôi giá gốc, và giá thuê B300 đã tăng hơn 60% kể từ khi ra mắt.
Doanh thu từ AI chủ quyền tăng hơn gấp ba lần so với cùng kỳ năm trước. Doanh thu tại chỗ của doanh nghiệp từ các dịch vụ tài chính đạt gần 5 tỷ USD tính theo 12 tháng gần nhất và tăng hơn gấp đôi, nhờ sự hỗ trợ từ hạ tầng giao dịch định lượng. Ban lãnh đạo dự kiến doanh thu từ khối doanh nghiệp sẽ tăng hơn gấp đôi trong năm tài chính 2028.
Doanh thu từ AI vật lý tăng gần gấp ba lần so với cùng kỳ năm trước và đang trên đà vượt 25 tỷ USD trong năm tài chính 2027. NVIDIA cũng dự kiến tổng doanh thu CPU sẽ tăng hơn gấp đôi trong năm tài chính 2028 khi việc ứng dụng Vera ngày càng mở rộng.
Công ty đã ký thỏa thuận chính thức vào tháng 9 để mua lại Hugging Face, nền tảng hiện phục vụ hơn 18 triệu nhà phát triển và 200.000 doanh nghiệp. NVIDIA định vị thương vụ này là một phần trong chiến lược mở rộng hệ sinh thái mô hình mở của mình.
Dự báo của Ban lãnh đạo
NVIDIA dự kiến doanh thu quý 4 năm tài chính 2027 đạt 130 tỷ USD, cộng hoặc trừ 3%, nhờ nhu cầu hạ tầng AI và đà tăng tốc sản xuất liên tục của dòng Vera Rubin. Triển vọng này bao gồm một tuần bổ sung nhưng loại trừ doanh số bán các sản phẩm Trung tâm Dữ liệu Hopper được cấp phép sang Trung Quốc.
Biên lợi nhuận gộp GAAP và non-GAAP dự kiến đạt trong khoảng 71% đến 72% trong quý 4. Ban lãnh đạo kỳ vọng quý 4 sẽ là mức đáy của biên lợi nhuận gộp, sau đó sẽ cải thiện lên 72%–73% trong năm tài chính 2028.
Chi phí hoạt động GAAP quý 4 dự kiến vào khoảng 11 tỷ USD, với chi phí hoạt động non-GAAP đạt khoảng 10,7 tỷ USD. Các ước tính này bao gồm tuần bổ sung và các chi phí liên quan đến quá trình chuyển giao lãnh đạo.
Đối với năm tài chính 2028, ban lãnh đạo đã nâng triển vọng tăng trưởng doanh thu từ 70% lên khoảng 75% so với cùng kỳ năm trước, tính từ mức nền cơ sở cao hơn của năm tài chính 2027. Dự báo phản ánh sự mở rộng nguồn cung kỳ vọng, mức độ sẵn sàng của trung tâm dữ liệu, nhu cầu của khách hàng cuối và khả năng tiếp cận vốn của khách hàng.
NVIDIA tiếp tục dự kiến thuế suất GAAP và non-GAAP cho năm tài chính 2027 ở mức 16%–18%, loại trừ các khoản mục phát sinh riêng lẻ và những thay đổi trọng yếu trong môi trường thuế.
Rủi ro và các yếu tố cần theo dõi
Quá trình chuyển giao ban lãnh đạo là vấn đề quản trị trọng tâm. Hội đồng Quản trị cho biết Jensen Huang đã từ chức sau khi có thông tin về hành vi cá nhân không phù hợp với quy tắc ứng xử, các tiêu chuẩn và chuẩn mực văn hóa của NVIDIA. Hội đồng Quản trị không có ý định phục chức cho ông và đang đánh giá các ứng viên nội bộ cũng như bên ngoài cho vị trí CEO chính thức.
NVIDIA cũng thành lập Ủy ban Giám sát Kiểm soát Xuất khẩu và An ninh Quốc gia. Ban lãnh đạo cho biết họ vẫn đang đánh giá các biện pháp kiểm soát xuất khẩu mới của Bộ Thương mại Mỹ. Doanh số các sản phẩm Trung tâm Dữ liệu Hopper được cấp phép bán sang Trung Quốc trong quý 3 đạt tổng cộng 2,4 tỷ USD, nhưng NVIDIA hiện không còn xuất xưởng các sản phẩm đó sang Trung Quốc và đã loại trừ chúng khỏi dự báo.
Nguồn cung vẫn thấp hơn nhu cầu. Các cam kết mua hàng của NVIDIA đã tăng lên 390 tỷ USD, trong khi các cam kết thuê trung tâm dữ liệu và dịch vụ đám mây đã vượt mức 100 tỷ USD. Các cam kết liên quan đến neocloud đạt tổng cộng 32 tỷ USD, và tổng giá trị bảo lãnh lên tới 118 tỷ USD, bao gồm 105 tỷ USD hỗ trợ tín dụng cho khu phức hợp PORTS-Pike.
Nguồn tài trợ của khách hàng, tiến độ xây dựng, khả năng tiếp cận vốn cùng sự sẵn có của đất đai, nguồn điện và mặt bằng thô vẫn là những biến số triển khai quan trọng. Ban lãnh đạo nhấn mạnh rằng một số khoản bảo lãnh nhất định chỉ bắt đầu có hiệu lực sau khi các đối tác đảm bảo được nguồn tài trợ và hoàn thành các cột mốc hạ tầng theo yêu cầu.
Điểm nhấn phiên hỏi đáp với chuyên gia phân tích
Ban lãnh đạo cho rằng triển vọng tăng trưởng cao hơn trong năm tài chính 2028 là nhờ nhu cầu tiếp tục mở rộng và nguồn cung được bổ sung. Ban lãnh đạo bày tỏ sự tự tin vào mục tiêu 75%, đồng thời thừa nhận rằng tác động từ các biện pháp kiểm soát xuất khẩu mới vẫn cần được xem xét thêm.
Về vị thế cạnh tranh, bà Kress nhấn mạnh vào đội ngũ nhân viên, lộ trình công nghệ, mối quan hệ với khách hàng cũng như quyền tiếp cận được đảm bảo đối với bộ nhớ và công nghệ đóng gói tiên tiến của NVIDIA. Hội đồng Quản trị cho biết CEO chính thức phải là người có năng lực định hướng cả về công nghệ bán dẫn tiên tiến lẫn các yếu tố an ninh quốc gia.
Về biên lợi nhuận, ban lãnh đạo duy trì triển vọng biên lợi nhuận gộp cho năm tài chính 2028 ở mức 72%–73%. Ban lãnh đạo cho biết quý 4 phản ánh các đơn đặt hàng hiện có của khách hàng và chi phí cung ứng cao hơn, trong khi chính sách giá mới sẽ bắt đầu có hiệu lực khi NVIDIA bước sang năm tài chính tiếp theo.
NVIDIA cho biết việc phân bổ công suất phụ thuộc chủ yếu vào việc khách hàng đã chuẩn bị sẵn sàng trung tâm dữ liệu, có đủ nguồn vốn và cam kết bao tiêu đáng tin cậy hay chưa. Ban lãnh đạo khẳng định việc phân bổ không dựa trên mức giá phân biệt và các khách hàng đều nhận được mức giá như nhau.
Các ưu tiên phân bổ vốn bao gồm hoàn trả vốn cho cổ đông, hỗ trợ công suất của nhà cung cấp và tài trợ có chọn lọc cho khách hàng hoặc cơ sở hạ tầng. NVIDIA dự kiến sẽ hoàn thành hạn mức mua lại cổ phiếu hiện tại vào cuối năm tài chính 2028, đồng thời tiếp tục đánh giá riêng rẽ từng thỏa thuận bảo lãnh và tài trợ.
Toàn văn biên bản cuộc họp công bố kết quả kinh doanh
Toàn văn cuộc gọi công bố kết quả kinh doanh
Phần trình bày của ban lãnh đạo
Operator
Good afternoon. My name is Sarah, and I will be your conference operator today. At this time, I would like to welcome everyone to NVIDIA's Third Quarter Financial Results Conference Call. [Operator Instructions] Thank you.
Toshiya Hari, you may begin your conference.
Toshiya Hari
Thank you. Good afternoon, and welcome to NVIDIA's conference call for the third quarter of fiscal 2027. With me today from NVIDIA are Colette Kress, Interim Chief Executive Officer and Executive Vice President and Chief Financial Officer; and Harvey Jones, Chairman of the Board.
Our call is being webcast live on NVIDIA's Investor Relations website. The webcast will be available for replay until the conference call to discuss our financial results for the fourth quarter of fiscal 2027. The content of today's call is NVIDIA's property. It can't be reproduced or transcribed without our prior written consent.
During this call, we may make forward-looking statements based on current expectations. These are subject to a number of significant risks and uncertainties, and our actual results may differ materially. For a discussion of factors that could affect our future financial results and business, please refer to the disclosure in today's earnings release, our most recent Forms 10-K and 10-Q and the reports that we may file on Form 8-K with the Securities and Exchange Commission.
All our statements are made as of today, November 17, 2026, based on information currently available to us. Except as required by law, we assume no obligation to update any such statements. During this call, we will discuss non-GAAP financial measures. You can find a reconciliation of these non-GAAP financial measures to GAAP financial measures in our CFO commentary, which is posted on our website.
With that, let me turn the call over to Harvey for some opening remarks.
Harvey Jones
Thank you, Toshiya. Good afternoon, everybody, and thank you for joining us today. As Chairman of NVIDIA's Board of Directors, I want to say a few words on the Board's behalf. NVIDIA is an exceptionally strong company. Its strength comes from its extraordinary people, groundbreaking technology and exemplary track record of execution. Our financial performance speaks for itself.
On November 12, our Board of Directors announced that after disclosure of personal behavior inconsistent with the company's code of conduct, standards and cultural norms, Jensen Huang voluntarily resigned his positions as President, CEO and Board member. The NVIDIA Board unanimously accepted his resignation and does not intend to reinstate Jensen as CEO. The Board appointed Colette Kress, NVIDIA's longtime and well-respected CFO, as Interim CEO. In addition, the Board announced it was initiating a search for a permanent Chief Executive, and we have since retained Heidrick & Struggles to lead the search.
While this was an unexpected transition, the Board acted deliberately and in the best interest of the company's employees, customers, partners and shareholders. This morning, the company announced that our Board has established a National Security and Export Control Oversight Committee that will include a select group of our independent Board Directors, senior external advisers and legal counsel. This committee will enhance our engagement with the U.S. government, including all relevant agencies. We will not be addressing or taking questions on today's call regarding the circumstances of Jensen's resignation and the ongoing investigation. Our Board will communicate further when appropriate.
On behalf of our shareholders, the Board is proactively dealing with the immediate consequences facing the company as well as enhancing our oversight capacity going forward. We have an incredibly talented leadership team, and we have every confidence in Colette's leadership, integrity and ability to deliver as interim CEO. Colette has been at the center of this company's extraordinary growth for the last 13 years, instrumental in executing strategy and with a deep understanding of our business, customers, operations and financial position. The Board and the entire leadership team remain laser-focused on driving NVIDIA forward. Our commitment to generating value for shareholders, customers and partners remains unchanged.
With that, I will turn it over to Colette.
Colette Kress
Thank you, Harvey. Our momentum is real. With the market-leading multiyear road map, supply agreements extending through to the end of the decade and sustainable demand, we are well positioned to accelerate NVIDIA's leadership position. Our ambition will not change. Our passion to innovate will not change. And our commitment to customers, partners and shareholders will not change.
Now let me transition to our financials. We delivered another outstanding quarter, marked by record revenue, operating income and EPS. Total revenue of $116 billion, more than doubled year-over-year and grew 20% on a sequential basis. Vera Rubin is off to an extraordinary start. In its first quarter of production shipments, the Rubin platform generated nearly $20 billion in revenue, making it the fastest product ramp in NVIDIA's history.
We're addressing strong demand for compute across a diverse and expansive set of end customers, including hyperscalers, frontier labs, AI natives, enterprises and governments. This, combined with increased support from our supply base, we are raising our fiscal year 2028 outlook. We expect revenue to grow approximately 75% year-over-year, up from 70% and off what is now a higher fiscal 2027 base. Our outlook considers data center readiness, the strength and durability of end customer demand and our customers' access to capital.
Q3 Data Center revenue increased 21% quarter-over-quarter to $108 billion with balanced growth across Hyperscale and AI Clouds, Industrial & Enterprise subsegments. Hyperscale revenue of $60 billion grew 23% sequentially, driven by the Vera Rubin ramp and sustained strength in Grace Blackwell. Our Hyperscale customers' cloud revenue growth accelerated for the sixth consecutive quarter. Cloud industry backlog now sits at $2.5 trillion, up from $2 trillion just 3 months ago, underscoring the significant multiyear opportunity ahead of our partners and of NVIDIA.
Proprietary models from OpenAI and Anthropic continue to advance the frontier with OpenAI launching its latest GPT-6 models and Anthropic releasing its Claude 5.5 models. As frontier models become more capable, adoption and revenue are scaling rapidly with Anthropic's annualized revenue reaching $81 billion at the end of October. We expect the major frontier labs to contribute approximately 15% of our Data Center revenue this year.
Open models are accelerating. AI adoption and weekly token volume on OpenRouter has surged 35x year-over-year and 3.5x in just this past 3 months. Agentic workloads now account for approximately 80% of tokens. A rapidly expanding ecosystem of inference providers is bringing these models to companies everywhere. At NVIDIA, we are committed to fostering a vibrant open model ecosystem so every country, company and industry can build its own intelligence.
In September, we entered into a definitive agreement to acquire Hugging Face, the home of open model community, serving more than 18 million developers and 200,000 companies. Together, we will scale its platform and make intelligence more open, more capable and more accessible. Nemotron is now the most widely used U.S. open model on OpenRouter and is where developers start to build their own intelligence. That foundation is empowering the ecosystem. This past quarter, leading enterprise software and security companies, including Microsoft, Salesforce, Palantir, CrowdStrike, Cadence and Synopsys, announced products that use post-trained open models tailored to domain-specific use cases.
Training partners such as Scale, Mercor, Prime Intellect and Applied Compute are helping enterprises turn open models into their own using their data, tools and workflows and start-ups across specialized domains, including Lila Sciences, Harvey, Therna and Abridge, are using post-trained open models to deliver results in science, law and health care. All of this is creating a powerful flywheel. Token prices have fallen nearly 60% since May, accelerating token usage and demand for compute. As a result, GPU rental prices are rising across generations. Hopper GPUs have recontracted at up to twice their original price under the multiyear agreements, while the B300 rates have risen more than 60% since their introduction.
NVIDIA compute remains fully utilized across every hyperscaler and neocloud we serve, underscoring the productivity, durability and fungibility of our infrastructure. As agents run longer and act with greater autonomy, organizations need stronger controls to ensure safety and security. The NVIDIA Open Agent Safety Platform, which we announced alongside more than 100 partners, enables these controls. OpenShell contains agents by limiting their access to files and tools, while NVIDIA Sentry running on BlueField-4 monitors the behavior in silicon and stops agents from overstepping limits in milliseconds.
OpenShell is already gaining broad ecosystem support from partners across AI models, enterprise software, cybersecurity, cloud infrastructure and robotics. Companies working with NVIDIA include Anthropic, AWS, CoreWeave, CrowdStrike, Dell, Google Cloud, HPE, IREN, JPMorgan Chase, Lenovo, Microsoft, Nebius, OpenAI and SpaceXSI. Advances in agents harnesses are accelerating the adoption of agentic AI. Meta's Muse agent launched on NVIDIA's infrastructure surpassed 10 million setups in the United States and Canada within the first 6 weeks. To support this massive scale and future growth, Meta extends to expand its deployment of Vera Rubin hardware with the goal of maximizing performance and efficiency for its next generation of AI agents.
Agentic AI is driving extraordinary demand for CPUs, far outstripping supply. NVIDIA Vera CPU meets the growing demand, delivering leading performance at data center processing and scientific simulation workloads. CoreWeave, Meta, Nscale and OpenAI have now adopted Vera, joining AWS, OCI and SpaceXSI, the leading partners we announced last quarter. We continue to see very strong demand for Vera and expect total CPU revenue to more than double in fiscal year '28. AI Clouds, Industrial & Enterprise or ACI&E revenue of $47 billion increased 18% sequentially and nearly tripled year-over-year. Sequential growth was driven by Vera Rubin shipments to top leading neocloud customers and the start of our frontier labs building their own data centers.
Neoclouds remain an increasingly important part of the global AI infrastructure ecosystem, providing additional customers with access to NVIDIA compute and expanding the reach of our platform. We have supported the development of this ecosystem, and we will continue to be disciplined and selective in our support. NVIDIA's full stack AI technology platform uniquely positions us to address opportunities across sovereign, enterprise and physical AI. Sovereign AI revenue more than tripled year-over-year in Q3. In September, we announced a collaboration with multiple NVIDIA cloud partners in Australia to develop up a 2 gigawatts of land, power and shell capacity by the end of calendar '27 with an opportunity to host multiple generations of NVIDIA DSX AI factories.
In enterprise, financial services delivered the strongest growth with on-prem revenue reaching nearly $5 billion on a trailing 12-month basis and more than doubled year-over-year as more customers leverage NVIDIA infrastructure to accelerate quantitative trading. We expect enterprise revenue to be more than double in fiscal year '28. Physical AI revenue nearly tripled year-over-year in Q3 and is now on track to exceed $25 billion for the full year. At GTC Berlin, we expanded our NVIDIA Isaac robotics platform with a new humanoid reference design built on Isaac GR00T and Jetson Thor.
In automotive, NVIDIA Hyperion and Alpamayo are enabling Mercedes-Benz and Bolt with Lucid to build robotaxi fleets, Einride to develop Level 4 autonomous trucks and Volvo Cars to advance safe, automated driving. Q3 Edge Computing revenue of $8.2 billion increased 14% sequentially and 42% year-over-year. Sequential growth was driven by stronger enterprise demand for Blackwell workstations and edge compute for physical AI applications. Q3 sales of licensed Hopper Data Center products to China was $2.4 billion. Our guidance does not have any licensed Hopper Data Center products included, and we are no longer shipping licensed Hopper Data Center products to China.
Transitioning to an update on our supply chain, financial commitments and guarantees. Expanding our global supply chain to support customers' growth remains a top priority. With demand continuing to outpace supply, we are working closely with our partners to accelerate capacity expansion across critical components, including wafers, substrates and advanced packaging. Today, we are announcing new agreements with Kinsus and Samsung Electro-Mechanics in substrates and SPIL in advanced packaging, further strengthening our global supply chain capacity and resiliency. These build on partnerships we announced last month with Amkor, IBIDEN, KYEC and Unimicron.
In Q3, our purchase commitments increased to $390 billion, reflecting our confidence in the long-term sustainability of demand for NVIDIA infrastructure. We are securing data center and cloud capacity required to support our R&D efforts across hardware and software and increasingly in open models. Together, our data center lease and cloud service commitments increased to a little over $100 billion, driven primarily by our long-term leases.
Commitments associated with our neocloud agreements totaled $32 billion at the end of Q3. Under these agreements, we earn revenue from the upfront sale of our infrastructure and participate in upside in rental revenue above a minimum commitment. Total guarantees were $118 billion at the end of Q3, including the $105 billion in credit support for land, power and shell capacity for the PORTS-Pike campus we discussed last quarter. The site will exclusively host NVIDIA infrastructure under a 20-year leases to OpenAI. SoftBank Energy plans are on track with capacity expected to come online in phases beginning in calendar 2028.
We continue to work closely with our infrastructure capital partners to stand up the financing platforms. These partners are instrumental in financing the build-out of AI infrastructure. As we advance this work, we remain thoughtful in how we participate and disciplined in how we commit capital. We are proud of the role that we have played so far in supporting the ecosystem. As the market matures and additional sources of capital become available, we are rigorously evaluating future investments, commitments and guarantees.
NVIDIA's ability to extreme co-design across compute, networking, systems, algorithms and software are unmatched, driving the X-factor leaps in performance per watt and per dollar on an annual cadence. On the MLPerf Endpoints benchmark, Vera Rubin delivered 10x more tokens per watt over GB300. With Vera Rubin still early in its life cycle, we expect continuous software optimization to deliver even greater performance over time. The transition from Grace Blackwell to Vera Rubin has been exceptionally smooth. Just 3 months after we began production shipments in August, production ramps are landing in data centers at every major hyperscaler and leading neoclouds.
Our Hyperscale partners brought Vera Rubin into production in about 1 month, in half the time required with Grace Blackwell. Our networking franchise continues to scale with Q3 revenue exceeding $21 billion. Our co-packaged optics road map remains on track with CPO scale-out in volume production. At Supercomputing '26, Lambda announced the first production deployment of NVIDIA Photonics CPO switches, connecting more than 10,000 GB300 GPUs. Compared with pluggable optics, NVIDIA Photonics switches delivered 5x better power efficiency, 15x higher reliability and 10% lower latency. We remain on track to launch Vera Rubin Ultra next year and Feynman in 2028.
The next big leap is on the horizon. At GTC Berlin, we introduced DSX POD, a full stack modular AI factory reference architecture. Just as Grace Blackwell NVL72 moved rack assembly from the data center into the supply chain, DSX POD extends that approach to the entire cluster. Pods arrive ready to connect, cutting deployment times from months to weeks, accelerating time to first token with annual revenue generating from 1 gigawatt AI factory approaching $50 billion, reaching first token just 1 quarter earlier could generate more than $10 billion in additional revenue for the operator. We are working more than a dozen partners, including Dell, Foxconn, Schneider Electric and Vertiv. We plan to ship our first pods to customers in the U.S. and Central Europe early next year.
I'm going to move to the rest of the P&L. GAAP and non-GAAP gross margins were 74.8% and 74.9% and exceeded our guidance due to better product mix. GAAP and non-GAAP operating expenses were up 9% and 7% sequentially, respectively, on higher compute and infrastructure spend. Operating expenses came in marginally below our outlook, primarily due to the timing of NRE expenses. Our non-GAAP effective tax rate of 17% increased from a year ago, primarily due to higher revenue.
In Q3, we returned a record $28 billion to shareholders, $22 billion through share repurchases and $6 billion in cash dividends. We have returned 65% of free cash flow on a year-to-date basis. And going forward, we intend to return all excess free cash flow, net of strategic uses, to shareholders during this last quarter and going forward. During the quarter, we authorized an additional $150 billion increase to our share repurchase program, the largest in history, and we exited Q3 with $227 billion remaining under our authorization. We continue to plan to execute our buyback authorization by the end of fiscal 2028.
Now let me turn to the outlook for the fourth quarter. Please note, we will have an additional week in Q4. Total revenue is expected to be $130 billion, plus or minus 3%. We expect sequential growth to be driven by robust demand for AI infrastructure and the continued ramp of Vera Rubin. New export controls were released by the U.S. Commerce Department this week. However, with demand far exceeding supply, we expect our performance to remain strong. As mentioned earlier, we now expect fiscal 2028 revenue to grow approximately 75% off what is now a higher fiscal '27 base.
GAAP and non-GAAP gross margins are expected to be between 71% and 72%, consistent with the outlook we provided last quarter. We expect Q4 to mark the bottom of gross margins before improving to the 72% to 73% range in fiscal year '28. GAAP and non-GAAP operating expenses are expected to be approximately $11 billion and $10.7 billion, respectively, impacted by the additional week and transition-related costs. We will continue to prioritize investments that support market growth and further strengthen our market position.
For full year fiscal 2027, we continue to expect GAAP and non-GAAP tax rates to be between 16% and 18%, excluding any discrete items and material changes to our tax environment. Finally, we plan to file our 10-Q in the coming weeks. In the interim, we have provided a supplemental document with additional financial information on our investor website.
Toshiya Hari
Thank you, Colette. We will now transition to Q&A. Ian Buck, Vice President of Hyperscale and High Performance Computing, will be joining Colette to answer your questions. Operator, please poll for questions.
Operator
[Operator Instructions] Your first question comes from Joseph Moore with Morgan Stanley.
Phần hỏi đáp
Joseph Moore
Great. I wonder if you could talk about the next year outlook for 75% growth. I know you were constrained largely wafers, but there were other constraints as well. What gives you the confidence to raise that number? And do you still feel like -- you said that demand is going to be stronger than that number. At one point, you talked about 100% with the new BIS rules. Do you still think that, that 100% is kind of the right way to think about demand?
Colette Kress
Yes. Thank you, Joe, for the question. That is correct. We are looking into our next fiscal year. We had originally indicated that we would be at about 70% growth year-over-year, and now we are increasing that to 75% growth. We took a lot of time working into both our demand that continues to be growing as well as our additional supply that we are able to provide to our different customers. Those 2 things together continue to give us more and more growth opportunity for us.
Additionally, we did look at an assessment in terms of the current export controls. We're still going to spend some time studying those to see the different pieces of it. But we feel confident in terms of both the demand that we have in front of us as well as the supply and additional supply that we can do the 75%.
Operator
Your next question comes from Vivek Arya with Bank of America Securities.
Vivek Arya
Colette, I appreciate that some decisions are still evolving, but I'm curious, what is your message to NVIDIA investors about just the structural moats and the talent retention at the company. What kind of remains durable? What could change during this leadership transition? Just how do you and the Board envision the company's competitive positioning and the financial strategy as you work through this leadership transition?
Colette Kress
Let me see if I will start, and I'll have Harvey see if he could say a few words as well. Probably the most important thing that we think through right now is our employees. Our employees are absolutely important, instrumental to everything in this company. We are always thinking about our employees. Like every community that there is out there, there are diverse opinions. Our employees, though, we believe, are committed. They are committed to doing their life's work at NVIDIA, and that has been true about our company from the very beginning. We will take care of our employees during this transition to make sure that they are safe and making sure the work can be done in the current environment.
I think there's also a time to think about our leadership and how important our leadership is across this organization. Our leadership has stood up so strongly over the last week to support both what we are doing here at earnings, but also supporting all of our different employees around the world. Supporting them and moving to work their jobs in every single day as they always do is what I highly understand them to do always. But honestly, they have been even greater than that in terms of the work working seamlessly in terms of what we need to complete. Remember, we have the best technologists on the planet with a world-class multiyear road map, and I think that's important.
From a business perspective, it's business as usual. We've engaged with our customers. We've engaged with our suppliers, our partners. They're all 100% up behind us. Customers have expressed their support to the business outlook and helping us through that. They've been working closely with our suppliers to narrow the supply and demand gaps as well. And I see right now the leadership even from TSMC that was in our offices just this week, also fully engaged in helping us. We've secured memory. We've secured advanced packaging. But I think the most important thing is where we are internally as a company and the strength that we have to continue.
And I'll turn it over to Harvey to see if he has more.
Harvey Jones
I think Colette said it very well. From the Board's point of view, we take the search very seriously. We're going to look into internal candidates as well as external candidates. The Board appreciates that we're living in a world where the advanced silicon technologies and national security dynamics are making major changes in our marketplace. Our new CEO will have to navigate those waters very carefully. So the search will be thorough, and we want to install the CEO, the permanent CEO as quickly as possible, but we won't expedite the CEO before finding exactly the right person.
Operator
Your next question comes from Timothy Arcuri with UBS.
Timothy Arcuri
Colette, I wanted to ask about this new gross margin comment that you'll be in the 72% to 73% range for calendar 2027. And I wanted to see if you can speak to your ability to kind of maintain pricing power and margins over the longer term. And I ask because if you look at the Frontier model guys, they're leaning in even more aggressively on ASIC and they're talking about compute efficiency as new KPIs. Is there an acceptable rate of margin compression that you might be willing to live with to protect your competitive position given all that's going on?
Colette Kress
Thanks, Tim, for the question. In our last earnings, we did discuss where we saw our gross margins, not only for the last 2 quarters of fiscal year '27, but we also gave an outlook in terms of fiscal year '28. If you recall at the time, there was a significant change in terms of the cost of supply throughout the quarter that we were in as well as what we were going to move in terms of fiscal year '28. So it was important for us to rework in what we thought would be important. Yes, we thought about in terms of our prices and pieces of that and a lot of the costs were some of the things that we absorbed in those numbers that we included.
What we provided here in Q4 is our bottom. If you recall, that was our bottom as we honored many of the different POs that we had already received from our customers, and we just wanted to make sure that, that was held solid as we move to a new pricing as we moved into the new year. Your statement regarding the changes that we are making. Those changes regarding fiscal year '28 are in line with what we communicated last quarter that we see in fiscal year '28. We see no changes at this time. We feel confident in terms of those numbers that we will be able to do.
For the long term and our thoughts in terms of pricing going forward, remember, when you think about our full road map and our systems, probably one of the most important parts is they are very, very world-class. And everything that we enable within those systems is so important. But again, it drives such an important piece of the work that is completed. And I think our pricing is well thought through, but that is something for the future, we'll continue to think through what that pricing might be.
Operator
Your next question comes from C.J. Muse with Cantor Fitzgerald.
Christopher Muse
ACIE is a very important growth vector for NVIDIA. I'm hoping you could update us on your appetite for future vendor financing. And if there is any change here, what are the implications for growth for this important segment, particularly for your NCP customers?
Colette Kress
Thanks for the question, C.J. So let's kind of break down that group of folks that are in that AI&E (sic) [ ACI&E ] piece of it. One of the more important areas are always together. And one of those is going to be, yes, our neoclouds. Our neoclouds have been so instrumental in terms of both supporting enterprises and supporting many of the different countries around the world, particularly on the sovereign perspective. And we continue to be working with them. But always remember, there's also other different types of groups within that category as well. We spend a lot in terms of where we are shipping in terms of our enterprises customers.
Our enterprises and building their AI factories has been important for them and the neoclouds enable them to essentially have a private cloud for them to complete that. We've also seen in terms of many of our physical AI and others that are working in terms of that piece as well. So our Edge Computing can sometimes also just have Data Center help included in that. So we actually have a vast group of different types of customers that are within that group to watch.
Now how do we think about in terms of the future and potential financing that will be necessary. The most important part of the financing that we would probably review but have to review in detail to make sure it is the right thing is focusing on land, power, shell. The availability of land, power, shell is such a critical part of the whole structure in terms of building out those data centers. So they may need the assistance. That's a great asset for anyone to hold. So we will look at that as probably one of our top priorities.
We -- as you heard in our announcement here that we've been discussing about many of our financing funds that have been helping us, and they can be a very big part of that work as well in terms of the neoclouds or even any of the enterprises in the work that they want to do. We have a strong amount of companies that we work with that are not only working in terms of in enterprises or in big businesses, but keep in mind, they are also IG customers. Those IG customers can be very helpful, particularly when it comes to enabling the amount of money that is required. We can look at them in terms of the assistance of meeting in terms of prepaying, which can help in terms of their build of that data center.
So right now, I think we would look at this and taking a very strong look in terms of the types of companies that we would want to help in terms of the financing as a lot of things have changed over this last quarter, and we know that, that's a priority of ours, but we do have some very, very strong existing customers that we will be continuing to work with.
Operator
Your next question comes from Stacy Rasgon with Bernstein Research.
Stacy Rasgon
Colette, I wanted to follow up on that. In the CFO commentary, it looks to me a new sentence or new paragraph on the neoclouds. It says your commitments take effect only if the neoclouds secure financing to purchase the data center infrastructure products and related compute capacity deployed. Maximum commitment, blah, blah, blah, it says certain commitments may be reduced or eliminated if they don't do this. I guess I'm asking, is this -- it's a new sentence in the CFO disclosure. Is this something new in how you're doing business with the neoclouds? Or is this something that you've been doing all along and you're just telling us about it now? Because it sounds like you guys are not the ones that are actually looking to provide the financing. It reads like they have to secure their own financing in order to get the backstops and commitments from you.
Colette Kress
Thanks, Stacy, for the question. Let me see if I can clarify for everyone about what that means. When we are looking at many of our different types of guarantees, much of the land, power, shell that has to be built and put together, there is construction, there is the raising of capital that they may need through equity and/or debt in order to build out any of the buildings, get them ready for use, but also the ability to procure the compute, install the compute and get ready.
What we are indicating in our statement there is that is when our guarantee starts. So if all of that begins and unfortunately, they are delayed or they have a challenge in terms of building that, our guarantee doesn't exist anymore. Our guarantee for most of these are really about when they are ready to go, particularly when we think about our one in terms of Portsmouth in terms of what we could do.
These are important things to understand because that shows what we have done in terms of really thoughts in terms of very, very strong businesses that we felt could actually maintain creating the debt and equity that they needed to get that done and create the data center such that we felt very confident in terms of that guarantee. Is it new? No. Did we change anything? No. What this is only saying is that we wanted to make sure that the clarity, there was other communication in another part that would have already communicated this, but we just wanted to make sure it was very clear as we had been receiving some of the questions on how that exactly worked.
Operator
Your next question comes from Ben Reitzes with Melius Research.
Benjamin Reitzes
Nice presentation tonight, Colette and team. I wanted to ask about if you could give me any idea on what you think about this idea that there needs to be a big slowdown in 2028 due to the change in government and some of the policies that we may see. There's obviously -- I'm sure you're aware of the big fears that there could be a slowdown after your great calendar '27. Is there something that you're seeing increased confidence in this? Have you spoken to anyone in the administration or not administration, but in government since the blue wave? And anything that gives you confidence beyond '27 that you might want -- might be able to tell us about?
Colette Kress
Yes. So thanks, Ben, for the question. Where we stand right now, there is so much demand with such the largest companies continuing to work on so many agentic solutions, new models that are also available. And that's been our focus to assist. I'd say even in the last week, the continued ask, how are we going to receive more and more supply to help them with the demand that they have in front of them. So they're all getting ready for that.
I know there's been a lot of focus in terms of the data centers within the United States as well. Are they well positioned to take care of the communities that are around that? And I think that's very important that all companies are working together so that the communities understand both the jobs that are going to be newly available, but giving them the confidence regarding how they will be perceived and working within those areas as well.
So I think that is an area to focus on, but I think there is a lot of data centers already in use or ready to go that might be that case. We're always going to be engaged with the government, making sure that we are all aligned in terms of what we see in terms of the future. But I think the demand and what we have in terms of agentic will continue not only in the '27 perspective, but moving to the '28 as well as we'll continue to carry over the amount of demand that's in front of us.
Operator
Your next question comes from Joshua Buchalter with TD Cowen.
Joshua Buchalter
I want to offer a sincere good luck to you and the employees during this transition. In the prepared remarks, you were clear that demand continues to meaningfully outstrip supply. But should we expect any changes to how NVIDIA allocates its available capacity across the key buckets of end demand like hyperscalers, AI labs and neoclouds? You mentioned that AI labs are likely to be about 15% of revenue this year, which is, I think, roughly flat with some disclosures you gave over the summer. But how big could that bucket be, for instance, over time?
Colette Kress
Thanks for the question. So when we think about our allocation, we do think about very important parts that we need to understand to make sure that in the quarter that we need to ship, they will be able to use and use very efficiently in the very beginning. What that means is that land, power, shell, whether or not it's structured, ready to go, is already set up and completed to do so. That's one of the most important things.
The second thing, though, is also to think about their capital needs. Their capital, it needs to be set up, ready to go for us to work in terms of that allocation. Many of our customers outside of the IG will want to provide and think through in terms of their offtake as we will be needing an understanding of do they have that solid offtake that they will be able to run as well. So our allocation runs from we're here for everyone, particularly when there is already a PO available, and we feel those means in terms of capital will happen, and we will continue to be allocating as quickly as possible when we know the amount of supply that we will think we will have.
So I believe this is a process that will continue to grow as more supply comes in. We'll continue working in terms of more allocation. But right now, that's probably the best way that I would describe all of our process that we are doing. It is not anything in terms of any types of different pricing. All are getting the same pricing that is super important for us, and we just need to assure that they are ready on their side in terms of being able to accept the compute.
Operator
Your next question comes from Jim Schneider with Goldman Sachs.
James Schneider
I was wondering if you can maybe address capital allocation and overall capital priorities. It sounds like there's no change to your existing commitments to partners or supply chain partners or customers, and you've reiterated the buyback commentary you talked about previously. Can you maybe just kind of clarify whether there is anything about the hurdle rates or thresholds for customer financing specifically that is changing going forward? Or is that unchanged? And can you maybe also address the timing or relative sequencing of buybacks versus customer financing or the partner financing?
Colette Kress
Okay. So let me clarify for you, Jim, where we stand in terms of returning capital to shareholders. Super top priority for us, first and foremost, to understand. We already are with the $150 billion increase in our opportunity. It is going to be used and hopefully, in terms of fully used for the full fiscal year '28. And that is going to be something that we want to do, and we believe we have the capability to do so.
But equally, there are important areas that we need to keep in mind for the capital use. The top priority is going to be our suppliers. Our suppliers are working feverishly to build more demand for us. And some of that demand needs assistance -- excuse me, some of that supply needs assistance in terms of capital from any forms of different types of prepaids to help them in terms of capacity. We want to make sure that we're there for them so they can move as quickly as possible. I think that is a very great use of our capital to help support that part of it.
Additionally, there are areas where we will see some of our expected committed investments that we need to make that we may need to finish. Not a very large piece of it, but something that we need to do. When you think about just on the edge of a set of types of very strong customers that may need assistance in terms of financing of some type of guarantee, that is something we will do. It is a process that we will look very deeply in terms of the company, the strength of what they have planned as well as everything that they need in terms of that data center to do so.
So there will be a portion that we will do. But again, something other -- a lot of work to understand if those are the right types of investments that we need to make. So thinking about all of that together, I think there is definitely the room to continue all of the stock repurchases plans that we want to do and complete those before the end of fiscal year '28.
Operator
There are no further questions at this time. Colette Kress, I turn the call back over to you.
Colette Kress
Thank you. Thank you all for joining us today. Our momentum is real. Our demand is accelerating. Our supply is growing. Our customers and our partners are with us and employees are committed. We've got this. Thank you so much.
Operator
This concludes today's conference call. You may now disconnect.
Bài viết này có thể bao gồm nội dung do AI tạo ra hoặc dịch bởi AI và được biên tập lại bởi con người, bài viết chỉ nhằm mục đích tham khảo và cung cấp thông tin chung, không phải là khuyến nghị đầu tư.
Bài viết đề xuất












Bình luận (0)
Nhấn vào nút $ , nhập ký hiệu, và chọn để liên kết với một cổ phiếu, ETF, hoặc mã khác.