NVIDIA Corporation (NVDA) Narrative

Sources: fiscal 2026 Form 10-K (year ended January 25, 2026), accession 0001045810-26-000021; Form 10-Q for the quarter ended April 26, 2026, accession 0001045810-26-000052; and current reports on Form 8-K as cited.

---

Business

From the fiscal 2026 10-K, accession 0001045810-26-000021.

NVIDIA describes itself as having moved past its origins as a GPU designer to become "a data center scale AI infrastructure company." The company was incorporated in California in April 1993, reincorporated in Delaware in April 1998, and is headquartered in Santa Clara, California.

The platform. The technology stack is anchored by CUDA, the development platform that runs on every NVIDIA GPU, plus hundreds of domain-specific libraries, frameworks, SDKs and APIs (CUDA-X, NVIDIA AI Enterprise, NeMo, NIM, Omniverse, DRIVE, Clara). Beginning with the Blackwell architecture, NVIDIA sells at data-center scale rather than at chip scale: chips, networking, systems, software and algorithms are co-designed so that hundreds of thousands of GPUs operate as a single computer. Management stresses that the same programmable architecture serves several multi-billion-dollar end markets, and that more than half of NVIDIA's engineers work on software. Cumulative R&D since inception exceeds $76.7 billion. Over 7.5 million developers use CUDA.

Reporting segments (as reported through Q1 FY2027). Two operating segments:

• Compute & Networking, Data Center accelerated computing and networking platforms, AI

solutions and software, and Automotive/autonomous-vehicle platforms. FY2026 revenue $193.5 billion (up 67%), operating income $130.1 billion.

• Graphics, GeForce GPUs for gaming and PCs, and Quadro/NVIDIA RTX GPUs for enterprise

workstation graphics. FY2026 revenue $22.5 billion (up 57%), operating income $9.2 billion.

Four end markets as described in the 10-K: Data Center, Gaming, Professional Visualization, and Automotive.

• Data Center, rack-scale systems, subsystems and modules combining GPUs, Grace CPUs, DPUs,

NVLink interconnects and switches, InfiniBand and Ethernet platforms, plus software and services. Customers include all major public and private cloud providers, AI model makers, enterprises, startups and public-sector entities, reached both directly and through OEMs, ODMs, system integrators and distributors. NVIDIA launched Blackwell in fiscal 2025, launched and scaled Blackwell Ultra in fiscal 2026, introduced NVLink Fusion (letting hyperscalers and custom ASIC designers attach their own CPUs/XPUs to the platform), and unveiled the Rubin platform, expected to enter production shipment in the second half of fiscal 2027, claiming up to a 10x reduction in cost per token versus Blackwell.

• Gaming, GeForce RTX GPUs for desktops and laptops, GeForce NOW cloud gaming, and SoCs and

development services for game consoles. Includes RTX ray tracing and DLSS.

• Professional Visualization, RTX PRO GPUs for design, engineering and content creation,

increasingly pulled by on-premises enterprise AI work.

• Automotive, the DRIVE Hyperion end-to-end autonomous-driving stack (DRIVE AGX hardware,

DRIVE OS, reference sensor set, modular AV software), sold through several hundred ecosystem partners including OEMs, robotaxi operators, tier-1 suppliers and startups.

Manufacturing. Fabless. NVIDIA partners with suppliers for all phases, wafer fabrication, assembly, test and packaging. Named suppliers in the 10-K: TSMC and Samsung for wafers; SK Hynix, Micron and Samsung for memory; Hon Hai (Foxconn), Wistron and Fabrinet for assembly, testing and packaging. CoWoS is used for advanced packaging. The supply chain is "mainly concentrated in Asia," with expansion under way into the U.S. and Latin America. In periods of growth the company places non-cancellable inventory orders ahead of historical lead times and pays premiums or deposits to secure capacity.

Competition. Named competitors: AMD, Huawei and Intel in GPUs/custom accelerators; Alibaba, Alphabet, Amazon, Baidu, Huawei and Microsoft among cloud companies designing their own internal accelerator hardware and software; Amazon, Huawei and Microsoft in Arm-based CPUs. NVIDIA identifies performance, breadth of offering, channel access, software support, API conformity, manufacturing capability, processor pricing and total system cost as the principal competitive factors.

Seasonality. Desktop gaming products typically see stronger revenue in the second half of the fiscal year; management cautions historical seasonality may not repeat.

People. Approximately 42,000 employees in 38 countries at fiscal 2026 year-end, 31,000 in research and development, 11,000 in sales, marketing, operations and administration. Roughly 5,900–6,000 employees are in Israel, primarily supporting the networking business.

IP. Protection rests on patents, trademarks, trade secrets, NDAs and licensing. Issued patents expire between March 2026 and June 2045.

---

Risk factors

From the fiscal 2026 10-K, accession 0001045810-26-000021, as updated by Part II, Item 1A of the Q1 FY2027 10-Q, accession 0001045810-26-000052.

Export controls and China, the dominant regulatory risk

NVIDIA describes "a series of shifting and expanding export control restrictions" over three years:

The blunt conclusion in the filings: "Under the current rules and geopolitical landscape, we are unable to create and deliver a competitive product for China's data center market that receives approval from both the USG and the Chinese government." As of the end of Q1 FY2027, NVIDIA was able to ship uncontrolled products to China (gaming and workstation GPUs) but was "effectively foreclosed from competing in China's data center computing/compute market," and states that this foreclosure "helped our competitors build larger developer and customer ecosystems to challenge us worldwide."

Related exposures management flags: possible penalties or business restrictions if Chinese regulators conclude the Mellanox commitments were breached; controls that could extend to networking products (high-speed interconnects) or to additional gaming GPUs; the French Competition Authority questioning whether gaming and data center GPUs are separate product categories, which could change which export controls apply in France and Europe; possible mandated chip tracking/throttling mechanisms, which NVIDIA argues would introduce system vulnerabilities and liability; China's public questioning of whether H20 contains built-in vulnerabilities (NVIDIA publicly denied this); a Chinese government "Action Plan" endorsing compute-per-watt and per-memory-bandwidth standards for accelerators in Chinese data centers; USG "Connected Vehicle" import restrictions that could affect automotive solutions; and proposed U.S. legislation (the Remote Access Security Act) that could bar cloud services to companies ultimately parented in China. Export controls also encourage customers outside China to "design out" U.S. semiconductors, and increase the risk that a product becomes unsellable between design and launch. Product warehousing and distribution for a substantial portion of NVIDIA's products runs through Hong Kong.

Customer concentration

• FY2026: one direct customer was 22% of total revenue and another 14%, both

primarily Compute & Networking. (FY2025: 12%, 11%, 11%. FY2024: a single customer at 13%.)

• Q1 FY2027: three direct customers were 21%, 17% and 16% of total revenue (Q1 FY2026: two

at 16% and 14%).

• Accounts receivable at April 26, 2026: three direct customers were 30%, 18% and 16% of the

balance (at January 25, 2026: 25%, 18% and 13%).

• Indirect customers, those buying through OEMs, ODMs, system integrators and distributors, also include names individually representing 10% or more of revenue. NVIDIA estimates that "one

AI research and deployment company" contributed a meaningful amount of revenue by purchasing cloud services from NVIDIA's customers, in both FY2026 and Q1 FY2027.

• Most sales are on a purchase-order basis, so customers can generally cancel, change or delay

commitments with little notice and without penalty. Losing a large customer, or being barred from selling to one by trade restrictions, would likely harm results.

Supply, demand and capacity commitments

Manufacturing lead times have exceeded 12 months. To secure supply, NVIDIA pays premiums, provides deposits, and enters long-term supply and prepaid manufacturing/capacity agreements. Management warns that these risks "have increased and may continue to increase as our purchase obligations and prepaids have grown and are expected to continue to grow and become a greater portion of our total supply." Because data center products are one component of a complex buildout, a shortage of any single component can have a broad revenue impact; supply availability and rising prices for memory and other components may push data center buildout prices higher. Publicly announced government or corporate intentions to buy NVIDIA products are often non-binding and complicate demand estimation. The 10-K states supply constraints were expected to be a headwind to Gaming in Q1 FY2027 and beyond. Historical precedent cited: Q2 FY2025 gross margin was hurt by inventory provisions for low-yielding Blackwell material.

The annual product cadence, a new data center architecture each year, magnifies transition risk: qualification time, channel inventory drawdowns ahead of launches, customers whose data center infrastructure is not ready, and unanticipated quality or yield problems.

The infrastructure-buildout dependency

A distinctive risk in these filings: NVIDIA's revenue now depends on its customers' ability to build data centers. "The availability of data centers, energy, and capital to support the buildout of NVIDIA AI infrastructure by our customers and partners is crucial." Expanding energy capacity is a multi-year process with regulatory, technical and construction obstacles, and access to capital is "particularly constrained for less-capitalized companies." Data centers also depend on clean water and predictable energy; power/water shortages, land or permitting constraints could cap customers' capacity and therefore NVIDIA's sales.

Ecosystem investments, guarantees and counterparty risk

NVIDIA has taken on financial exposure to the customers and partners it sells to. It invested $17.5 billion in private companies and infrastructure funds in FY2026, including AI model makers that buy its products directly or through cloud providers, mostly illiquid and non-marketable. It also provided $3.5 billion of land, power and shell guarantees to early-stage companies over multi-year periods; if escrow and partner operations do not cover a default, NVIDIA may assume the underlying leases or sublease them. The 10-K states NVIDIA "is finalizing an investment and partnership agreement with OpenAI," with no assurance any agreement is signed or completed. Commercial arrangements, long-term capacity purchase obligations and financial guarantees, expose NVIDIA to counterparty inability to fund commitments, project delays, distress or insolvency. NVIDIA has been asked to offer customer financing arrangements but states it has not entered into any; if it did, it could reduce upfront cash flows and increase credit risk. The investment portfolio also carries industry-sector concentration risk.

Technology, competitive and execution risks

• Groq license. NVIDIA entered a non-exclusive intellectual property license arrangement with

Groq, Inc. requiring significant, nonrefundable payments. Incorporating the licensed technology requires substantial engineering effort, may not hit expected timelines or achieve customer/ecosystem adoption, and NVIDIA may be unable to recover the associated costs.

• Customer in-house silicon. The Q1 FY2027 update sharpens this: "Some of our customers are

developing their own ASICs and other products, including designs optimized for certain workloads that may not require all of the features and functionality our data center systems provide."

• Open-source models. High-quality open-source foundation models are making advanced AI broadly

accessible; if deployed on competitors' platforms they could reduce demand for NVIDIA products. Conversely, restrictions on supporting China-origin open models (DeepSeek, Qwen, KIMMI) would also hurt.

• Multi-year cloud service agreements entered to support R&D have shifted in timing and

availability and may not deliver anticipated benefits.

• Standalone software (NVIDIA AI Enterprise, Omniverse, DRIVE) may fail to produce meaningful

revenue.

Other substantive risks

• Product defects and warranty. Complex hardware/software has contained defects; NVIDIA

generally remains responsible to the customer for warranty defects even where components come from partners.

• Security. Third-party supply-chain compromise; ransomware; AI-assisted attacks; prior

incidents at GeForce NOW exposing third-party game source code held by NVIDIA; insider errors such as employees posting company data to third-party sites.

• Geographic concentration. Headquarters, a large portion of current data center capacity, and

much R&D sit in California (earthquake, wildfire, power-shutoff exposure); critical operations, finished goods and suppliers are concentrated in Asia, Taiwan and South Korea supply dependence is called out explicitly. The Israel operation supports the networking business, and employees there have periodically been on extended active military duty.

• ERP transition. A phased enterprise-resource-planning upgrade is under way; implementation

problems could affect the ability to operate and to report results accurately.

• Tax. Most income is taxed in the U.S., with a significant portion qualifying as

foreign-derived deduction eligible income (FDDEI), reduction of that deduction would hurt results. OECD Pillar One/Pillar Two adoption adds uncertainty. Stock-price declines reduce future stock-based-compensation tax benefits and raise the effective rate.

• Litigation. A putative securities class action (*In re NVIDIA Corporation Securities

Litigation*) over statements about channel inventory and cryptocurrency-mining demand in 2017–2018, plus related derivative suits.

• Change of control. Under the Xbox agreement with Microsoft, if someone offers to buy at least

30% of NVIDIA's outstanding common stock, Microsoft may have first and last rights of refusal, alongside Delaware anti-takeover law and charter/bylaw provisions.

• Regulatory scrutiny of the AI position. Broad information requests from competition

regulators in the EU, U.S., UK, China and South Korea covering GPU sales, supply allocation, foundation-model investments and partnerships, roadmaps, and customer/supplier agreements, described as expensive, burdensome, and likely to recur.

---

Management's discussion and analysis, fiscal 2026

From the fiscal 2026 10-K, accession 0001045810-26-000021. Fiscal year ended January 25, 2026.

Headline results

By market: Data Center up 68%; Gaming up 41% on Blackwell demand; Professional Visualization up 70% on Blackwell demand and the DGX Spark launch; Automotive up 39% on self-driving platform adoption. Within Compute & Networking, Data Center computing revenue grew 59% and Data Center networking revenue grew 142%, the latter driven by NVLink compute fabric for GB200/GB300 systems plus Ethernet and InfiniBand.

Revenue from customers headquartered outside the United States was 31% of total in FY2026, down from 41% in FY2025, a direct read-through of the China foreclosure.

Margin bridge

Gross margin fell from 75.0% to 71.1% for two reasons management names: the business model transitioned from selling Hopper HGX systems to selling Blackwell full-scale data center solutions, and the $4.5 billion H20 charge. Provisions for inventory and excess inventory purchase obligations totaled $7.2 billion in FY2026 versus $3.7 billion in FY2025 (FY2026 including the $4.5 billion H20 amount); releases from selling previously reserved inventory and settling excess purchase obligations were $1.5 billion versus $689 million. Net unfavorable gross-margin impact: 2.6% in FY2026, 2.3% in FY2025.

Operating expenses rose 41%. R&D was driven by a 29% increase in compensation and benefits and a 79% increase in compute and infrastructure spending.

Below the line

Total other income, net was $11.1 billion versus $2.6 billion, driven primarily by unrealized gains on non-marketable and publicly-held equity securities, including gains on the previously announced investment in Intel common stock. Income tax expense was $21.4 billion (15.1% of pre-tax income) versus $11.1 billion (13.3%); the rate rose because stock-based-compensation, FDDEI and R&D-credit benefits grew more slowly than pre-tax income. The One Big Beautiful Bill Act (OBBBA), enacted July 2025, is reflected for currently effective provisions. NVIDIA released $711 million of valuation allowance on certain state deferred tax assets. Unrecognized tax benefits at year-end were $4.0 billion including $374 million of interest and penalties; the IRS is examining fiscal years 2023 and 2024.

Cash, capital and commitments

• Cash, cash equivalents and marketable securities at January 25, 2026: $62.6 billion

(approximately $1.7 billion held outside the U.S. without accrued repatriation taxes).

• Operating cash flow rose on higher revenue. Investing outflows rose on higher purchases of

equity investment securities and the execution of the non-exclusive license agreement with Groq. Financing outflows rose mainly on higher share repurchases.

• Buybacks: the Board approved an additional $60.0 billion authorization on August 26, 2025.

NVIDIA repurchased 282 million shares for $40.4 billion in FY2026, leaving $58.5 billion authorized at year-end, and repurchased a further 8 million shares for $1.5 billion between January 26 and February 20, 2026 under a pre-established trading plan.

• Dividends: $974 million paid in FY2026.
• Debt: $8.5 billion of senior notes outstanding at fiscal year-end, all fixed rate. In January

2026 the commercial paper program was increased to $25.0 billion, with none outstanding at year-end.

• Capital expenditures: $6.1 billion in FY2026 versus $3.4 billion in FY2025, with management

expecting a further increase in FY2027.

• Purchase commitments at January 25, 2026: manufacturing, supply and capacity commitments of

$95.2 billion, substantially all payable through fiscal 2027; multi-year cloud service agreement commitments of $27 billion; investment commitments of $11.4 billion.

• Guarantees: agreements entered in FY2026 to guarantee partners' facility lease obligations in

the event of default, in exchange for warrants, maximum gross exposure $3.5 billion, reducing as partners pay lessors over 5-to-7-year terms, with $712 million placed in escrow by partners. Classified as credit derivatives; fair-value changes were not material.

Management's stated challenges

Beyond export controls, the 10-K highlights: the availability of data centers, energy and capital to support customer buildouts; the complexity of annual product transitions (Blackwell Ultra/GB300 began production shipment in Q2 FY2026; Rubin follows); the rise of open-source foundation models; supply-chain expansion from Asia into the U.S. and Latin America and whether local ecosystems can ramp on time; and macroeconomic factors including tariffs, inflation, rates and capital-market volatility.

---

Current quarter, Q1 fiscal 2027

From the 10-Q for the quarter ended April 26, 2026, accession 0001045810-26-000052.

Results

New reporting framework

The most consequential disclosure change this quarter: NVIDIA is transitioning to a new market- platform reporting framework "that better reflects our current and future growth drivers." Two market platforms replace the previous four markets:

• Data Center, split into two sub-markets, Hyperscale (public clouds and the world's

largest consumer internet companies) and ACIE (AI Clouds, Industrial, and Enterprise, addressing purpose-built AI data centers and AI factories across industries and countries).

• Edge Computing, devices for agentic and physical AI: PCs, game consoles, workstations,

AI-RAN base stations, robotics and automotive.

Prior periods were recast:

Legacy reportable segments still disclosed: Compute & Networking revenue $74,550M (up 88% YoY), operating income $53,335M (up 142%); Graphics revenue $7,065M (up 58%), operating income $2,941M (up 79%).

What drove the quarter

Data Center growth came from the ramp of Blackwell 300 products and demand for InfiniBand, Spectrum-X Ethernet and NVLink. Hyperscaler revenue rose sequentially and held at roughly 50% of Data Center revenue, with the other half from a broadening base of AI Clouds, industrial, enterprise and sovereign customers, the diversification management is now organizing its reporting around.

China: zero. "No shipments of Data Center Hopper products to China occurred during the quarter, compared with $4.6 billion in the first quarter of fiscal year 2026." Revenue from customers headquartered outside the United States fell to 22% of total in Q1 FY2027 from 42% a year earlier. Under the February 2026 H200 license program, no revenue has been generated to date, and NVIDIA does not know whether imports into China will be allowed; H200s must clear a U.S. inspection process and carry a 25% U.S. import tariff.

Edge Computing rose 29% year over year on robust Blackwell workstation demand, partially offset by slower consumer PC demand "tempered by elevated memory and systems prices."

Gross margin of 74.9% was up 14.4 points year over year almost entirely because the prior-year quarter absorbed the $4.5 billion H20 charge, and was roughly flat sequentially. Inventory and excess-purchase-obligation provisions were $1.1 billion this quarter versus $5.3 billion a year ago; net unfavorable gross-margin impact 1.2% versus 11.0%.

Operating expenses rose 52% year over year. R&D of $6,321M (up 58%) was driven by a 112% increase in compute and infrastructure, a 31% increase in compensation and benefits, and a 204% increase in engineering development materials for new product introductions, a visible tell on the pace of the Rubin ramp.

Below the line, the investment portfolio now moves earnings

Total other income, net was $16,367M versus $272M a year ago. The swing came almost entirely from unrealized gains on publicly-held equity securities of $13.4 billion and on non-marketable equity securities of $2.6 billion. Net income of $58.3 billion is therefore materially inflated relative to operating income of $53.5 billion by mark-to-market gains, not operations.

The balance-sheet counterpart: non-marketable equity securities grew from $22,251M at the start of the quarter to $42,336M, driven by $17,899M of net additions plus $2,603M of unrealized gains. NVIDIA also holds $1.0 billion of infrastructure-fund investments accounted for under the equity method, with maximum loss exposure (invested plus committed) of $2.3 billion. MD&A states NVIDIA made $18.6 billion of investments in private companies and infrastructure funds during the quarter, some of them AI model makers that may indirectly buy or use NVIDIA products in the cloud. A hypothetical 10% decline in publicly-held equity securities would cut their fair value by $3.9 billion (versus $1.8 billion at January 25, 2026).

Income tax expense was $11.6 billion, 16.6% of pre-tax income, versus 14.3% a year ago; the rate rose because stock-based-compensation benefits grew more slowly than pre-tax income. Unrecognized tax benefits at quarter-end were $4.5 billion including $439 million of interest and penalties.

Balance sheet, cash and commitments

• Cash, cash equivalents and marketable debt securities: $50.3 billion (from $49.7 billion),

plus $30.2 billion of marketable equity securities.

• Operating cash flow $50,344M (from $27,414M); investing outflow $26,429M (from $5,216M) on

higher equity-security purchases; financing outflow $21,283M (from $15,553M) on higher buybacks.

• Inventories $25,797M (from $21,403M), with raw materials nearly doubling to $6,647M, a

build ahead of the Rubin ramp. Inventory provisions of $0.8 billion were recorded in the quarter.

• Accrued purchase consideration of $3,957M relates to the Groq non-exclusive license

agreement. Excess inventory purchase obligations accrued: $3,121M.

• Taxes payable jumped to $10,638M from $2,669M; no federal income tax payments were made in Q1

FY2027, and Q2 FY2027 is scheduled to include two payments.

• Commitments at April 26, 2026 grew sharply: manufacturing, supply and capacity commitments of

$119 billion (up from $95.2 billion at January 25, 2026), of which $95 billion falls due in the remainder of fiscal 2027 and the balance across fiscal 2028–2031; multi-year cloud service commitments of $30 billion ($6B remainder of FY2027, then $7B, $7B, $5B, $3B, $2B); other vendor commitments of $6 billion; and investment commitments of $27 billion (up from $11.4 billion), expected to be made through the remainder of fiscal 2027.

• Debt unchanged at a $8,470M net carrying amount across seven note series; commercial paper

capacity $25.0 billion with nothing outstanding. Estimated fair value of debt $7.4 billion.

• Product warranty liability $2,948M.
• Buybacks: 108 million shares for $20.2 billion in the quarter (Q1 FY2026: 126 million for

$14.5 billion), leaving $38.5 billion authorized at April 26, 2026. Average prices paid ranged from $183.38 to $186.35 across the three monthly windows, with 91.1 million of the 108.3 million shares repurchased in the March 23 – April 26 window. A further ~12 million shares worth $2.1 billion were withheld through net share settlement for employee equity tax obligations. Dividends paid: $243 million.

Guidance and forward statements in the filing

The 10-Q contains no numeric revenue guidance. Forward-looking statements made: the Rubin platform is expected to start shipping in the second half of fiscal year 2027; capital expenditures are expected to increase in fiscal 2027; ecosystem investment is expected to continue; and management repeats that customer access to data centers, energy and capital is the gating factor on the buildout.

Acquisitions during the quarter

Two small acquisitions closed during the quarter, disclosed only through the unregistered-securities item and without target names or total purchase prices:

• February 17, 2026, acquired a company; issued 37,890 shares to a key employee, valued at

approximately $7 million at the closing price on the issuance date.

• April 10, 2026, acquired a company; issued 72,972 shares to key employees, valued at

approximately $14 million at the closing price on the issuance date.

Both issuances were made under the Section 4(a)(2) private-placement exemption.

Litigation update

In the securities class action (In re NVIDIA Corporation Securities Litigation, N.D. Cal. 4:18-cv-07669-HSG), the district court granted plaintiffs' motion for class certification on March 25, 2026, certifying a class of purchasers of NVIDIA common stock between August 10, 2017 and November 15, 2018. On April 8, 2026, NVIDIA petitioned the Ninth Circuit for permission to appeal that order under Rule 23(f). Related derivative actions in the Northern District of California, the District of Delaware and the Delaware Court of Chancery remain stayed or administratively closed pending final resolution. No contingent liabilities are accrued for these matters; NVIDIA considers loss reasonably possible but not probable, and not reasonably estimable.

Insider trading arrangements

Director Tench Coxe adopted a Rule 10b5-1 arrangement on March 19, 2026 for up to 8,000,000 shares (expiring October 30, 2027), stated to be solely for gifts to charitable donor-advised funds. CFO Colette Kress terminated a Rule 10b5-1 arrangement on April 10, 2026 covering 500,000 shares; no shares had been sold under it.

---

Subsequent events

Events after the April 26, 2026 quarter end.

Capital return, announced May 18, 2026 (10-Q, accession 0001045810-26-000052)

• The Board approved an additional $80.0 billion share repurchase authorization, without

expiration, on top of the $38.5 billion remaining at quarter end.

• NVIDIA increased the quarterly cash dividend 25-fold, from $0.01 to $0.25 per share, for

shareholders of record on June 4, 2026, paid June 26, 2026.

$25.0 billion senior notes offering completed June 18, 2026 (Form 8-K filed June 18, 2026, accession 0001193125-26-275783), NVIDIA's largest debt raise to date, roughly tripling outstanding debt from the $8.5 billion carried at quarter end. Seven tranches:

Offered under shelf registration statement File No. 333-287619. The underwriting agreement, dated June 15, 2026, was entered with Goldman Sachs & Co. LLC, J.P. Morgan Securities LLC and Morgan Stanley & Co. LLC as representatives of the underwriters. The notes were issued under the September 16, 2016 base indenture with Computershare Trust Company, N.A. as trustee (successor to Wells Fargo Bank, N.A.), supplemented by an officers' certificate dated June 18, 2026.

Management and board changes

• Chief Accounting Officer transition *(Form 8-K filed April 27, 2026, accession

0001045810-26-000026)*, Donald Robertson, VP and CAO, notified the company on April 24, 2026 of his retirement from that role effective May 4, 2026, remaining as VP, Finance until July 1,

1. Scott Gawel, 55, was appointed VP and CAO effective on commencement of employment

May 4, 2026, serving as principal accounting officer. Gawel joined from Intel, where he was Corporate VP and CAO from 2022 to 2026, and previously held senior controller roles at Oracle. Annual base salary $800,000 plus two new-hire equity grants.

• Board addition (Form 8-K filed May 8, 2026, accession 0001045810-26-000028), on May 7,

2026 the Board appointed Suzanne Nora Johnson as a director and Audit Committee member, effective July 13, 2026, expanding the Board from ten to eleven seats. Compensation: an initial RSU grant with a $255,000 target value vesting semi-annually over three years, a pro-rated annual RSU grant with a $255,000 annualized target value, and a pro-rated annual cash retainer with an $85,000 annualized value.

• Head of worldwide sales succession *(Form 8-K filed July 2, 2026, accession

0001045810-26-000060)*, on June 28, 2026, Ajay K. Puri, Executive Vice President, Worldwide Field Operations and a 21-year NVIDIA executive, notified the company of his intention to retire from the role effective when his successor starts; he is expected to remain in a senior advisory capacity. On July 1, 2026, NVIDIA appointed Nicholas Parker, 55, as Executive Vice President, Worldwide Field Operations, effective on commencement of employment anticipated August 24, 2026. Parker joins after 26 years at Microsoft, most recently as Executive Vice President and Chief Business Officer of Microsoft's Worldwide Sales & Solutions organization (2025–2026), and previously President, Industry & Partner Sales (2022–2025) and Vice President, Global Partner Solutions (2020–2022). Compensation: $1,000,000 base salary, a $5,000,000 sign-on bonus, target annualized cash variable compensation of $1,500,000 under the fiscal 2027 plan (pro-rated), a $35,000,000 target-value RSU grant vesting over approximately four years, and a $5,000,000 target-value multi-year performance RSU grant vesting on total shareholder return relative to the S&P 500 over three years.

2026 Annual Meeting held June 24, 2026 (Form 8-K filed June 30, 2026, accession 0001045810-26-000056), shareholders elected all ten director nominees to serve until the 2027 annual meeting. Founder and CEO Jen-Hsun Huang drew the strongest support of the slate (16.65 billion shares for, 166 million against); directors Stephen C. Neal (2.23 billion against), Harvey C. Jones (1.57 billion against), A. Brooke Seawell (1.50 billion against) and Tench Coxe (1.40 billion against) drew the most opposition.

No acquisitions, divestitures or litigation settlements were disclosed as occurring after the April 26, 2026 quarter end.