revenue $81.6B, net income $58.3B. NVIDIA Q1 FY27: Revenue +85% on Blackwell ramp; net income +211% on non-operating gains

The district court granted class certification on March 25, 2026, allowing the securities lawsuit to proceed as a class action covering investors who purchased NVIDIA stock between August 10, 2017 and November 15, 2018. NVIDIA filed a petition to appeal this certification order on April 8, 2026. Class certification significantly expands potential liability exposure and litigation costs, as the company now faces claims on behalf of a broad investor class rather than individual plaintiffs.

One plaintiff (City of Westland Police and Fire Retirement System) was voluntarily dismissed with prejudice from the Delaware Chancery derivative action on August 11, 2025. This reduces the number of parties pursuing derivative claims against NVIDIA officers and directors, though other plaintiffs remain in the case.

Manufacturing, supply, and capacity commitments increased from $29.8 billion to $119 billion year-over-year, a 300% increase. The current filing attributes this to "data center-scale production and longer future ordering horizons across current and future product architectures." $95 billion of the $119 billion is due in the remainder of fiscal 2027. This reflects significantly expanded production scale and longer-term supply agreements, likely driven by AI data center demand.

Cloud service commitments increased from $10.6 billion to $30 billion and are now disclosed separately with a detailed payment schedule extending through fiscal 2032 and beyond. The current filing notes these agreements "will be primarily used to support our research and development efforts," dropping the prior reference to DGX Cloud offerings. The increase reflects expanded cloud infrastructure needs, likely for AI model training and development.

Product warranty liabilities increased from $2.1 billion (April 27, 2025) to $2.9 billion (April 26, 2026), a 38% year-over-year increase. Both filings note the additions primarily relate to the Compute & Networking segment. The increase reflects higher warranty reserves as shipment volumes grow, consistent with expanded data center product deployments.

The loss contingency disclosure now explicitly references "regulatory and government inquiries and investigations" in addition to ordinary-course legal actions. This language was not present in the baseline. The addition suggests NVIDIA is subject to regulatory or government inquiries beyond the disclosed litigation, though the company maintains these matters are not expected to have a material adverse effect.

The current filing states that Blackwell now accounts for the majority of system shipments, a milestone not mentioned in the baseline (which only noted Blackwell's ramp was "extending beyond large cloud service providers"). This indicates Blackwell has become the dominant product in the mix, a significant product-transition milestone.

The current filing expands the infrastructure-constraint disclosure to detail the multi-year complexity of energy-capacity expansion (regulatory, technical, construction challenges) and the capital-access difficulties for less-capitalized companies. The baseline mentioned the constraint in one sentence; the current filing provides operational and financing color on the bottlenecks.

The current filing discloses a new product platform (Rubin) with a shipping timeline (second half of FY2027). The baseline made no mention of Rubin. This is forward-looking product-roadmap guidance.

The current filing discloses a new USG licensing program (started February 2026) allowing limited H200 shipments to China, with no revenue generated to date, an inspection requirement, and a 25% tariff on U.S. import. The baseline contained no mention of this program. This is a new regulatory development affecting China revenue potential.

The current filing discloses $18.6 billion in Q1 FY27 investments in private companies and infrastructure funds (some involving AI model makers that may indirectly purchase NVIDIA products) and notes investments in publicly-held equity securities with significant value fluctuation risk. The baseline contained no comparable disclosure. This is a new, material capital-allocation action.

The current filing introduces a new revenue-reporting framework: two market platforms (Data Center, Edge Computing), with Data Center split into Hyperscale and ACIE (AI Clouds, Industrial, Enterprise). The baseline had no such framework. This is a presentational change to align disclosure with the company's view of its growth drivers; the underlying business segments (Compute & Networking, Graphics) remain unchanged.

Data Center revenue grew from $39.1B (Q1 FY26) to $75.2B (Q1 FY27), a 92% YoY increase (vs. 73% in the prior year). The current filing attributes the growth to Blackwell 300 ramp and networking solutions (InfiniBand, Spectrum-X Ethernet, NVLink). The baseline attributed growth to "demand for our accelerated computing platform used for large language models, recommendation engines, and generative and agentic AI applications." The dollar figures and growth rates are materially higher.

The current filing states zero Data Center Hopper shipments to China in Q1 FY27, vs. $4.6B of H20 sales in Q1 FY26 (before the April 2025 export-control action). This is a year-over-year comparison showing the revenue impact of the export controls, not a new development within the current quarter.

The current filing reports Edge Computing revenue ($6.4B, +29% YoY, +10% QoQ) as a new market-platform category, driven by Blackwell workstation demand and offset by slower consumer PC demand. The baseline had no Edge Computing category (it reported Gaming, Professional Visualization, and Automotive separately). This is a presentational change; the underlying products are the same.

Gross margin was 74.9% in Q1 FY27 vs. 60.5% in Q1 FY26 (a 14.4-point YoY increase). The current filing attributes the increase to the non-recurrence of the prior year's $4.5B H20 charge. The baseline (Q1 FY26) reported a gross-margin decrease vs. the prior year (Q1 FY25: 78.4%) due to the H20 charge. This is a year-over-year comparison showing the margin recovery after the one-time charge.

Operating expenses grew 52% YoY in Q1 FY27 (vs. 44% in Q1 FY26), driven by compensation, compute/infrastructure, and engineering development. The current filing does not break out the sequential drivers as granularly as the baseline. The dollar amounts are $7.6B (Q1 FY27) vs. $5.0B (Q1 FY26).

R&D expense growth drivers in Q1 FY27: compute/infrastructure +112% (vs. +70% in Q1 FY26), compensation +31% (vs. +34%), engineering development +204% (vs. +182%). The current-year increases in compute/infrastructure and engineering development are materially higher, reflecting accelerated investment in new product development.

Other income swung from a $180M loss in Q1 FY26 (driven by unrealized losses in publicly-held equity securities, particularly CoreWeave and Arm) to a $15.9B gain in Q1 FY27 (driven by $13.4B unrealized gains in publicly-held equity securities and $2.6B in non-marketable equity securities). This is a $16.1B favorable swing, reflecting mark-to-market volatility in the investment portfolio.

The effective tax rate rose from 14.3% (Q1 FY26) to 16.6% (Q1 FY27), primarily due to stock-based compensation tax benefits not keeping pace with the increase in pretax income. The baseline's rate increase (13.9% to 14.3%) was also driven by lower stock-comp benefits, partially offset by FDII. The current filing does not mention FDII as an offset.

The current filing reports $50.3B in cash/cash equivalents/marketable debt securities plus $30.2B in marketable equity securities (total $80.5B). The baseline reported $53.7B in cash/cash equivalents/marketable securities (no separate equity-securities line). The current filing's separate equity-securities disclosure reflects the $18.6B investment activity and the $13.4B unrealized gain in publicly-held equity securities. The cash/debt-securities balance declined slightly ($53.7B to $50.3B), but total liquid assets increased materially when equity securities are included.

Q1 FY27 repurchases: 108M shares for $19.3B (vs. 126M shares for $14.5B in Q1 FY26). Remaining authorization: $38.5B (vs. $24.3B). The dollar amount of repurchases increased 39% YoY, reflecting higher share prices and continued capital return. The higher remaining authorization reflects the May 2026 $80B authorization increase (disclosed separately).

The current filing discloses a new $80B share-repurchase authorization approved May 18, 2026, with no expiration. The baseline contained no such disclosure. This is a new, material capital-allocation action.

The current filing discloses a 25x dividend increase (from $0.01 to $0.25 per share) effective May 18, 2026. The baseline reported $244M in Q1 FY26 dividends but did not disclose a per-share rate or an increase. This is a new, material capital-allocation action.

The commercial paper program size increased from $575M (baseline) to $25.0B (current). This is a material increase in short-term borrowing capacity, though no commercial paper is outstanding in either period.

Unrecognized tax benefits increased from $2.5B (Q1 FY26) to $4.5B (Q1 FY27), an 80% increase. The current filing separately discloses $439M in related interest and penalties. This reflects the growth in the company's tax reserves as income scales.

The current filing discloses an ongoing IRS examination for fiscal years 2023 and 2024. The baseline contained no such disclosure. This is a new tax-examination disclosure.

NVIDIA now holds $30.2B in publicly-held equity securities (marketable + other assets), with $27.4B subject to short-term lock-up restrictions and $8.9B locked through December 2027. The baseline filing reported only $1.3B in publicly-held equity securities (CoreWeave reclassification). This represents a massive expansion in equity-method and strategic equity investments, concentrated in companies whose shares NVIDIA cannot immediately liquidate. Net unrealized gains on these holdings were $13.4B in Q1 FY27 vs. $222M unrealized losses in Q1 FY26.

Non-marketable equity securities (privately-held companies) grew from $3.2B (baseline Q1 FY26) to $42.3B (current Q1 FY27), with $5.3B cumulative unrealized gains. NVIDIA added $1.0B in infrastructure-fund equity-method investments (max exposure $2.3B) and disclosed $27B in future investment commitments through remainder of FY27. This reflects a strategic pivot toward venture-style equity stakes in AI infrastructure, data-center operators, and private AI companies—illiquid positions that tie NVIDIA's balance sheet to the success of its ecosystem partners.

NVIDIA disclosed $119B in manufacturing, supply, and capacity commitments as of April 26, 2026, with $95B due in the remainder of FY27 and the balance through FY2031. The baseline filing reported $29.8B in inventory purchase and long-term supply obligations. This fourfold increase reflects data-center-scale production and longer ordering horizons across current (Hopper) and future (Blackwell, next-gen) architectures. The commitments are partially cancellable or adjustable, but changes may result in additional costs.

NVIDIA committed to $30B in multi-year cloud service agreements (up from $10.6B in baseline), with $6B due in remainder of FY27 and the balance through FY2032+. The agreements are primarily for R&D (not DGX Cloud resale). This reflects NVIDIA's need for massive compute capacity to train and test next-generation AI models, and signals that the company is locking in cloud infrastructure at scale—potentially to avoid spot-market pricing volatility or capacity constraints.

NVIDIA guaranteed up to $3.5B of partners' facility lease obligations (5-7 year terms) in exchange for warrants, with $712M in escrow to mitigate exposure. The guarantees are classified as credit derivatives with fair-value changes in Other income (expense). This structure suggests NVIDIA is backstopping data-center or manufacturing facility leases for ecosystem partners (likely ODMs, CSPs, or AI infrastructure startups) to secure capacity or strategic alignment, taking on credit risk in exchange for equity upside.

NVIDIA expects to commence leases with $32.4B in future obligations (Q2 FY27 through FY2033), primarily for data-center leases supporting R&D, with 3-20 year terms. The baseline filing disclosed $7.4B in expected future lease obligations (Q2 FY26 through FY2030). This 4.4x increase reflects NVIDIA's buildout of internal AI training and inference infrastructure at unprecedented scale, and signals that the company is committing to long-term data-center capacity to support next-generation model development.

Inventory provisions declined from $2.3B in Q1 FY26 (including $1.9B for H20 product, part of a $4.5B total H20 charge) to $0.8B in Q1 FY27. The current filing does not break out product-specific provisions. This suggests the H20 inventory issue (China-market export-controlled product) has been substantially resolved, and NVIDIA is now provisioning at a more normalized run-rate. The absence of a similar product-specific callout in the current period is a positive signal on inventory management.

Excess inventory purchase obligation charges declined from $3.0B in Q1 FY26 (including $2.6B for H20 product orders) to $0.3B in Q1 FY27. The liability balance increased from $4.3B (Apr 27, 2025) to $3.1B (Apr 26, 2026), reflecting utilization of prior-period accruals. This confirms that the H20 excess-order issue (part of the $4.5B total H20 charge) has been substantially worked through, and NVIDIA is now accruing for excess purchase obligations at a much lower run-rate.

Product warranty liabilities increased from $2.1B (Apr 27, 2025) to $2.9B (Apr 26, 2026), with additions of $330M in Q1 FY27 vs. $870M in Q1 FY26. The sequential increase (Jan 25, 2026: $2.8B → Apr 26, 2026: $2.9B) is modest, but the year-over-year comparison shows a 41% increase in the liability balance. This reflects higher warranty exposure on the growing installed base of Compute & Networking products (Hopper, Blackwell), and may signal elevated field-failure rates or extended warranty terms on data-center-scale deployments.

Customer advances declined sharply: Q1 FY27 added $1.7B (vs. $6.2B in Q1 FY26) and recognized $1.6B (vs. $6.0B in Q1 FY26). The ending balance of customer advances was $297M (Apr 26, 2026) vs. $287M (Apr 27, 2025). This suggests NVIDIA is no longer requiring large upfront deposits from customers—likely because supply constraints have eased and the company has greater confidence in its ability to fulfill orders without pre-funding. The normalization of customer-advance activity is a sign of a healthier, less supply-constrained market.

The district court granted class certification on March 25, 2026, certifying a class of investors who purchased NVIDIA stock between August 10, 2017 and November 15, 2018. NVIDIA filed a Ninth Circuit petition to appeal the certification order on April 8, 2026 under Rule 23(f). Class certification is a procedural milestone that significantly increases litigation risk and potential settlement pressure, as it aggregates claims from all class members and makes the case more expensive to defend. The Ninth Circuit may or may not grant permission to appeal.

NVIDIA changed its revenue-by-market presentation in Q1 FY27, replacing the prior "End Market" breakdown (Data Center split into Compute/Networking, plus Gaming, ProViz, Auto, OEM) with a new "Market Platform" breakdown (Data Center split into Hyperscale vs. AI Clouds/Industrial/Enterprise, plus Edge Computing). The new presentation aggregates Gaming, ProViz, Auto, and OEM into "Edge Computing" ($6.4B) and splits Data Center by customer type (Hyperscale CSPs vs. AI Clouds/enterprises). This reflects NVIDIA's strategic focus on data-center customer segmentation and de-emphasizes the legacy Gaming/ProViz businesses.

Stock-based compensation expense increased from $1.5B (Q1 FY26) to $1.9B (Q1 FY27), and unearned SBC expense increased from $15.3B to $19.3B, with a weighted average recognition period of 2.6 years (up from 2.3 years). This reflects higher grant-date fair values (weighted average $181.73 per share for Q1 FY27 grants vs. $106.43 for Q1 FY26 grants) and a larger outstanding equity award base. The increase in unearned SBC signals continued dilution pressure, though the company is offsetting this with share repurchases.

NVDA added a new standalone competition risk factor. It describes competitive pressures from companies with fabrication facilities, longer histories, and larger IP portfolios, as well as customers developing their own ASICs and cloud-based AI services that may not require NVDA's full feature set. The disclosure highlights the risk that NVDA may be unable to establish sufficient market share in AI cloud services to meet business objectives.

NVDA added a new paragraph describing infrastructure constraints (data centers, energy, capital) that could limit customer deployments. The disclosure emphasizes that energy expansion is a multi-year process with regulatory and construction challenges, and that less-capitalized companies may struggle to secure financing. This is a new articulation of demand-side constraints beyond NVDA's direct supply chain.

NVDA changed "plan to increase" to "are increasing" (present tense), indicating the U.S. manufacturing expansion is now underway. The current filing also adds a new sentence acknowledging that NVDA "may experience delays or difficulties in scaling production as planned," a more explicit risk disclosure than the baseline's conditional framing.

NVDA reframed "product transitions" as "introducing or offering multiple architectures concurrently" and narrowed the scope from "Data Center, Gaming, Professional Visualization, and Automotive" to "Data Center and Edge Computing." The current filing also adds a specific forward-looking reference to the Rubin platform, expected to ship in the second half of fiscal 2027, providing a concrete roadmap milestone.

NVDA updated the China antitrust disclosure to reflect a specific date (September 15, 2025) and a "preliminary finding" that NVDA's compliance with U.S. export controls violated the terms of the Mellanox acquisition approval. The baseline described an ongoing "investigation" into whether compliance "discriminates unfairly"; the current filing reports that regulators have now "published their preliminary finding" that it does. This is a progression from inquiry to formal preliminary adverse finding.

This is a new development since the baseline filing and introduces a potential revenue-sharing obligation with the government.

NVDA added disclosure about a new H200 licensing program that began in February 2026. The program requires U.S. inspection before shipment, which subjects H200 products to a 25% tariff upon U.S. importation. NVDA has not yet generated revenue under this program and does not know if imports into China will be allowed. The disclosure warns that NVDA may be unable to pass the tariff cost to customers, potentially harming margins and competitive position.

NVDA changed the framing from "may be unable" (future conditional) to "were effectively foreclosed" (past/present fact) as of Q1 FY2027. The current filing adds that this foreclosure has already helped competitors build larger ecosystems to challenge NVDA worldwide, and that the lost opportunity will have a material adverse impact unless NVDA can return with a product approved by both governments. This is a shift from hypothetical risk to disclosed current status.

NVDA updated the AI Diffusion IFR discussion to add GB300 to the list of affected products and to remove the technical ECCN classification detail. More significantly, the current filing adds two new legislative developments: the GAIN AI Act (passed by the Senate in October 2025, which would restrict the Trump Administration's ability to adapt export controls and allow private parties to review licensing decisions) and the Remote Access Security Act (RASA, under consideration, which could prohibit cloud services to companies with Chinese ultimate parents). These are new potential restrictions not mentioned in the baseline.

NVDA added a specific date reference ("new state laws that took effect on January 1, 2026") to the state AI regulation disclosure, indicating that some of the previously-anticipated regulations have now taken effect. This is a progression from future conditional to current compliance obligation.

NVDA added a new paragraph describing how draft U.S. legislation calling for mandatory chip features prompted China's government to publicly question whether H20 products have built-in vulnerabilities, discouraging customer purchases. NVDA states it provided a public response clarifying that its GPUs do not include such vulnerabilities. This is a new disclosure of a specific reputational and demand impact from legislative proposals, even before enactment.

NVDA added a sentence emphasizing that demand for open-source foundation models "promotes use of our products worldwide," framing open-source as a positive demand driver. The current filing also adds "KIMMI" to the list of Chinese foundation models (alongside DeepSeek and Qwen). The addition of the positive framing makes the subsequent risk of regulatory restriction more material, as it would cut off a growing demand source.

NVDA added a conditional framing to the Chinese Action Plan disclosure: "Although we are already effectively foreclosed from the China market by U.S. export controls, if those controls changed to allow us to return to the market, the Chinese government could modify or implement the Action Plan..." This acknowledges that the Action Plan risk is currently moot due to U.S. export controls, but would become material if NVDA regains access to the China data center market.

NVDA expanded the compliance disclosure to add that it "has provided and will continue to provide assistance to authorities regarding attempted diversion," and that it must rely on customer/partner compliance programs since it lacks physical control post-sale. The current filing also adds that "reports of diversion of controlled products, even when unsubstantiated and untrue, or any compliance failure at a customer or partner" may harm NVDA, and adds a new sentence about "incorrect allegations that legitimate and appropriate business is using supposed 'loopholes' in the export controls." This is a more defensive framing, anticipating criticism of NVDA's compliance approach.