August 26, 2026
Nvidia earnings test valuations as custom chips alter hardware exposure
NVDA, AAPL, PSKY, VEEV, AMD, AVGO, TSM, SPCX, AMZN

Executive Summary
Rebound Driven: Nvidia led semiconductor gains ahead of earnings as compressed valuations spurred market recovery.
Silicon Shift: Custom chips from OpenAI and Apple highlight an industry move toward specialized inference computing.
Private Capital: Funding rounds and upcoming IPOs reveal strong demand across AI infrastructure, defense, and wearables.
Economic Risks: Retaliatory Canadian tariffs and AI debt maturities introduce near-term macro headwinds.
Corporate Dynamics
NVDA (Nvidia): Compressed Valuations and Networking Growth Frame Earnings Expectations
News: Snapped a 7-session losing streak—recovering from a $400 billion market cap loss—gaining 2% during the session ahead of quarterly earnings. Its forward price-to-earnings ratio compressed to 18–19x (roughly 16x 2028 earnings), aligning its valuation with the S&P 500. Expected quarterly revenue ranges from $92 billion to nearly $100 billion (doubling revenue and adjusted earnings year-over-year), with top-line whisper numbers at $107 billion–$108 billion, quarterly free cash flow at $40 billion–$50 billion (finishing fiscal year 2026 with nearly $100 billion in free cash flow), and $140 billion in commitments. Its data center compute segment generates ~$70 billion quarterly (77% year-over-year growth), while networking grows ~200% year-over-year. To offset memory costs and protect ~75% gross margins amid a Vera Rubin architecture ramp, reports indicate the company implemented 15% price increases. Nvidia is also introducing a reporting split between hyperscalers and its ACIE group (AI clouds, industrial, enterprise) to show customer diversification, while China revenue expectations remain zeroed out (restricted to nominal H200 sales). Despite beating guidance targets by 2.5% to 5.6% historically, the stock closed lower following four of its last five earnings reports. Investor focus remains on its circular financing ecosystem (compared by Franklin Templeton's Sara Araghi to John Malone’s 1980s TCI model) and its $250 billion to $500 billion float capacity used to fund customer buildouts.
Outlook: Absorbing memory cost inflation via price increases defends industry-leading gross margins, while hyper-growth in networking and customer reporting splits help offset projected AI market share losses (expected to drop from 90% to 70%) to custom ASICs.
OpenAI: Proprietary Silicon Entry Targets Inference Cost Reductions
News: Unveiled its first custom inference chip, "Jalapeno," featuring an integrated HBM4 and SRAM architecture designed to eliminate data movement. Benchmarked via the Inference X framework against Nvidia's GB300 system, it demonstrated 1.8x to 4x higher performance-per-watt, higher throughput, and lower latency. Designed in-house with Broadcom handling physical design/TSMC manufacturing masks and Celestica handling rack system co-design, VP of Hardware Richard Ho noted the chip specifically targets inference to lower token costs for weekly active users while relying on Nvidia for training. Development is already underway for second and third-generation chips.
Outlook: Deploying custom inference hardware lowers serving costs for rapidly expanding user bases and diminishes operational reliance on third-party silicon suppliers.
Get instant access to AI-driven US tech market research.
Executive Summary
Rebound Driven: Nvidia led semiconductor gains ahead of earnings as compressed valuations spurred market recovery.
Silicon Shift: Custom chips from OpenAI and Apple highlight an industry move toward specialized inference computing.
Private Capital: Funding rounds and upcoming IPOs reveal strong demand across AI infrastructure, defense, and wearables.
Economic Risks: Retaliatory Canadian tariffs and AI debt maturities introduce near-term macro headwinds.
Corporate Dynamics
NVDA (Nvidia): Compressed Valuations and Networking Growth Frame Earnings Expectations
News: Snapped a 7-session losing streak—recovering from a $400 billion market cap loss—gaining 2% during the session ahead of quarterly earnings. Its forward price-to-earnings ratio compressed to 18–19x (roughly 16x 2028 earnings), aligning its valuation with the S&P 500. Expected quarterly revenue ranges from $92 billion to nearly $100 billion (doubling revenue and adjusted earnings year-over-year), with top-line whisper numbers at $107 billion–$108 billion, quarterly free cash flow at $40 billion–$50 billion (finishing fiscal year 2026 with nearly $100 billion in free cash flow), and $140 billion in commitments. Its data center compute segment generates ~$70 billion quarterly (77% year-over-year growth), while networking grows ~200% year-over-year. To offset memory costs and protect ~75% gross margins amid a Vera Rubin architecture ramp, reports indicate the company implemented 15% price increases. Nvidia is also introducing a reporting split between hyperscalers and its ACIE group (AI clouds, industrial, enterprise) to show customer diversification, while China revenue expectations remain zeroed out (restricted to nominal H200 sales). Despite beating guidance targets by 2.5% to 5.6% historically, the stock closed lower following four of its last five earnings reports. Investor focus remains on its circular financing ecosystem (compared by Franklin Templeton's Sara Araghi to John Malone’s 1980s TCI model) and its $250 billion to $500 billion float capacity used to fund customer buildouts.
Outlook: Absorbing memory cost inflation via price increases defends industry-leading gross margins, while hyper-growth in networking and customer reporting splits help offset projected AI market share losses (expected to drop from 90% to 70%) to custom ASICs.
OpenAI: Proprietary Silicon Entry Targets Inference Cost Reductions
News: Unveiled its first custom inference chip, "Jalapeno," featuring an integrated HBM4 and SRAM architecture designed to eliminate data movement. Benchmarked via the Inference X framework against Nvidia's GB300 system, it demonstrated 1.8x to 4x higher performance-per-watt, higher throughput, and lower latency. Designed in-house with Broadcom handling physical design/TSMC manufacturing masks and Celestica handling rack system co-design, VP of Hardware Richard Ho noted the chip specifically targets inference to lower token costs for weekly active users while relying on Nvidia for training. Development is already underway for second and third-generation chips.
Outlook: Deploying custom inference hardware lowers serving costs for rapidly expanding user bases and diminishes operational reliance on third-party silicon suppliers.
AAPL (Apple): Desktop Silicon Overhaul Addresses On-Device AI Demand
News: Announced desktop silicon upgrades for the Mac Mini and Mac Studio, available September 22. The Mac Mini debuts the M6 chip (with M5 Pro options), raising base pricing from $600 to $900. The Mac Studio features the M5 Max and new top-tier M5 Ultra chip (offering up to 80 CPU cores and half a terabyte of memory capacity), starting at approximately $5,500. Upgrades were driven by demand for local AI processing alongside industry-wide memory and silicon shortages.
Outlook: Shifting product lineups toward higher-tier processing configurations captures premium demand for local AI capabilities while preserving profit margins against rising component supply costs.
Oura: Capitalizing on Health-Tech Demand via Public Listing
News: Preparing an initial public offering as early as September to raise up to $3 billion, targeting a market valuation exceeding $16 billion.
Outlook: Transitioning to public markets provides substantial liquidity to accelerate hardware development and scale consumer health platform footprint.
Lambda: Funding Round Lays Groundwork for Infrastructure Expansion
News: In discussions to raise up to $3 billion in a private funding round valuing the neocloud provider at $12 billion or more, setting up a potential IPO next year.
Outlook: Accumulating private capital fuels physical data center expansions required to meet surging enterprise compute demand ahead of a public listing.
Gatik: Contracted Backlog Secures Capital-Light Autonomous Logistics
News: Closed a $200 million funding round led by Qatar Investment Authority (QIA) and Koch Disruptive Technologies. Holds over $600 million in contracted backlog via multi-year, asset-light take-or-pay agreements with partners including PepsiCo, Isuzu, and Nvidia.
Outlook: Securing long-term take-or-pay contracts ensures multi-year revenue visibility while an asset-light model shields balance sheets from heavy fleet capital expenditures.
Alice: Capital Injection Focuses on Autonomous AI Safety
News: Closed a $140 million funding round led by Apax Digital to expand research into model safety, stress-testing, and preventing autonomous agent breaches.
Outlook: Enterprise deployment of autonomous agents creates direct demand for specialized security validation, positioning the firm to capture early cybersecurity spend.
Waymo: International Fleet Expansion Target European Urban Centers
News: Announced plans to expand its robotaxi network into Munich, Germany by the end of 2027, marking its third international market following initial manual mapping runs.
Outlook: Geographic expansion into international markets increases total addressable ride-hailing revenues while spreading platform development overhead across broader municipal networks.
SpaceX (SPCX): Satellite Data Center Roadmap Extends Infrastructure Reach
News: Shares rose 2% as Elon Musk updated timelines for satellite-based data centers, stating SpaceX’s first AI satellites equipped with Nvidia chips will launch at the end of 2027 and reach significant scale by 2028, following a successful IPO market window.
Outlook: Deploying orbital compute infrastructure establishes a differentiated cloud hosting layer that captures high-density enterprise workload demand.
PSKY (Paramount): Regulatory Friction Pauses Media Consolidation
News: Planned acquisition of Warner Bros. stalled as regulatory talks with the State of California are on hold. California Attorney General Rob Bonta cited unauthorized leaks and rejected Paramount's proposed behavioral commitments, demanding structural remedies across theatrical, blockbuster, and cable distribution markets. Separately, AG Bonta confirmed Meta CEO Mark Zuckerberg is on the state's witness list in an ongoing trial against Meta.
Outlook: Stricter regulatory mandates requiring structural asset sales increase transaction complexity and delay cost-synergy realization across legacy media portfolios.
Ursa Major: SPAC Transaction Scales Defense Propulsion Capacity
News: Agreed to go public via a SPAC transaction at a post-deal valuation of $2.3 billion, targeting Q1 completion. Proceeds will expand its Colorado facility from 10 to 500 hypersonic units per year, fund a 400-acre solid rocket motor energetics campus, and scale its Havoc missile system, with management projecting positive cash flow by the end of 2027.
Outlook: Transforming manufacturing throughput 50-fold enables the company to convert defense backlog into high-volume delivery, supporting its goal of positive cash flow by 2027.
LEGO Group: Interactive Hardware Integrations Expand Consumer Licensing
News: CEO Niels B. Christiansen outlined the expansion of its interactive "Smart Play" platform following launches across six markets with Star Wars sets. The company integrated its interactive "Smart Brick" into its Pokémon line (featuring an interactive Pikachu) with international rollouts planned for next year.
Outlook: Layering interactive hardware onto physical build sets increases average selling prices and opens high-margin cross-brand licensing opportunities.
VEEV (Veeva Systems): Vertical AI Software Demonstrates Defensive Operating Metrics
News: Highlighted by Baird tech strategist Ted Mortonson as a top risk-reward candidate in healthcare software due to its positioning in early vertical AI adoption.
Outlook: Maintaining 40%+ operating margins and strong free cash flow shields the business from macro volatility while reinforcing its market position in healthcare software.
AMD (Advanced Micro Devices): Market Share Gains Support Semiconductor Surge
News: Shares surged 5%, leading semiconductor sector gains alongside Lumentum (+4% to +6%), Marvell (+4%), Western Digital (+3.5%), Lam Research (+1%), and TSMC (+1%).
Outlook: Capturing merchant market share in data center processors diversifies revenue streams as cloud hyperscalers seek alternative silicon vendors.
Software Sector Peers: Divergent Rebounds Highlight Selective Institutional Flows
News: Oracle rose nearly 2% (1.5% in later trading), Microsoft gained 0.5% to 1%, and Cisco rose 0.8%, alongside Ark Innovation ETF gains in Robinhood (+8%), Beam (+8%), DNA (+10%), and CRISPR (+6%). Conversely, CrowdStrike, Intuit, and Palo Alto Networks declined by over 3%, with Zoom, Thomson Reuters, and Zscaler also trading lower despite a broader software rebound off mid-July lows.
Outlook: Institutional capital is selectively rotating into high-margin "systems of record" while penalizing high-valuation software providers facing extended sales cycles.
Industry Trends
Custom ASIC Adoption and Hardware Architectural Shifts
Analysis: Driven by high chip costs and supply constraints, cloud hyperscalers and AI labs are shifting workloads toward proprietary custom ASICs (such as OpenAI's Jalapeno, GOOGL (Google)'s TPU, and AMZN (Amazon)'s Trainium). Concurrently, data center physical infrastructure is undergoing two major architectural pivots: converting copper interconnects to co-packaged optics (CPO) and stepping up power management architectures from 40-volt to 800-volt systems to handle extreme thermal loads.
Outlook: Market share for merchant AI data center chips is projected to shift from 90% down to approximately 70% over the next two years. This transition expands addressable markets for specialized optical and power component suppliers like AVGO (Broadcom), ADI (Analog Devices), and MPWR (Monolithic Power Systems).
Hyperscaler Capital Spending vs. Infrastructure Credit Risks
Analysis: Hyperscaler capital expenditures are projected to climb from approximately $800 billion this year to $1.1 trillion in 2027 to establish foundational agentic AI infrastructure. However, software valuation multiples compressed significantly relative to semiconductors—at peak divergence, the semiconductor index (SOX) was up 90%–100% while the software index (IGV) was down 30%, before normalizing to +61% and flat, respectively. Sycamore Tree Capital Partners CIO Trey Parker warned that AI data center credit markets mirror late-1990s telecom debt maturity risks, where debt schedules threatened to outpace monetization timelines.
Outlook: High capex cycles heighten credit risks for debt-financed infrastructure buildouts, accelerating institutional preference toward software platforms that demonstrate immediate, high-margin enterprise adoption.
Cross-Border Trade Escalation and Supply Chain Protections
Analysis: Trade disputes are driving structural cost increases across consumer hardware and industrial manufacturing. Canadian Prime Minister Mark Carney announced dollar-for-dollar retaliatory tariffs on $20 billion worth of US goods taking effect September 8. The structure doubles existing tariffs on US steel and aluminum to 50%, applies a new 50% tariff to US furniture, clothing, and electronics, and imposes a 25% tariff on appliances, dairy, fish, and metal derivatives. Concurrently, US export controls have zeroed out China financial expectations for top-tier chipmakers.
Outlook: Retaliatory tariffs and export limits inflate hardware component costs, compress international profit margins, and force domestic governments to issue dedicated aid packages to support affected industries.
Market Sentiment
Pre-Earnings Skepticism vs. Fundamental Multiples
Analysis: Despite Nvidia snapping its 7-session losing streak ($400 billion valuation drop) and displaying compressed forward P/E multiples of 18–19x (and ~16x 2028 earnings), price action reflects investor nervousness. Historical data shows the stock fell post-earnings in four of the last five quarters despite exceeding revenue guidance by 2.5% to 5.6%. Market focus has shifted heavily toward whisper numbers ($107B–$108B), circular financing structures (such as consortium investments), and gross margin headwinds from memory price increases.
Outlook: Investors are prioritizing sustainable cash-flow metrics, margin preservation, and clear customer diversification over simple top-line earnings beats.
Enterprise Cybersecurity Premiums vs. Pipeline Conversion Delays
Analysis: Emerging AI threat models (such as Mythos) that actively probe hardware and router architectures have elevated demand for foundational security providers like NET (Cloudflare), CRWD (CrowdStrike), and PANW (Palo Alto Networks). However, market sentiment remains cautious due to high sector valuations (trading near 30x EV/Sales) and insider selling. Strategist Ted Mortonson highlighted that architectural security shifts impact forward sales pipelines rapidly, but create uncertainty regarding how quickly pipelines convert into Annual Recurring Revenue (ARR).
Outlook: High valuation multiples render cybersecurity equities vulnerable to pullbacks if pipeline opportunities face extended enterprise procurement cycles.
Private Funding Rebound vs. Truncated IPO Windows
Analysis: Large private funding rounds (Oura's $3B IPO target, Lambda's $3B round, Gatik's $200M round, and Alice's $140M round) indicate healthy private equity liquidity. However, market experts point out that public market liquidity remains constrained to a narrow, truncated 6 to 7-week fall IPO window post-Labor Day ahead of US midterms.
Outlook: Capital flows will heavily favor mature, midcap-to-megacap issuers ($10 billion to $20 billion market caps) with clear paths to cash flow positivity, while higher-risk growth issuers face tighter public market conditions.
Important Disclaimer
This newsletter is for informational purposes only and does not constitute investment advice. Content is generated by AI and may contain inaccuracies; always verify data independently before trading. Investing involves significant risk of loss. AlchemyJ is not a registered financial advisor. By reading this, you agree to our terms.
