{"id":13659,"date":"2026-07-17T16:00:00","date_gmt":"2026-07-17T08:00:00","guid":{"rendered":"https:\/\/ai-stack.ai\/?p=13659"},"modified":"2026-07-14T16:38:41","modified_gmt":"2026-07-14T08:38:41","slug":"jensen-huang-nvidia-morgan-stanley-roadshow-2026","status":"publish","type":"post","link":"https:\/\/ai-stack.ai\/en\/jensen-huang-nvidia-morgan-stanley-roadshow-2026","title":{"rendered":"Jensen Huang at Morgan Stanley: NVIDIA Nears $100B Quarter, Denies Rubin Ultra Delay, and Reveals the ASIC Reversal"},"content":{"rendered":"<style>table{border-collapse:collapse;width:100%;margin:1em 0}th,td{border:1px solid #ddd;padding:8px 12px;text-align:left}th{background-color:#f5f5f5;font-weight:bold}tr:nth-child(even){background-color:#fafafa}<\/style>\n<h1 id=\"jensen-huang-at-morgan-stanley-nvidia-nears-100b-quarter-denies-rubin-ultra-delay-and-reveals-the-asic-reversal\"><strong>Jensen Huang at Morgan Stanley: NVIDIA Nears $100B Quarter, Denies Rubin Ultra Delay, and Reveals the ASIC Reversal<\/strong><\/h1>\n<p>In early July 2026, NVIDIA CEO Jensen Huang, alongside CFO Colette Kress, appeared at a Morgan Stanley Non-Deal Roadshow (NDR) in California. Before an audience of institutional investors, Huang delivered a message of striking confidence: <strong>NVIDIA\u2019s quarterly revenue is approaching $100 billion \u2014 and growth is still accelerating.<\/strong><\/p>\n<p>The timing was critical. Over the preceding month, NVIDIA had faced a barrage of market skepticism on three fronts: rumors that the next-generation Rubin Ultra chip would be delayed to 2028, fears that custom ASICs were systematically eating into GPU market share, and mounting questions about whether the AI capex cycle had peaked.<\/p>\n<p>Huang addressed all three head-on with data, technical specifics, and an uncharacteristically combative tone. Morgan Stanley emerged from the meeting maintaining its <strong>Overweight<\/strong> rating and <strong>$288 price target<\/strong> \u2014 representing approximately 42% upside \u2014 while reiterating NVIDIA as its top semiconductor pick.<\/p>\n<p>Here is what was said, what it means, and why it matters for enterprise AI strategy.<\/p>\n<h2 id=\"the-100-billion-quarter-inside-nvidias-accelerating-growth-curve\"><strong>1. The $100 Billion Quarter: Inside NVIDIA\u2019s Accelerating Growth Curve<\/strong><\/h2>\n<p>NVIDIA is approaching a milestone that even Huang described as \u201chard to fathom\u201d: a $100 billion quarter.<\/p>\n<p>The most recent fiscal data (Q1 FY2027, reported May 2026) paints the picture:<\/p>\n<table>\n<thead>\n<tr>\n<th>Metric<\/th>\n<th>Q1 FY2027<\/th>\n<th>YoY Change<\/th>\n<\/tr>\n<\/thead>\n<tbody>\n<tr>\n<td>Total Revenue<\/td>\n<td><strong>$81.6B<\/strong><\/td>\n<td>\u2191 85%<\/td>\n<\/tr>\n<tr>\n<td>Data Center Revenue<\/td>\n<td><strong>$75.2B<\/strong><\/td>\n<td>\u2191 92%<\/td>\n<\/tr>\n<tr>\n<td>Networking Revenue<\/td>\n<td><strong>$14.8B<\/strong><\/td>\n<td>\u2191 199%<\/td>\n<\/tr>\n<tr>\n<td>Net Income<\/td>\n<td><strong>$58.3B<\/strong><\/td>\n<td>3\u00d7<\/td>\n<\/tr>\n<tr>\n<td>GAAP Gross Margin<\/td>\n<td><strong>74.9%<\/strong><\/td>\n<td>from 60.8%<\/td>\n<\/tr>\n<tr>\n<td>Free Cash Flow<\/td>\n<td><strong>$48.6B<\/strong><\/td>\n<td>\u2191 85%<\/td>\n<\/tr>\n<\/tbody>\n<\/table>\n<p>Q2 guidance stands at approximately <strong>$91 billion<\/strong> \u2014 notably excluding any Data Center sales to China, suggesting potential upside. The $100 billion quarterly threshold is no longer a question of \u201cif\u201d but \u201cwhen.\u201d<\/p>\n<p>Morgan Stanley analyst Joseph Moore provided longer-range projections:<\/p>\n<table>\n<thead>\n<tr>\n<th>Fiscal Year<\/th>\n<th>Estimated Revenue<\/th>\n<th>YoY Growth<\/th>\n<\/tr>\n<\/thead>\n<tbody>\n<tr>\n<td>FY2026<\/td>\n<td><strong>$215.9B<\/strong><\/td>\n<td>+82%<\/td>\n<\/tr>\n<tr>\n<td>FY2027<\/td>\n<td><strong>$393.0B<\/strong><\/td>\n<td>+52.4%<\/td>\n<\/tr>\n<tr>\n<td>FY2028<\/td>\n<td><strong>$598.8B<\/strong><\/td>\n<td>+52.4%<\/td>\n<\/tr>\n<tr>\n<td>FY2029<\/td>\n<td><strong>$783.9B<\/strong><\/td>\n<td>+30.9%<\/td>\n<\/tr>\n<\/tbody>\n<\/table>\n<p>Calendar 2027 EPS is estimated at <strong>$13.08<\/strong>. At 22\u00d7 earnings, the $288 price target requires no valuation multiple expansion \u2014 pure earnings-driven upside. For enterprises evaluating GPU investments, \ud83d\udd17 <a href=\"https:\/\/ai-stack.ai\/en\/gpu-roi\"><u>understanding NVIDIA\u2019s pricing power and margin structure<\/u><\/a> is essential to forecasting your own AI infrastructure costs.<\/p>\n<h2 id=\"rubin-ultra-on-track-the-technical-truth-behind-the-delay-rumors\"><strong>2. Rubin Ultra On Track: The Technical Truth Behind the Delay Rumors<\/strong><\/h2>\n<p>In late June, SemiAnalysis published a report that shook NVIDIA\u2019s stock: the original 4-die Rubin Ultra packaging was reportedly canceled due to TSMC CoWoS-L substrate warping and signal integrity failures, forcing a downgrade to a 2-die design, with shipments potentially slipping to <strong>2028<\/strong>.<\/p>\n<p>At the Morgan Stanley roadshow, Huang was unequivocal: <strong>\u201cRubin Ultra will ship next year, on schedule.\u201d<\/strong><\/p>\n<p>He explained that the Kyber rack system is being replaced with an improved architecture designed to support larger compute domains \u2014 but framed this as routine system-level optimization, not a product delay. Three facts support this timeline:<\/p>\n<ol type=\"1\">\n<li><strong>800V high-voltage power delivery<\/strong> is progressing on schedule \u2014 a critical enabler for rack-scale Rubin Ultra deployments<\/li>\n<li><strong>Inter-rack optical interconnects<\/strong> are replacing copper, enabling million-GPU clusters<\/li>\n<li><strong>TSMC CoWoS capacity<\/strong> is expanding ~80% year-over-year, providing the physical foundation for mass production<\/li>\n<\/ol>\n<p>\ud83d\udd17 <a href=\"https:\/\/ai-stack.ai\/en\/gtc-2026-nemoclaw\"><u>NVIDIA\u2019s GTC 2026 Rubin architecture reveal<\/u><\/a> and \ud83d\udd17 <a href=\"https:\/\/ai-stack.ai\/en\/blackwell-vs-mi300x\"><u>Blackwell vs.\u00a0MI300X performance benchmarks<\/u><\/a> provide the technical context for understanding Rubin Ultra\u2019s generational leap.<\/p>\n<h2 id=\"the-asic-reversal-when-custom-chip-customers-come-back-to-gpu\"><strong>3. The ASIC Reversal: When Custom Chip Customers Come Back to GPU<\/strong><\/h2>\n<p>The most explosive disclosure of the roadshow concerned ASIC competition. Huang revealed that <strong>a frontier AI lab previously running its flagship models almost entirely on custom ASICs \u2014 widely believed to be Anthropic \u2014 has shifted nearly 50% of its compute back to NVIDIA GPUs.<\/strong><\/p>\n<p>This single data point upends the prevailing narrative that custom silicon is on an inexorable march to replace general-purpose GPUs. The ASIC camp\u2019s most prominent customer is buying GPUs again \u2014 and in large quantities.<\/p>\n<p>Huang\u2019s explanation: \u201cThe customer\u2019s evaluation criteria isn\u2019t chip unit price \u2014 it\u2019s <strong>total cost per token.<\/strong>\u201d At deployment scale, NVIDIA\u2019s ecosystem maturity (CUDA, cuDNN, TensorRT), supply chain reliability (TSMC CoWoS capacity), and developer efficiency (millions of CUDA engineers) translate into a genuine total-cost advantage.<\/p>\n<p>However, the ASIC threat is not imaginary. The data tells a nuanced story:<\/p>\n<table>\n<thead>\n<tr>\n<th>Year<\/th>\n<th>NVIDIA AI Accelerator Share<\/th>\n<th>Total Market Size<\/th>\n<\/tr>\n<\/thead>\n<tbody>\n<tr>\n<td>2024<\/td>\n<td><strong>87% (peak)<\/strong><\/td>\n<td>~$115B<\/td>\n<\/tr>\n<tr>\n<td>2025<\/td>\n<td>81%<\/td>\n<td>~$160B<\/td>\n<\/tr>\n<tr>\n<td>2026E<\/td>\n<td><strong>75%<\/strong><\/td>\n<td>$200B+<\/td>\n<\/tr>\n<\/tbody>\n<\/table>\n<p>\ud83d\udd17 <a href=\"https:\/\/ai-stack.ai\/en\/asic-vs-gpu\"><u>ASIC vs.\u00a0GPU: the technology battle<\/u><\/a> is fundamentally about the bifurcation of the AI compute market into training and inference. NVIDIA\u2019s training dominance (90%+ share) remains nearly unassailable, but inference is where Google TPUs, Amazon Trainium, Microsoft Maia, and Broadcom custom ASICs are gaining ground.<\/p>\n<p>\ud83d\udd17 <a href=\"https:\/\/ai-stack.ai\/en\/google-tpu-vs-nvidia-gpu\"><u>Google TPU vs.\u00a0NVIDIA GPU: a complete comparison<\/u><\/a> illustrates why purpose-built silicon can win on cost in specific inference workloads \u2014 though Huang\u2019s point stands that when you factor in the full software stack and engineering overhead, GPU economics often still prevail.<\/p>\n<h2 id=\"three-growth-engines-mapping-the-global-demand-for-compute\"><strong>4. Three Growth Engines: Mapping the Global Demand for Compute<\/strong><\/h2>\n<p>Huang segmented NVIDIA\u2019s revenue into three distinct growth engines, a strategic framing that itself carries analytical weight:<\/p>\n<table>\n<colgroup>\n<col style=\"width: 25%\" \/>\n<col style=\"width: 43%\" \/>\n<col style=\"width: 31%\" \/>\n<\/colgroup>\n<thead>\n<tr>\n<th>Engine<\/th>\n<th>Revenue Share<\/th>\n<th>Dynamics<\/th>\n<\/tr>\n<\/thead>\n<tbody>\n<tr>\n<td><strong>AI Labs<\/strong><\/td>\n<td>~20%<\/td>\n<td>Frontier model builders (OpenAI, Anthropic); lumpy demand but massive individual orders; recent trend: ASIC-to-GPU reversion<\/td>\n<\/tr>\n<tr>\n<td><strong>Hyperscale Cloud<\/strong><\/td>\n<td>~50%<\/td>\n<td>Microsoft, Meta, Amazon, Google; steady expansion; growth extending from compute to networking and CPUs<\/td>\n<\/tr>\n<tr>\n<td><strong>Sovereign AI + Industrial AI + New AI Clouds<\/strong><\/td>\n<td>~30%<\/td>\n<td>Fastest-growing segment; least affected by ASIC substitution; driven by data localization, energy\/land constraints, and national AI autonomy<\/td>\n<\/tr>\n<\/tbody>\n<\/table>\n<p>The third engine \u2014 sovereign AI \u2014 deserves particular attention. Huang emphasized that governments are treating AI infrastructure as a national security asset, with an investment logic distinct from commercial cloud: <strong>\u201cThey\u2019re not optimizing for cost. They\u2019re optimizing for autonomy.\u201d<\/strong><\/p>\n<p>An underappreciated highlight: NVIDIA\u2019s <strong>CPU business<\/strong>. Huang projected ~<strong>$20 billion<\/strong> in CPU revenue this fiscal year, nearly half from standalone Vera CPU racks \u2014 NVIDIA\u2019s first serious push into the general-purpose server CPU market, directly challenging Intel Xeon and AMD EPYC.<\/p>\n<p>\ud83d\udd17 <a href=\"https:\/\/ai-stack.ai\/en\/gpu-npu-tpu-lpu\"><u>GPU, NPU, TPU, LPU: the complete 2026 AI processor landscape<\/u><\/a> explains why NVIDIA is expanding from GPUs into CPUs and DPUs \u2014 not diversification, but platform construction.<\/p>\n<h2 id=\"risks-and-challenges-why-the-market-stays-anxious\"><strong>5. Risks and Challenges: Why the Market Stays Anxious<\/strong><\/h2>\n<p>Despite Huang\u2019s confident presentation, Morgan Stanley\u2019s report enumerated four key risks:<\/p>\n<ol type=\"1\">\n<li><strong>Faster-than-expected hardware supply release<\/strong> \u2014 if TSMC CoWoS expansion outpaces demand or competitors (Samsung 2nm GAA) achieve yield breakthroughs, GPU oversupply could erode NVIDIA\u2019s pricing power<\/li>\n<li><strong>Dramatic reduction in AI R&amp;D costs<\/strong> \u2014 a DeepSeek-style architectural innovation that slashes training compute requirements would reduce per-unit demand<\/li>\n<li><strong>Competitor breakthrough products<\/strong> \u2014 AMD, Intel, or emerging ASIC vendors could deliver solutions significantly better than GPUs in specific use cases<\/li>\n<li><strong>Cloud customers becoming competitors<\/strong> \u2014 Google TPU is the most mature ASIC ecosystem; Meta\u2019s Iris chip enters mass production in September; the risk of major customers turning into rivals persists<\/li>\n<\/ol>\n<p>\ud83d\udd17 <a href=\"https:\/\/ai-stack.ai\/en\/meta-compute-ai-cloud\"><u>Meta Compute: Zuckerberg\u2019s $145B AI bet and the global semiconductor sell-off<\/u><\/a> demonstrated how \u201ccompute oversupply\u201d fears can vaporize hundreds of billions in market value in a single day \u2014 even when those fears prove unfounded.<\/p>\n<p>Yet each of these risks, paradoxically, underscores the depth of NVIDIA\u2019s moat. <strong>Every threat scenario presumes competitors first reaching technological, ecosystem, and scale parity with NVIDIA<\/strong> \u2014 which remains the hardest part of the equation.<\/p>\n<h2 id=\"conclusion-from-gpu-supplier-to-computing-platform\"><strong>Conclusion: From GPU Supplier to Computing Platform<\/strong><\/h2>\n<p>What Huang presented at Morgan Stanley was, at its surface, a rebuttal of three bear cases. But the deeper message was NVIDIA\u2019s evolution <strong>from a GPU supplier into an integrated computing platform spanning GPUs, CPUs, DPUs, and networking.<\/strong><\/p>\n<p>The strategic implication: when customers buy rack-scale systems \u2014 Vera CPU to Rubin Ultra GPU to Spectrum-X networking \u2014 the substitutability of any single component plummets. NVIDIA isn\u2019t selling chips; it\u2019s selling an <strong>integrated computing architecture that resists disaggregation.<\/strong><\/p>\n<p>For enterprises planning AI infrastructure, the roadshow\u2019s core takeaways are:<\/p>\n<ol type=\"1\">\n<li><strong>Compute supply is moving from scarcity to abundance, but frontier compute remains concentrated<\/strong> \u2014 procurement strategy must distinguish \u201ccommodity compute\u201d from \u201cfrontier compute\u201d<\/li>\n<li><strong>ASICs and GPUs are not a zero-sum game<\/strong> \u2014 both will coexist; choose based on workload (training vs.\u00a0inference)<\/li>\n<li><strong>NVIDIA\u2019s pricing power derives from the ecosystem, not the silicon<\/strong> \u2014 CUDA\u2019s network effects resist near-term replication<\/li>\n<li><strong>Sovereign AI and industrial AI represent the next wave<\/strong> \u2014 these segments have more fragmented customer bases and more diverse demand, implying a more stable revenue foundation for NVIDIA<\/li>\n<\/ol>\n<p>Huang\u2019s closing line captured the moment best: <strong>\u201cI don\u2019t know anyone in the industry who thinks they have too much compute.\u201d<\/strong><\/p>\n<p><em>This article was compiled by the INFINITIX team. Sources include <a href=\"https:\/\/in.investing.com\/news\/stock-market-news\/morgan-stanley-reiterates-overweight-on-nvidia-stock-citing-diversified-growth-93CH-5491461\" target=\"_blank\" rel=\"noopener\">Morgan Stanley NVIDIA Research Report<\/a>, <a href=\"https:\/\/dataconomy.com\/2026\/07\/13\/nvidia-ceo-quarterly-revenue-100-billion\/\" target=\"_blank\" rel=\"noopener\">DataConomy Roadshow Coverage<\/a>, <a href=\"https:\/\/www.edgen.tech\/zh\/news\/post\/nvidias-jensen-huang-denies-rubin-delay-sees-asic-rival-shift-50-to-gpus\" target=\"_blank\" rel=\"noopener\">EdgeN ASIC Competition Analysis<\/a>, <a href=\"https:\/\/semianalysis.com\" target=\"_blank\" rel=\"noopener\">SemiAnalysis Rubin Ultra Technical Analysis<\/a>, and multiple financial media reports.<\/em><\/p>\n","protected":false},"excerpt":{"rendered":"<p>In July 2026, NVIDIA CEO Jensen Huang appeared at a Morgan Stanley closed-door investor roadshow, revealing that quarterly revenue is approaching $100 billion, Rubin Ultra remains on schedule for 2027, and a major ASIC-focused AI lab has shifted nearly 50% of its compute back to NVIDIA GPUs. This article breaks down the five key signals from the roadshow and what they mean for the AI infrastructure landscape.<\/p>\n","protected":false},"author":253372376,"featured_media":13671,"comment_status":"closed","ping_status":"open","sticky":false,"template":"","format":"standard","meta":{"_acf_changed":false,"_jetpack_newsletter_access":"","_jetpack_dont_email_post_to_subs":false,"_jetpack_newsletter_tier_id":0,"_jetpack_memberships_contains_paywalled_content":false,"_jetpack_feature_clip_id":0,"_jetpack_memberships_contains_paid_content":false,"footnotes":"","jetpack_post_was_ever_published":false},"categories":[96987604,96987592,96987604,96987592],"tags":[96987769,96988087,96988087],"class_list":["post-13659","post","type-post","status-publish","format-standard","has-post-thumbnail","hentry","category-ai-news","category-featured-articles","tag-nvidia-2","tag-gpu-2-en"],"blocksy_meta":[],"acf":[],"jetpack_featured_media_url":"https:\/\/i0.wp.com\/ai-stack.ai\/wp-content\/uploads\/2026\/07\/en-1.jpg?fit=1920%2C1080&quality=100&ct=202603031250&ssl=1","jetpack_shortlink":"https:\/\/wp.me\/ph344V-3yj","jetpack_sharing_enabled":true,"_links":{"self":[{"href":"https:\/\/ai-stack.ai\/en\/wp-json\/wp\/v2\/posts\/13659","targetHints":{"allow":["GET"]}}],"collection":[{"href":"https:\/\/ai-stack.ai\/en\/wp-json\/wp\/v2\/posts"}],"about":[{"href":"https:\/\/ai-stack.ai\/en\/wp-json\/wp\/v2\/types\/post"}],"author":[{"embeddable":true,"href":"https:\/\/ai-stack.ai\/en\/wp-json\/wp\/v2\/users\/253372376"}],"replies":[{"embeddable":true,"href":"https:\/\/ai-stack.ai\/en\/wp-json\/wp\/v2\/comments?post=13659"}],"version-history":[{"count":1,"href":"https:\/\/ai-stack.ai\/en\/wp-json\/wp\/v2\/posts\/13659\/revisions"}],"predecessor-version":[{"id":13662,"href":"https:\/\/ai-stack.ai\/en\/wp-json\/wp\/v2\/posts\/13659\/revisions\/13662"}],"wp:featuredmedia":[{"embeddable":true,"href":"https:\/\/ai-stack.ai\/en\/wp-json\/wp\/v2\/media\/13671"}],"wp:attachment":[{"href":"https:\/\/ai-stack.ai\/en\/wp-json\/wp\/v2\/media?parent=13659"}],"wp:term":[{"taxonomy":"category","embeddable":true,"href":"https:\/\/ai-stack.ai\/en\/wp-json\/wp\/v2\/categories?post=13659"},{"taxonomy":"post_tag","embeddable":true,"href":"https:\/\/ai-stack.ai\/en\/wp-json\/wp\/v2\/tags?post=13659"}],"curies":[{"name":"wp","href":"https:\/\/api.w.org\/{rel}","templated":true}]}}