newsroom
$NVDA · Jensen Huang – Will Nvidia’s moat persist?
now playing · $NVDA
$NVDA···bullish· high· posjensen huang
Huang: Electrons-to-tokens transformation is Nvidia's defensible moat
The transformation of electrons to valuable tokens requires extreme co-design across architecture, system, fabric, libraries, and algorithms — a full-stack capability no competito…
$NVDA···bullish· high· posjensen huang
Huang: CUDA ecosystem moat — install base, programmability, and cloud ubiquity
Nvidia's CUDA moat rests on three pillars: hundreds of millions of GPUs installed across every cloud and on-prem, a programmable architecture that enables rapid algorithm inventio…
$NVDA···bullish· high· posjensen huang
Huang: Nvidia has best performance-per-TCO in the world; no competitor demonstrates better
Nvidia's stack delivers the highest tokens-per-watt and lowest cost per token; Huang challenges any competitor (TPU, Trainium) to run InferenceMAX or MLPerf benchmarks — none have…
$ANTHROPICneutral· medium· posjensen huang
Huang: Anthropic's TPU/Trainium usage is unique instance, not a trend; Nvidia now invested
Anthropic's use of Google TPUs and AWS Trainium was driven by early strategic investments from those clouds when Nvidia lacked capital to compete; Nvidia has since invested in Ant…
$OPENAIbullish· high· posjensen huang
Huang: Nvidia invested $30B in OpenAI; regrets not investing earlier in Anthropic
Nvidia now directly invests in frontier labs (OpenAI $30B, Anthropic $10B) to secure compute demand and align roadmaps; Huang admits missing Anthropic early due to capital constra…
$CRWV···bullish· high· posjensen huang
Huang: Nvidia invested $2B and backstopped $6.3B to bootstrap CoreWeave as AI cloud
Nvidia deliberately seeded neoclouds (CoreWeave, Nscale, Nebius) to expand GPU demand surface area; CoreWeave would not exist without Nvidia's supply allocation and financial supp…
$NBIS···bullish· medium· posjensen huang
Huang: Nebius transformed with Nvidia support; now doing fantastically
Nvidia's ecosystem support helped Nebius (formerly Yandex Cloud) pivot and scale as an AI infrastructure provider, demonstrating Nvidia's ability to catalyze new cloud entrants.
$NSCALEbullish· medium· posjensen huang
Huang: Nvidia support enabled Nscale's growth as AI cloud provider
Nscale is another neocloud that Nvidia backed with supply and ecosystem support to broaden the GPU install base and demand diversity.
$TSM···bullish· high· posjensen huang
Huang: 30-year trust-based partnership with TSMC; Nvidia is largest N3 customer
Nvidia-TSMC relationship operates without a legal contract but with deep mutual trust; Nvidia's demand visibility lets TSMC scale CoWoS, packaging, and leading-edge nodes (N3, N2)…
$NVDA···bullish· high· posjensen huang
Huang: Annual architecture cadence (Vera Rubin, Feynman) creates dependable supply foundation
Predictable yearly leaps (Blackwell → Vera Rubin → Vera Rubin Ultra → Feynman → next) let customers bet their business on Nvidia's roadmap; no other ASIC team offers comparable an…
$GROQbullish· medium· posjensen huang
Huang: Nvidia folding Groq into CUDA ecosystem for ultra-low-latency inference segment
As token value rises, inference market segments into high-throughput vs ultra-low-latency (high ASP); Groq's architecture addresses the latter, and Nvidia is integrating it rather…
$CDNS···bullish· mediumjensen huang
Huang: Cadence tools will see exponential demand from AI agent workflows
Same agent-driven tool usage thesis applies to Cadence EDA tools; software tool companies benefit from agent proliferation, not commoditization.
$SNPS···bullish· mediumjensen huang
Huang: Agent-driven tool usage will skyrocket Synopsys Design Compiler instances
AI agents will exponentially increase usage of EDA tools (Synopsys Design Compiler, floor planners, layout tools, DRC) as they explore design spaces far beyond human engineer capa…
$MU···bullish· mediumjensen huang
Huang: Early partnership with Micron on HBM/LPDDR was pivotal; they doubled down on AI memory
Nvidia's early roadmap sharing convinced Micron (Sanjay Mehrotra) to invest heavily in HBM and LPDDR for AI, creating a memory supply advantage that has been 'tremendous for the c…
$COHR···bullish· mediumjensen huang
Huang: Coherent part of silicon photonics ecosystem build-out for AI networking
Coherent, alongside Lumentum, received Nvidia investment to scale silicon photonics capacity critical for NVLink and Spectrum-X, addressing the interconnect bottleneck in AI facto…
$LITE···bullish· mediumjensen huang
Huang: Lumentum and Coherent investments reshaped silicon photonics supply chain
Multi-year investments in Lumentum, Coherent, and silicon photonics ecosystem (including COUPE with TSMC, double-sided probing) pre-fetched networking bottlenecks and enabled NVLi…
$ASML···neutral· mediumjensen huang
Huang: EUV supply is a 2-3 year solvable bottleneck if TSMC is convinced of demand
EUV machine scaling is not a fundamental limit; with a clear demand signal from TSMC (driven by Nvidia's commitments), ASML can replicate capacity quickly — 'once you can build on…
$AMZN···neutral· mediumjensen huang
Huang: AWS Trainium growth also 100% driven by Anthropic; Nvidia still dominant in AWS
Trainium adoption is similarly concentrated; most Nvidia GPUs in AWS serve external customers, not Amazon internal workloads, reinforcing Nvidia's merchant silicon model.
$GOOGL···neutral· mediumjensen huang
Huang: Google TPU growth driven entirely by Anthropic; not a broad market trend
TPU adoption is concentrated in Anthropic (due to Google's early investment), not a general shift away from Nvidia; Google's internal TPU use doesn't reflect merchant market dynam…
$AMD···neutral· lowjensen huang
Huang: OpenAI's AMD Titan accelerator is minor; OpenAI remains vastly Nvidia
OpenAI's custom silicon effort with AMD is acknowledged but dismissed as insignificant relative to their massive Nvidia footprint; Nvidia expects to continue deep partnership.
$AVGO···neutral· mediumjensen huang
Huang: Broadcom ASIC margins ~65%; not a structural threat to Nvidia's 70% margins
Custom ASICs for hyperscalers (like Anthropic's Broadcom deal) carry high margins themselves (~65%), so the cost savings vs Nvidia are minimal; plus Nvidia's TCO advantage include…
$HUAWEIneutral· highjensen huang
Huang: Huawei shipped millions of AI chips in record year; China has abundant energy and manufacturing
China's AI chip capacity (Huawei, SMIC, HBM2) is already substantial; with abundant energy and 7nm+ manufacturing, they can scale compute by ganging chips — export controls accele…