Walnut Ave Value
Value Investor & Writer
Nvidia's push into AI models and agent software looks weak even though its GPUs and racks are great. Its Nemotron model is mediocre and its LangChain-based agent stack won't make it. If Nemotron were capable of frontier performance it would be there already, and Nvidia doesn't seem to know how to get there. x.com x.comÂ
Nvidia's $6B push into open-weight AI directly helps Palantir: the stronger and more widely adopted Nemotron becomes, the more valuable Palantir is as the platform deploying it. Per @StockSavvyShay, Nvidia is licensing Poolside's technology and moving 100+ staff onto Nemotron. Nemotron is already built into Palantir's AIP and Ontology, and since June it has powered custom US government models. As open models mature, the scarce piece is safely wiring them into enterprise data and workflows. x.comÂ
Nvidia's AI profits are real, but its compute demand depends heavily on fragile venture funding. Per Chamath, 1% of firms account for 90% of AI spending, and 72 cents of each dollar comes from unprofitable VC-backed firms. Venture money passes through startups to Nvidia and cloud suppliers; if the next raises fail, concentrated orders could contract together. x.com x.com x.comÂ
Nvidia's system components can reduce the idle time that erodes returns on expensive AI chips. CPUs, accelerator interconnects and server networking help operators such as Nebius and IREN keep machines doing useful work; solving the next bottleneck could win Nvidia more of the data-center equipment budget even when a rival chip looks appealing alone. Buyers must compare the cost of completed work across the whole system. x.comÂ
Nvidia faces risks on both sides of the AI cycle. Vendor financing is a concern at the center of the buildout: strong AI sentiment brings competition, while weaker sentiment brings a slowdown. That makes it the largest holding of greatest concern. x.comÂ
Nvidia's case is a lot weaker than two years ago, because large providers such as OpenAI and Meta are likely to design custom chips tuned to their own models. Nvidia wins only if workloads and models stay diverse. The claim that frontier labs are decoupling from Nvidia is doubtful, since OpenAI's recent efficiency gains likely came from Blackwell, but it holds for Google and Amazon. Nvidia also needs to own the open-source ecosystem, which is contradictory by definition. x.com x.comÂ
Enterprise agents widen NVIDIA's opportunity beyond inference because reasoning, tool use and company-data workflows need infrastructure and controls outside the model. Satya Nadella describes those needs; Vera CPUs, OpenShell and Sentry on BlueField are the NVIDIA read-through. x.comÂ
NVIDIA loses from the current AI drama because it is the king of frontier AI training, and a real slowdown would hurt it. As Elon Musk, Dario Amodei and Sam Altman pivot to warning that AI could kill people, AI, already polling badly, will struggle to attract the investment it was getting. If that money dries up, labs with no business model beyond AGI are cooked, and Broadcom, Marvell and Micron lose too. x.comÂ
Nvidia could become the weakest link in AI's circular financing economy as it takes on more customer risk. It is turning to insurers to cover customers' GPU debt, and if insurers require Nvidia to keep backstopping deals, its liabilities will keep growing on top of $245B of supply chain commitments and prepayments as of last quarter. Such large counterparty conflicts are unwise. Amazon is also looking to move $8B of Nvidia Blackwell GPUs into an SPV to take on more debt for its AI buildout. … Show more
AI infrastructure bottlenecks could give eight stocks as much as 18x upside. Nvidia supplies frontier-model GPUs; Broadcom and Marvell supply custom silicon and networking; TSMC makes advanced chips; Micron supplies sold-out HBM; Vertiv provides power and cooling; Palantir monetizes enterprise AI; Alphabet spans TPUs, research, cloud and apps. The list tracks where AI infrastructure spending is flowing. x.com x.comÂ
Nvidia and Micron face the risk that AI's true running costs eventually hit the semiconductor buildout. Microsoft canceled internal Claude licenses over escalating token expenses, a sign that running models at scale costs far more than flat-rate revenue, so either companies curb usage and slow frontier labs' revenue ramp or labs absorb losses. High memory prices are also driving workarounds, such as optical interconnects that allow more HBM. For now, semis remain the only game in town. x.comÂ
