Nvidia Corporation (NASDAQ:NVDA) already sells the chips that train and run the world’s AI models. On July 26, the company made a case for owning a lot more of the design process too, expanding NVIDIA Agent Toolkit for engineering with the PhysicsNeMo and CUDA-X libraries that let software agents run physics simulations, solvers, and even quantum chemistry work largely on their own. It sounds technical, and it is, but the pitch is simple. Instead of engineers manually running simulations for weeks, an AI agent can now do the grunt work.
The Bull Case: High Software Retention
Nvidia’s chip business is not slowing down. In its fiscal first quarter of 2027, the company reported revenue of $81.6 billion, with data center sales accounting for $75.2 billion of that, and it guided to roughly $91 billion the following quarter. However, the Agent Toolkit expansion points to something else: turning that hardware advantage into a software habit. Siemens is already using the toolkit inside its Fuse EDA AI Agent and reports more than 10x faster chip library characterization, with token costs cut by a similar margin. Samsung is applying the same libraries to lithography and chip-scale thermal analysis across models containing up to 10 billion cells. Once an engineering team builds its workflow around Nvidia’s agent tools, the switching cost to a rival platform gets steep, which is the same kind of stickiness that turned CUDA into a decade-long advantage in the first place.
The Bear Case: Revenue Concentration Risk
None of that has translated into a hot stock this year. Nvidia shares are up only modestly for 2026 so far (around 17% as of August 6), trailing the S&P 500, even as the company keeps beating growth expectations. Part of the hesitation is concentration risk. Data center sales made up 92% of last quarter’s revenue, and that revenue rests heavily on a handful of hyperscaler customers whose capital budgets get revisited every year. A broader semiconductor sell-off in early June wiped out more than a trillion dollars in combined chip stock value, underscoring how quickly sentiment in the sector can shift.
Wall Street Is Pricing Nvidia And AMD Very Differently
Institutional interest in Nvidia has grown, with the hedge fund count rising from 264 to 275 in the most recent 13F filings, a signal that more institutional holders were accumulating positions as of that reporting date. Advanced Micro Devices (NASDAQ:AMD), the peer chasing Nvidia hardest in this same window, saw a smaller increase, from 132 to 134 funds.
AMD earned that comparison by launching its own agentic AI infrastructure push on July 22, headlined by Helios rack-scale systems now in production and a multiyear deal with Anthropic to deploy up to 2 gigawatts of Instinct GPUs. Under this partnership, Anthropic will begin deploying the first gigawatt of MI450 Series GPUs in H1 2027, backed by AMD’s commitment to make a strategic equity investment of up to $5 billion in the AI company. The alliance also includes a multi-year engineering collaboration to utilize Claude to optimize workloads for AMD Instinct hardware, accelerate ROCm software development, and deploy Claude broadly across AMD’s internal engineering and product development teams.
Short interest tells a different story. Only about 1.39% of Nvidia’s float is sold short, against roughly 2.45% for AMD. Valuation is where the gap is widest: as of August 4, Nvidia trades at a forward P/E ratio around 22x, cheap for a company still growing revenue at double-digit sequential rates, while AMD trades at roughly 66x forward earnings as it rolls out its Helios and MI400 platforms.
The Debate Comes Down To Patience
Nvidia’s Agent Toolkit expansion is a bet that owning the tools engineers use to design the next generation of chips is worth as much as owning the chips themselves, and early adopters like Siemens and Samsung suggest the idea has traction. For the bull case to win, that traction needs to show up in results as a business that Nvidia can quantify on an earnings call.
While we acknowledge the risk and potential of NVDA and AMD as an investment, our conviction lies in the belief that some AI stocks hold greater promise for delivering higher returns and doing so within a shorter time frame. If you are looking for an AI stock that is more promising than NVDA and AMD and that has 10,000% upside potential, check out our report about this cheapest AI stock.
READ NEXT: 10 Best Future Stocks to Buy Under $10 and 12 Best Performing Semiconductor Stocks to Invest In.
Disclosure: None. Follow Insider Monkey on Google News.
