The first half of 2026 was unusual for Nvidia’s stock. Around the middle of the year, the company, which had been the undisputed darling of the AI trade for years, found itself in an awkward position: it was still profitable and dominant, but it wasn’t exciting enough for a market that had fallen in love with memory chips and AI infrastructure plays. By late June, shares had increased by about 12%, while the larger semiconductor ETF had risen by almost 85%. For a brief moment, it seemed as though everyone had moved on to the next glamorous thing. It was short-lived.
Nvidia was comfortably close to its all-time high by early September, trading at about $230, up roughly 22% for the year. The late August earnings report caused the kind of harm to skepticism that only unadulterated data can. Fiscal 2027’s second quarter revenue was $96.2 billion, a 106% increase over the same period last year. There isn’t much room for ambiguity in that type of figure. Just the data center segment brought in $89 billion, a 117% increase from the previous year. The underlying business never truly slowed down, despite the stock’s mid-year softness.
During the earnings announcement, Jensen Huang said something that stuck: “compute is revenue.” It’s the best way to describe what Nvidia has become, even though it sounds almost too straightforward. These days, the company offers more than just chips. It sells the infrastructure that powers the AI economy as a whole. It is hard to claim that demand is anything but structural when Google agrees to pay SpaceX almost $1 billion a month to rent computing capacity, and those contracts are powered by Nvidia’s GPUs.
Nevertheless, there was some logic behind the stock’s mid-year decline. Nvidia’s flagship data center GPU, the B200 chip, dropped from a per-hour compute price of roughly $6.11 in late May to about $4.22 by late June, according to pricing data from platforms monitoring GPU rental costs. On prediction markets, some traders were actively wagering that GPU prices would keep declining.
Whether that pricing pressure is the result of a real oversupply, a normal market correction following a period of scarcity, or just a brief respite in a long-term increase is still up for debate. The truth might lie somewhere in the middle.
The trajectory Nvidia set for the third quarter is more difficult to contest. Just two years ago, the $108 billion in revenue that management guided for would have seemed almost unreal for any semiconductor company. Additionally, the company disclosed that its Vera Rubin platform is entering full production, with partners like Oracle Cloud, Google Cloud, and Microsoft Azure already ramping up. The new Vera CPU, which was created especially for AI agents, is expected to add an additional $20 billion to networking revenue, which currently stands at about $60 billion annually.
With a four-star rating and a fair value estimate of $310, Morningstar describes NVDA as moderately undervalued at current prices. The consensus price target on Wall Street is close to $326. That indicates that analysts think the market hasn’t fully priced in what’s still to come. It’s a significant difference from where shares are currently trading. Price targets are, of course, estimates, and the difference between the lowest ($180) and highest ($515) projections provides insight into the true extent of the disagreement.
The way Nvidia is able to feel both worn out and just getting started is what makes its story unique. Fortunes have already been made with the stock. Its market value has surpassed $5.5 trillion. However, based on the most recent quarter’s numbers, it appears that the business is still growing into something even bigger. The development of AI infrastructure is continuing. It’s speeding up. As of right now, Nvidia is still in the best position to profit from every dollar invested in it.
Nobody really knows if the stock will double from here, as some analysts predict, or stall in a range, as the cautious crowd anticipates. This uncertainty is one of the things that makes following NVDA truly fascinating—not just as a trade, but also as a window into the true direction of the technology economy.
