Nvidia shares hit a record high on Wednesday, closing at $154.31 and lifting the chipmaker’s market value to $3.77 trillion. The move allowed the AI titan to reclaim the title of the world’s most valuable company — but one firm thinks the rally isn’t even close to being over.
The stock extended its record high on Thursday, rising about 1% to $155.
The latest jump to record highs marks an impressive comeback for Nvidia after struggling against headwinds like the reveal of China’s DeepSeek AI and the impact of tariffs earlier in the year.
But the party is just getting started, according to one analyst.
Ananda Baruah of Loop Capital raised his Nvidia price target to $250, a bold move even among AI bulls. It’s the highest price target on the Street, implying an upside of 127% and a market cap of $6 trillion.
‘The math just works’
A $6 trillion market cap might sound crazy, but “the math just works,” according to Baruah. If you think AI infrastructure spending is robust now, brace yourself for even bigger capex from the AI hypercalers.
Loop Capital sees a coming surge in data center spending as companies like Amazon and Microsoft continue to increase their investments in the technology.
Loop Capital estimates that spending on non-CPU compute like GPUs and AI accelerators could soar as hyperscalers boost non-CPU compute from around 15% of their infrastructure to 50%-60% by 2028.
They said their work “suggests that Hyperscale & AI Factory (Sovereign, Neocloud & Enterprise) Gen AI & AI Accelerator compute spending ALONE could increase to~$2.0T by 2028 using current compute economics.”
Nvidia also anticipates rising demand for AI factories, which are vertically integrated data centers built to train AI models. Loop Capital’s projections suggest Nvidia has “line of sight” to tens of gigawatts of demand in the next two to three years.
Nvidia CEO Jensen Huang has stated that every gigawatt of demand translates into around $40 to $50 billion of Nvidia revenue, implying $450 to $900 billion of sales in the pipeline.
Additionally, the newest AI reasoning models are much more compute-intensive—and expensive—than anticipated, taking up to 150x more compute than traditional LLMs. That means Big Tech companies are growing their tokens at a rapid rate and spending more money on Nvidia AI server systems.
Baruah sees Nvidia’s data center revenue more than doubling in the next few years, from $115 billion in fiscal year 2025 to $367 billion in fiscal year 2028. He believes Nvidia is well-positioned to capture a significant chunk of that demand, as the company “remains essentially a monopoly for critical tech,” giving it pricing power and room for margin expansion.
“We’re entering the next ‘Golden Wave’ of Gen AI adoption,” Baruah wrote. “Nvidia is at the front-end of another material leg of stronger-than-anticipated demand.”
