The Bull Case for Nvidia: Numbers Don't Lie (Completely)
Nvidia closed Friday at $190.17, a 1.77% uptick. Bank of America is maintaining a "buy" rating with a target price of $275. That’s a hefty premium, and it's predicated on some serious growth projections. Specifically, BofA analysts bumped their fiscal year 2026/27/28 non-GAAP EPS estimates by 3%/12%/14% to $4.56/$7.02/$9.15, respectively. Sales estimates also got a facelift, rising from $204.14 billion to $208.48 billion in fiscal year 2026, from $271.62 billion to $300.19 billion in 2027, and from $343.59 billion to a whopping $383.95 billion in 2028.
The core argument rests on continued, explosive growth fueled by insatiable AI demand. And to be fair, the demand is there. But is it sustainable at these projected levels? That's the multi-billion dollar question.
BofA's $275 target is based on a 30x multiple of their calendar year 2027 price-to-earnings (P/E) estimate excluding cash. This, they claim, is within Nvidia's historical forward year P/E range of 25 to 56. Now, here's where things get interesting. A P/E of 56 represents peak exuberance, while 25 suggests a more sober outlook. Thirty is somewhere in the middle. But is "somewhere in the middle" justified given the current climate? Bank of America resets Nvidia stock forecast before earnings - Yahoo Finance
Cracks in the Foundation? AI Skepticism and Capital Expenditure Realities
Michael Burry, of "The Big Short" fame, is shorting Nvidia and Palantir. He alleges that AI hyperscalers are artificially inflating earnings. That's a bold claim, and it's worth dissecting. Is Burry seeing something the BofA analysts aren't? Or is he simply early (or wrong, which happens to even the best of us)?
One factor to consider is the seasonality of cloud capital expenditures. These tend to peak in Q4. So, if the hyperscalers are indeed "juicing" their numbers, the effect would be most pronounced in the current quarter. We'll need to see how Q1 2026 shapes up (or doesn't) to get a clearer picture. (Q1 numbers always tell a story).
Nvidia stock is currently trading about 8% below its recent peak of $207.04. That's not a crash, by any means, but it's a noticeable dip. It coincides with growing skepticism surrounding the long-term viability of the AI boom. Are we in a true paradigm shift, or are we witnessing a classic hype cycle?

SoftBank Group, notably, sold all 32.1 million of its Nvidia shares (a $5.8 billion stake) and plowed the money into OpenAI. This raises eyebrows. Was SoftBank simply diversifying, or did they lose faith in Nvidia's long-term prospects relative to OpenAI's? I've looked at hundreds of these filings, and this particular shift is unusual – it suggests a very specific bet on a different part of the AI ecosystem.
Nvidia has a strong product roadmap. They’ll have three chip generations – Blackwell, Blackwell Ultra, and Vera Rubin – by the second half of 2026. That's impressive. But technology roadmaps are a dime a dozen. The question is whether the demand will keep pace with the supply. And this is the part of the report that I find genuinely puzzling.
The AI Gold Rush: Fool's Gold?
Here's my take: the Bank of America analysts are likely correct in their direction, but potentially off on their magnitude. AI is here to stay, and Nvidia is undoubtedly a key player. But projecting near-hyperbolic growth for the next three years seems overly optimistic.
Consider this analogy: Nvidia is selling the picks and shovels during a gold rush. There's a huge initial surge in demand. But eventually, the low-hanging fruit is gone, the easily accessible gold is mined, and the demand for picks and shovels tapers off. The gold rush doesn't end entirely, but the exponential growth phase does.
The market is pricing in sustained, exponential growth. I believe a more realistic scenario involves a deceleration, perhaps to a still-impressive but less-frenzied pace. A P/E multiple of 30 might be justified for a company growing at, say, 20% annually. But at the projected 30%+? That's a big ask. And even Vivek Arya at Bank of America has had to address AI bubble fears more than once.
