#AIStockGuruReportedlyBullishOnAI

Calling someone bullish on artificial intelligence is easy. Proving that bullish view with numbers is much harder. For me, the AI thesis is becoming increasingly powerful because accelerating compute demand, massive hyperscaler investment and supply constraints across critical semiconductor components are working together. AI is no longer simply a software narrative; it is becoming one of the largest infrastructure build-outs of this generation. But markets do not reward narratives forever. Eventually, price, earnings, valuation, liquidity and real demand must confirm the story.
At the centre of this cycle stands Nvidia. The stock closed around $223.67 on September 9, 2026, down 0.91%, with a market capitalisation of roughly $5.39 trillion and turnover of approximately 79.27 million shares.

Just two sessions earlier it traded near $230.36, while September 2 delivered a 3.21% gain on about 156.55 million shares. On August 28, Nvidia dropped 4.57% while approximately 194.28 million shares changed hands. Its 52-week range is around $164.27–$236.54, leaving the stock close to its yearly high and still roughly 36% above its 52-week low.
The longer-term performance explains why Nvidia remains the flagship AI asset. Year to date, it is up approximately 17.36% versus 10.90% for the S&P 500. Over one year, Nvidia gained around 23.44% versus 16.22% for the index. Over three years, it is up approximately 380.58% compared with 70.31% for the S&P 500, while its five-year gain is roughly 875.72%.

Average daily volume around 130–133 million shares represents approximately $29–30 billion changing hands in a normal session. A five-year beta near 2.22 also confirms that Nvidia is a high-volatility growth engine rather than a defensive stock.
The fundamental numbers are even more impressive. Nvidia reported approximately $96.22 billion in fiscal second-quarter revenue, up 18% sequentially and an extraordinary 106% year over year. Data-centre revenue reached around $89 billion, while net income stood near $59.69 billion, producing a net margin above 62%. A multi-trillion-dollar company still producing triple-digit revenue growth demonstrates why the AI thesis cannot simply be dismissed as hype.

Valuation makes the story more interesting. Nvidia's trailing P/E is around 28, while its forward multiple is near 14. Street targets have clustered around $323–$328, implying roughly 45% upside from $223.67. A move to $250 represents about 12% upside, $275 around 23%, $300 approximately 34%, $325 about 45%, $350 around 56% and $400 nearly 79%. These are scenarios, not guarantees. Annualised volatility around 41.8% and an RSI near 52.7 also show that the stock was not in an obvious technical overbought condition at that snapshot.

One of the most interesting developments is that Nvidia has lagged the broader semiconductor sector. The Philadelphia Semiconductor Index gained roughly 61% during 2026 through late August, while Nvidia had advanced only around 11.7%, creating a relative-performance gap of nearly 49 percentage points. I do not see this as evidence that AI is weakening. I see it as evidence that the opportunity is broadening. Capital is moving into memory, networking, optics, power, cooling and semiconductor equipment. An AI data centre is an ecosystem, not a single chip.

Broadcom provides another powerful confirmation. The stock traded around $416.05, up approximately 20.65% year to date and 33.97% over twelve months. Quarterly revenue reached about $22.19 billion, up 47.9% year over year, while AI semiconductor revenue reached roughly $10.8 billion, an extraordinary 143% increase. When Nvidia and Broadcom are simultaneously reporting powerful AI growth, the thesis becomes much harder to describe as a one-company bubble.

Memory provides perhaps the clearest evidence that this demand is physical. DRAM prices reportedly surged around 90% in the first quarter of 2026, while broader DRAM, NAND and HBM pricing increased approximately 80–90%. Consumer DRAM spot prices were reported nearly 700% higher year over year. HBM demand is expected to grow around 70% in 2026 and consume approximately 23–25% of DRAM wafer output. The HBM market is estimated around $54.6 billion, up roughly 58%, while SK Hynix and Micron have indicated that 2026 HBM production is effectively sold out. That is a real supply-demand imbalance, not simply social-media excitement.

Micron trades around 6–7 times forward earnings, while SK Hynix carries a forward multiple near 5. Those valuations become particularly interesting alongside the rapid growth in AI-memory demand. When customers compete for limited capacity and prices rise, the market is showing us the physical economics behind the AI boom.

Then comes the fuel underneath the entire machine: capital expenditure. Combined 2026 spending guidance from the four largest hyperscalers has moved toward $720–745 billion. Including Oracle pushes the figure toward approximately $835 billion, while other estimates place the largest five around $775–800 billion, roughly 64% above the previous year. Expectations for 2027 have already moved above $1 trillion. If these budgets translate into real infrastructure, the beneficiaries extend far beyond GPU manufacturers into memory, servers, networking, electricity, transformers, cooling and data-centre construction.

This is why I believe the biggest AI question is not simply which single stock wins. The bigger question is how much capital continues flowing through the entire infrastructure chain. Every AI cluster requires HBM, networking, storage, racks, power delivery, cooling, grid capacity and physical data-centre space. That creates multiple routes to benefit from the same secular trend.

The physical infrastructure story is already visible. Vertiv increased quarterly sales around 24% year over year to roughly $3.27 billion and raised full-year guidance toward $14 billion. Dell entered its fiscal year with an AI-server backlog near $43 billion. These numbers demonstrate the multiplier effect of AI capex: money spent on compute eventually flows through servers, memory, networking, power systems and cooling.

But disciplined analysis must challenge the bullish narrative. The ECB has warned that previous technological revolutions show how stretched valuations can eventually correct, while the Financial Stability Board has highlighted leverage, elevated asset prices and interconnected AI investments. These warnings do not prove an AI crash is coming. They simply remind investors that extraordinary growth does not eliminate valuation risk.
Circular financing is another risk worth watching. Nvidia has committed substantial capital toward OpenAI, which is also a major Nvidia hardware customer, while other financing arrangements have developed across the AI ecosystem. This does not automatically make demand artificial, but investors should monitor how much demand is supported by genuine end-user cash generation versus increasingly complex financing structures.

Free cash flow, depreciation and GPU economics also matter. AI equipment represents a much larger share of hyperscaler capex than it did several years ago, while some compute assets can have relatively short economic lives. If financing conditions tighten or GPU rental rates fall sharply as supply expands, returns on new infrastructure could come under pressure. AI can remain technologically successful while the financial returns on infrastructure temporarily disappoint.

This is why I separate the AI thesis from the timing trade. I remain structurally bullish on AI over the multi-year horizon, but Nvidia does not need to rise every month or every quarter. A secular bull market can still contain 20%, 30% or even deeper corrections without necessarily destroying the long-term thesis.
My scenario framework is simple. In the bullish case, AI spending approaches $800 billion during 2026, exceeds $1 trillion in 2027, HBM remains constrained and Nvidia earnings continue compounding rapidly. Under that combination, the $320–$400 zone becomes a credible upside framework. From $223.67, $320 represents roughly 43% upside, $350 around 56% and $400 nearly 79%. In the base case, spending continues rising but growth slows toward 20–40%, allowing earnings to catch up with valuation. In the bearish case, financing tightens, GPU rental prices fall, depreciation assumptions change and valuation multiples contract faster than earnings grow.

The next major catalysts will be Nvidia's expected November 2026 results and major AI announcements such as Meta's September 23–24 developer event. I would personally watch hyperscaler capex, HBM pricing, GPU rental rates, cloud free cash flow, Nvidia guidance and the relationship between volume and price. If Nvidia repeatedly trades 130–190 million shares without breaking resistance, distribution could become a concern. If heavy volume accompanies breakouts, rising earnings expectations and improving semiconductor breadth, it would suggest institutional accumulation.

My final view is clear: the AI bullish thesis remains fundamentally powerful, but it should never be confused with blind optimism. Nvidia's 106% revenue growth, Broadcom's 143% AI semiconductor growth, memory pricing increases of 80–90%, consumer DRAM prices reportedly nearly 700% higher year over year, HBM demand growth near 70% and hyperscaler spending approaching $800 billion create an extraordinary fundamental backdrop.

For me, the strongest conclusion is not simply “AI will go up.” The stronger conclusion is that AI infrastructure demand remains powerful enough to justify a constructive multi-year outlook while the short-term path remains uncertain. Growth tells us why the sector can expand. Earnings tell us whether that growth is becoming economically real. Valuation tells us how much future growth is already priced in. Liquidity tells us how violently positioning can change.
I remain bullish because the numbers increasingly confirm the technology. AI is no longer merely changing software; it is reshaping the physical architecture of the global technology economy. Nvidia remains at the centre, but memory, networking, power, cooling, servers and semiconductor equipment are all becoming part of the same enormous investment cycle. The trend may still have significant room to run—but the numbers must continue earning that confidence.
Make the bullish and bearish cases clearer