AMD or SMCI: Billionaire Steve Cohen Loads Up on One Top AI Stock

The stock market’s recent ups and downs reflect a growing unease as investors debate whether an AI-driven bubble is forming. Alphabet CEO Sundar Pichai recently warned of “elements of irrationality” in the current environment, noting that no company would be fully insulated if a correction hits.

The heightened focus on Pichai’s comments echoes a broader point from billionaire Steve Cohen, who noted that markets are “somewhat headline-driven,” a challenging backdrop for disciplined investing.

Nevertheless, Cohen, who runs Point72 Asset Management, a hedge fund boasting $40 billion of assets under management, thinks AI isn’t going anywhere. “This is a technology cycle that’s probably going to be long-lasting, have duration, and going to have a massive effect on how we live our lives today,” Cohen, who has a net worth of ~23 billion, said.

Being one of Wall Street’s biggest names, Cohen is not hanging around watching the tech take off without throwing his hat in the ring, but he has been making some changes to his AI portfolio recently. Until not long ago, chipmakers Advanced Micro Devices (NASDAQ:AMD) and Super Micro Computer (NASDAQ:SMCI) both nestled in his portfolio, but that is no longer the case, with one taking more space and the other entirely absent.

Let’s take a closer look at these big AI names to see what’s behind Cohen’s moves, and with some help from the TipRanks database, we can also gauge general Street sentiment here.

AMD

We’ll start with AMD, a company with a well-earned reputation for comebacks and one that has repeatedly shown it should not be underestimated. A little over a decade ago, the chip giant was on the brink of bankruptcy and lagging far behind Intel in the CPU market. Then Lisa Su took over the reins and transformed the company’s fortunes.

By both offering superior products and making the most of a series of strategic mistakes at Intel, AMD closed the gap considerably on its once-far-bigger rival, becoming a force to be reckoned with in the CPU space.

Fast forward to just a year ago, and AMD once again found itself up against a dominant competitor – this time Nvidia. AMD was widely viewed as arriving late to the AI race and lacking the capabilities to mount a meaningful challenge in the data-center GPU market. But that perception is now shifting. The company recently secured major agreements with OpenAI and Oracle, and its upcoming MI450 GPU is viewed as a strong offering that could materially pressure Nvidia’s lead. At the same time, AMD believes it is on track to generate tens of billions in annual AI revenue by 2027 and more than $100 billion in data-center revenue within the next three to five years.

Watching all this unfold, Cohen must feel confident AMD will keep on delivering the goods. During Q3, via Point72, Cohen bought 406,063 shares, increasing his stake by 31%. The fund now holds a total of 1,743,993 shares, currently worth about $355 million.

Benchmark’s Cody Acree, an analyst who ranks among the top 1% on Wall Street, is also backing AMD’s chances, believing the market is once again underestimating the chipmaker.

“We don’t believe AMD’s shares yet fully reflect the company’s longer-term potential of becoming a true competitive source of AI solutions to the market leader, Nvidia,” the 5-star analyst said. “With a significantly increasing Data Center silicon TAM, which AMD now projects to be $1 trillion by the end of the decade versus its prior estimate of $500 billion, we believe AMD is just now in the very early stages of realizing the benefits of its years of investment and development work and the AI market’s rapid expansion of inference and training workloads.”

“We believe AMD’s current AI position versus Nvidia is very analogous to its earlier competition with the previously dominant CPU supplier, Intel, where AMD steadily and methodically chipped away at Intel’s market share through many years of increasingly competitive designs and a consistently aggressive development roadmap. While we don’t expect Nvidia to make the same technological competitive mistakes as Intel, we believe the customers of the AI industry are even more eager for a viable supply alternative compared to where the compute industry was just a few years ago,” Acree further added.

Quantifying his bullish stance, Acree rates AMD a Buy, while his $325 price target points toward 12-month returns of 59%. (To watch Acree’s track record, click here)

Elsewhere on the Street, the stock claims an additional 27 Buys and 10 Holds, for a Moderate Buy consensus rating. Going by the $284.67 average price target, a year from now, shares will be changing hands for ~40% premium. (See AMD stock forecast)

Super Micro Computer

Super Micro’s path in AI stands in marked contrast to AMD’s. The company was one of the earliest winners of the AI cycle, with its stock climbing rapidly on the strength of its role in supplying servers for AI-centric data centers and its close relationship with Nvidia. Sustaining that early momentum, however, has become increasingly challenging.

Auditing issues, delayed filings, and the threat of a NASDAQ delisting soon soured sentiment and the stock bled heavily. However, the company eventually got back on track and resolved its auditing issues, restated its financials, and regained compliance with NASDAQ listing requirements.

That restored investor confidence, but recently the shares took a hit due to a more pedestrian reason: a disappointing quarterly report.

In its September quarter (F1Q) readout, Supermicro reported revenue of $5 billion, down 15.3% year-over-year and falling short of the consensus estimate by $800 million. On the bottom line, adjusted EPS came in at $0.35, missing expectations by $0.04. Looking ahead, the company anticipates FQ2 revenue between $10 billion and $11 billion, with a midpoint of $10.5 billion — more than $2 billion above the Street’s $8.05 billion forecast. However, that impressive revenue outlook was paired with an adjusted EPS guidance of $0.46–$0.54, below the expected $0.62.

It’s the kind of mixed setup that likely validated Cohen’s decision to exit his entire SMCI position – 534,369 shares – during Q3.

That is a move that makes sense to Bank of America analyst Ruplu Bhattacharya, who highlights the various problems here.

“The AI server market is very competitive and large deals typically come with lower margins as it’s a competitive bidding process. Super Micro’s guide for Dec quarter gross margin (6.5%, -300bps q/q) is meaningfully lower than we expected and is impacted by engineering costs, expedite costs and overtime costs as Super Micro scales production of the new Nvidia GPU racks (Blackwell Ultra),” Bhattacharya explained. “In our opinion, such costs can occur again when subsequent generations of GPUs are launched (Rubin) where the rack structure is undergoing a meaningful change (Kyber vs. Oberon). Sovereign remains a longer-term opportunity but gross margin impact remains unclear (conceptually, if Super Micro can sell more services then that can benefit margins).”

As such, Bhattacharya rates SMCI stock as Underperform (i.e., Sell), while his $34 price target still implies ~6% upside from current levels. (To watch Bhattacharya’s track record, click here)

One other analyst shares BofA’s bearish view, while an additional 10 analysts are split evenly between Buys and Holds. Overall, SMCI carries a Hold (i.e., Neutral) consensus rating. However, the outlook remains constructive, as the $46.82 average price target suggests potential 12-month upside of about 45%. (See SMCI stock forecast)

To find good ideas for stocks trading at attractive valuations, visit TipRanks’ Best Stocks to Buy, a tool that unites all of TipRanks’ equity insights.

Disclaimer: The opinions expressed in this article are solely those of the featured analyst. The content is intended to be used for informational purposes only. It is very important to do your own analysis before making any investment.