Tom Lee, the head of research at Fundstrat Global Advisors, went on CNBC this week and read out a short list of stocks investors should consider owning heading into the fall. 

He also named one crowd favorite that he would leave alone.

Tom Lee has become one of the best known analysts for a reason. He’s been helping professional money managers navigate the markets since the early 1990s. Not only did he call last year’s bull run early, but he also leaned into AI and energy stocks long before either trade became the obvious consensus play.

Given his record, it may be worth considering what he says now. Here are his current picks for this fall, the reasoning behind them, and the risk that comes with owning them to help you decide which names could deserve a place in your portfolio.

Tom Lee points to Arista Networks demand surge

Lee’s first pick is Arista Networks (ANET), and the reason for this choice is the company’s numbers.

Arista makes the high-speed switches that move data inside large data centers. As companies build AI systems, they need far more of this networking gear.

On August 4, Arista reported its first-ever quarter above $3 billion in revenue. Sales reached $3.036 billion, up 37.7% from a year earlier, according to a press release.

The company’s management then raised its full-year 2026 revenue guidance to about $12.6 billion, which points to roughly 40% annual growth.

Joseph Terranova of Virtus Investment Partners agreed with Lee on CNBC’s Investment Committee. 

He noted that Arista’s revenue growth is speeding up rather than slowing, Insider Monkey reported.

More AI and Bank Stocks:

However, there is a catch, and Arista’s filings point it out.  

A small number of major customers drive most of that growth.

That creates a real risk. When those customers place big orders, Arista’s revenue jumps. If even one pulls back, the shortfall shows up fast.

Arista also warned that its gross margin slipped to 62.9% in the quarter, down from 65.2% a year earlier, as bigger customers received larger discounts, according to its SEC filing.

Arista sells directly into AI demand, and buying the stock means accepting customer concentration as the cost of that exposure.

Fundstrat’s Tom Lee laid out a focused set of stock ideas for the fall, spanning AI networking, banking, and optical components.

Cindy Ord / Getty Images

JPMorgan Chase is Tom Lee’s top bank stock

Lee’s second pick moves away from technology entirely.

JPMorgan Chase (JPM) is a bank, and Lee added it as his top name in financial services.

The timing follows a record quarter. On July 14, JPMorgan reported net profit of $21.1 billion for the second quarter, up 41% from a year earlier. 

The increase came from an 86% jump in equities trading revenue to $6 billion.

Kevin Simpson of Capital Wealth Planning backed the call, pointing to a rebound in initial public offerings as the main driver. 

JPMorgan runs many of those deals and collects large fees for the work.

That IPO pipeline is already active, and JPMorgan now sits within reach of a milestone no bank has ever hit.

The stock trades at about 15 times earnings, well under the multiples on the trillion-dollar technology names.

Related: Jim Cramer reveals 6 AI stocks to watch in 2026

That means investors are paying for current profit, not for a forecast. 

JPMorgan gives you a way to invest in the AI boom without buying a chipmaker directly. 

The bank profits from the wave of dealmaking that AI spending is funding, including IPOs and related trading activity.

That trading strength is also the risk. The 86% jump in equities trading revenue came from unusually active markets. When markets calm down, that growth pace can slow just as fast.

How Lumentum stock fits the AI infrastructure trade

Lee’s third named pick is Lumentum Holdings (LITE), and it plays a specific role in the AI buildout.

Lumentum makes optical components that move data as light instead of electrical signals. Hyperscalers need these parts to connect the servers inside AI data centers.

The company’s results support this.

On August 11, Lumentum reported fiscal fourth-quarter revenue of about $1.01 billion, up 109% from a year earlier.

Analysts responded quickly to the news. JPMorgan raised its price target to $1,280, and Citi lifted its target to $1,200.

Related: JPMorgan resets LLY stock target on drug demand

Nvidia has also placed a direct bet on the company. 

In March, Nvidia announced plans to invest about $2 billion into Lumentum to support its light-based technology for optical networking and AI processor connectivity.

The concern here is price. Lumentum shares have climbed more than 130% this year, and several analysts now say the stock costs more than its growth can justify.

Buying Lumentum at this price means paying a premium for a stock that has already climbed a lot. 

That bet only pays off if AI demand stays strong through 2027. Management expects that to happen, but no company can guarantee it.

Where Lee still sees room in energy and cyclical stocks

Beyond the three named stocks, Lee continues to favor broader energy exposure.

He flagged energy as a likely outperformer earlier this year, arguing that years of underperformance set the sector up for a rebound.

Data centers consume enormous amounts of electricity, and the companies that generate and deliver that power stand to benefit as buildouts continue.

For investors, energy offers a different angle on AI. Instead of buying the chips or the switches, you buy the power that runs them.

This is the part of Lee’s view that spreads risk across a sector rather than a single stock. 

If a single company in the sector has a bad quarter, the sector position does not automatically fall with it.

The one stock Tom Lee says to avoid right now

Lee did not only hand out buy ideas. He named Robinhood Markets (HOOD) as a stock to avoid in 2026.

His concern is valuation. Robinhood carries a forward price-to-earnings multiple of 33.7x, well above Charles Schwab at 15.2x and SoFi at 24.6x, according to Yahoo Finance.

Lee grouped Robinhood with other crypto-sensitive names, including Galaxy Digital (GLXY) and Riot Platforms (RIOT), where price swings tend to be sharp.

Not everyone agrees with that call. 

Kevin Simpson pushed back on the same broadcast, saying he still backs CEO Vlad Tenev and the company’s direction.

Simpson pointed to HOOD’s growth instead. 

Robinhood’s second-quarter revenue rose 32% from a year earlier to $1.31 billion, helped by an increase in prediction-market activity.

Here is the tension a buyer has to consider:

  • The bear case: the stock is priced for perfection, and any slowdown in crypto or trading volume hits it hard.
  • The bull case: Robinhood keeps expanding into new products, and younger investors stay loyal to the platform.

If you own Robinhood, the question is whether its growth can keep pace with a valuation that already sits far above its closest rivals.

What to do with Tom Lee’s fall 2026 list

Lee’s picks have something in common: most of them sell directly into AI infrastructure spending, and the bank he selected profits from the deals that fund it.

That focus is a strength and a weakness. If AI spending holds, these names benefit together. If it slows, they could also fall together.

A few practical steps can help you use this list without simply copying it:

  • Check the valuation before you buy. Arista and Lumentum have already run up sharply, so your entry price matters.
  • Size each position to your own risk tolerance. Customer concentration at Arista and price swings at crypto-linked names are real.
  • Treat the energy idea as a sector bet. It spreads risk across many companies rather than one.

An analyst’s buy list is a starting point for your own research, not a substitute for it. 

Lee himself pairs every pick with a reason and a risk, and that is the part worth copying.

Before you act on any of these names, match them against your own timeline and how much loss you could absorb if the AI trade cools.

Related: Micron stock draws aggressive target reset from 5-star analyst