The Best AI Stocks to Buy Now, Plus We Rank Our Top 4 Neoclouds

Listen to the audio version of this article (generated by AI).

When SpaceX Corp. (SPCX) went public on June 12, financial advisor Jeff Barnett got a call from a client. Had he bought shares for the portfolio? He had not. SpaceX came nowhere close to the client’s preset criteria on valuation and governance. The client listened, then asked: “Can we just buy 10 shares?”

Barnett got it done, reluctantly. He likened it to buying a lottery ticket the moment the jackpot hits $1 billion. Sure the odds are slim, but a small bet buys a share of the buzz… plus, it scratches that deep-down “what if” itch without putting the rest of the portfolio at risk.

That client understood that you do not need to pick the one winner in a young, fast-moving industry. You just need exposure to the field.

This week’s Being Exponential Q&A searches for that exposure in five separate corners of the market: Space communications, optics, semiconductor ETFs, neoclouds, and physical AI chips.

In an industry still writing its own rules, the investor who owns the field beats the investor who bets the farm on a single name. Watch this week’s latest podcast for more:

AST SpaceMobile vs. SpaceX: Why Choose?

A subscriber made a sharp observation this week: AST SpaceMobile Inc. (ASTS) is technologically ahead of SpaceX in direct-to-cell service. That observation is not wrong. AST SpaceMobile leads on commercial development and technical capability in that specific niche.

But betting against Elon Musk carries its own long, expensive history. Musk trailed on electric vehicles for years before Tesla Inc. (TSLA) became the EV standard-bearer. He trailed in other ventures, too, before pulling ahead. Direct-to-cell satellite communication remains an early industry, and early industries reshuffle their leaderboards constantly on the road to commercial maturity. SpaceX enters that fight freshly capitalized off its IPO, and launching satellites is not a cheap habit to fund.

This is the same logic behind holding NVIDIA Corp. (NVDA) and Advanced Micro Devices Inc. (AMD) side by side rather than going all-in on Nvidia alone. Yes, an all-in Nvidia bet outperforms in hindsight. But hindsight is not a strategy; it is a rearview mirror.

In the early stages of a massive growth story, you want multiple shots on goal, because you rarely know in advance which company survives to dominate the category. Recommend both AST SpaceMobile and SpaceX.

One is technologically ahead while the other brings a battle-tested track record and a mountain of capital.

There is room for both to win.

Corning: The Bounce Is Not Finished

A subscriber flagged Corning Inc. (GLW) as overvalued by 22%. Corning earns its place in this portfolio, and the stock’s recent chart proves why patience pays. After a strong earnings report shook out weak hands and dropped shares to the $110 to $115 range, Corning roared back to $152, with a recent high near $170.

The story here is bigger than one ticker. Lumentum, Coherent, Corning, and Applied Optoelectronics all delivered the same read-through this earnings season: The 800G-to-1.6T optics transition is happening in real time, and multiple growth vectors are stacking on top of each other inside this trade.

Look at the charts for Corning, Applied Optoelectronics, or Lumentum, and you see the same pattern repeating: a red-hot run, a selloff, a bounce off major support, and a rebound that continues.

Optics stands out as one of the strongest trades for the second half of the year, and Corning sits at the center of it.

Skip DRAM and Own the Whole Chip Trade Instead

One subscriber got burned holding Broadcom Inc. (AVGO) and wanted to know about sector-based ETFs, specifically a DRAM-focused fund. Here is the direct answer: DRAM burns you the same way Broadcom did, because memory names rank among the most volatile in the entire semiconductor complex. Sandisk swung from $200 to $2,000 to $1,000 to $1,800 within a single year. Micron, Western Digital, and Seagate all carry that same whipsaw reputation.

If you want AI infrastructure exposure without memory’s stomach-churning volatility, the VanEck Semiconductor ETF (SMH) and the PHLX Semiconductor Index (SOX) do the job. SOX runs about 1.5 times more volatile than SMH in either direction, so SMH becomes the calmer vehicle for the same underlying exposure.

Skip the in-between sector plays like DRAM or a hypothetical optics fund. Either build a basket of individual stocks, or park capital in SMH and let the broader AI buildout do the work.

Ranking the Neoclouds

On the neocloud question, the rising tide lifts every boat in this category over the next 12 months, but that does not mean every boat deserves an equal allocation. Ranked from favorite to least favorite: Nebius Group (NBIS), CoreWeave Inc. (CRWV), Applied Digital Corp. (APLD), and IREN Ltd. (IREN).

Nebius earns the top spot on the highest growth rate, the most financial firepower, and the highest overall business quality in the group.

CoreWeave lands at No. 2, still delivering strong growth off a fantastic recent earnings report, though a heavier debt load adds real risk to the story.

Applied Digital continues to post good growth with lingering leverage concerns of its own.

IREN sits at the bottom of this ranking, not because it is a bad stock, but because its Bitcoin-miner-turned-neocloud pivot signals less operating experience on the AI compute side relative to its peers.

That pivot makes strategic sense, and IREN remains a fine holding. It simply is growing slower growth than the other neoclouds.

Qualcomm: The Sleeping Dragon

Qualcomm Inc. (QCOM) delivers a genuine one-two punch that the market has not priced in yet. Punch one lands now: Qualcomm just entered the data center chip market, putting it directly into competition with Nvidia, AMD, and Intel for the first time. That development alone could drive earnings growth above consensus estimates while the stock’s cheap multiple expands to reflect the new opportunity.

Punch two lands later, tied to physical AI. Nine out of 10 prototype devices coming out of Big Tech right now run on Qualcomm’s Snapdragon processor, whether that final form factor turns out to be glasses, a necklace, or something nobody has invented yet. Most of those nine projects will not generate meaningful revenue.

But one of them becomes the next iPhone, and when that happens, Snapdragon carries a 90% probability of sitting inside it. That single device could do for Qualcomm from 2026 through 2035 what the original iPhone did for the stock from 2007 through 2019. Qualcomm remains a sleeping dragon, a stock worth recommending both for the near-term data center catalyst and the long-term physical AI bet.

Want the full breakdown, complete with live charts and Luke’s real-time read on every one of these tickers? Watch the complete Being Exponential Q&A episode now.