Markets
Los Angeles Dodgers World Series Championship parade
(Jessie Alcheh/Getty Images)

Retail traders’ success is thanks to doubling down on two things that have worked: “AI” and “buy the dip”

Main Street bought after weakness at three distinct moments early in the year.

This year, we’ve seen evidence that the increased presence of retail traders is changing how stocks behave around earnings announcements, and even forcing institutional investors to buy what they’re buying.

“2025 is set to be a record year for retail traders,” JPMorgan strategist Arun Jain wrote on the footprint of the retail community, noting that inflows by the cohort are “tracking at ~1.9 times the 5-year average, 50% above the levels seen last year and 12% above the previous peak seen during the retail mania of 2021.”

And for these traders, it’s also been a successful stretch because of a continued willingness to double down on a theme that’s been the biggest driver of market success in recent years (AI) and a tactic that hasn’t yet let them down (buy the dip).

“Retail investors began the year by sizeably buying the dip during three episodes of weakness (Post-DeepSeek correction, Momentum Unwind, and Liberation Day meltdown) — building 75% of their year to date single stock position during Jan-Apr and making Tech, particularly Nvidia and Tesla, clear winners of this trend,” he wrote.

JPM cumulative retail buying
JPMorgan

(Side note: poor Apple!)

For years, retail has been building an increasingly de facto “overweight AI, underweight everything else” position.

“In fact, retail investors have proved their conviction in the AI theme by funding large purchases in AI30/Mag 7 with holdings in the SPX 470,” Jain added. “This bifurcation has been persistent since 2023 following the launch of ChatGPT.”

JPM retail quarterly buying activity
JPMorgan

Since the release of ChatGPT on November 30, 2022, the maximum number of days between fresh highs in the S&P 500 has been 128 sessions (or a little over six months), a milestone-free dry spell that ran from February 19 to June 27 of this year. During that period and thereafter, AI-geared stocks have played a key role in fueling the market’s gains.

As such, buying the dip — and doing so across AI stocks in particular — has been an extremely potent combination.

More Markets

See all Markets
Vince Carter

Nvidia dunks on the doubters

CEO Jensen Huang and CFO Colette Kress dismantled most of the recent arguments and bear cases put forward by their naysayers.

markets

Cipher Mining surges on additional AI hosting deal

Bitcoin miner turned AI compute power provider Cipher Mining jumped early Thursday after announcing a deal that fully leases its Barber Lake data center in Colorado City, Texas.

The deal — which is also giving a lift to IREN, another miner turned compute provider — is an expansion of a previous agreement with Fluidstack, a UK-based provider of GPU-based cloud networks. The new deal amounts to roughly $830 million in additional revenue over 10 years, Cipher says.

The market clearly loves it. But it’s worth pointing out that this agreement is a pretty good example of the byzantine financial structures that are increasingly accompanying plans for many billions of dollars of spending on the AI boom.

For example, Cipher also announced Thursday that it would be borrowing $333 million to finance an expansion of that Barber Lake data center through a private placement of debt.

That offering will be secured, in part, by the warrants Google received to purchase Cipher common stock worth roughly 5.4% of the company. (Those warrants, by the way, look a lot more valuable today, with Cipher mining up double digits.) Google is also backstopping Fluidstack’s borrowing plans to finance its build-out to the tune of $1.4 billion.

For now, this makes financial sense. Alphabet — one of the most successful companies on the planet — needs the computing power to compete in the AI race. And the quickest way to get that capacity is to essentially cosign leases for the smaller companies taking the lead in that build-out, thereby lowering development costs and helping to bring projects into existence.

But in this deal alone, things get awfully complicated awfully quickly, as Alphabet is essentially the prime customer of, an important debt guarantor for, and potentially a significant owner in Cipher Mining, once it transfers the warrants into an ownership stake of more than 5%.

This isn’t, on its face, a terrible thing. There are precedents for circular funding relationships in industries like aerospace, as it developed from the 1920s to the 1950s.

But financial complexity does have a history of essentially hiding the level and locus of financial risks a system is building up, essentially during periods of heady optimism.

The market clearly loves it. But it’s worth pointing out that this agreement is a pretty good example of the byzantine financial structures that are increasingly accompanying plans for many billions of dollars of spending on the AI boom.

For example, Cipher also announced Thursday that it would be borrowing $333 million to finance an expansion of that Barber Lake data center through a private placement of debt.

That offering will be secured, in part, by the warrants Google received to purchase Cipher common stock worth roughly 5.4% of the company. (Those warrants, by the way, look a lot more valuable today, with Cipher mining up double digits.) Google is also backstopping Fluidstack’s borrowing plans to finance its build-out to the tune of $1.4 billion.

For now, this makes financial sense. Alphabet — one of the most successful companies on the planet — needs the computing power to compete in the AI race. And the quickest way to get that capacity is to essentially cosign leases for the smaller companies taking the lead in that build-out, thereby lowering development costs and helping to bring projects into existence.

But in this deal alone, things get awfully complicated awfully quickly, as Alphabet is essentially the prime customer of, an important debt guarantor for, and potentially a significant owner in Cipher Mining, once it transfers the warrants into an ownership stake of more than 5%.

This isn’t, on its face, a terrible thing. There are precedents for circular funding relationships in industries like aerospace, as it developed from the 1920s to the 1950s.

But financial complexity does have a history of essentially hiding the level and locus of financial risks a system is building up, essentially during periods of heady optimism.

markets

Odds of December Fed cut creep higher after unemployment rate unexpectedly rises in September

The September jobs report was a mixed bag: much better job growth than anticipated, but the unemployment rate unexpectedly edged higher.

The release of this data, which was delayed by the government shutdown, showed that nonfarm payrolls grew 119,000 (compared to the expected 51,000), but the unemployment rate crept up to 4.4%, while economists thought it would remain steady at 4.3%.

Event contracts show that the likelihood of the US central bank standing pat in December moderated to about 65% from around 75% prior to the release.

(Event contracts are offered through Robinhood Derivatives, LLC — probabilities referenced or sourced from KalshiEx LLC or ForecastEx LLC.)

Job growth for the prior two months was revised lower by 33,000.

The market-implied odds of a Fed cut in December tanked on Wednesday after the Bureau of Labor Statistics said that the updated employment statistics through November wouldn’t be published until December 16 — that is, the week after the US central bank’s last meeting of the year.

During the press conference that followed the October decision to lower rates by 25 basis points, Fed Chair Jerome Powell said that a dearth of fresh data “could be an argument in favor of caution about moving,” adding that a rate cut in December was “far from” a foregone conclusion.

Fedspeak since that October rate cut has generally tilted hawkish. Some voting members like Boston Fed President Susan Collins and Kansas City President Jeffrey Schmid (who dissented from the last cut) have signaled that they are unlikely to support an interest rate cut in December. Fed Governors Chris Waller and Stephen Miran have publicly endorsed another rate reduction, while other officials have yet to take a definitive stance. The minutes from the October meeting said that “many participants” thought “it would likely be appropriate to keep the target rate unchanged for the rest of the year.”

Walmart cart

Walmart beats Wall Street estimates, hikes sales forecast

The retail giant beat on earnings and revenue while also raising its sales forecast.

markets

Nvidia’s blowout earnings and guidance are lifting the entire AI supply chain

You might expect a company that recently reached a $5 trillion market cap (the first in history to do so) to be a relatively mature enterprise; slowing down, with a fat cash cow at its core, multiple divisions pulling in billions of dollars, and a few exciting nascent bets on the future. You probably wouldn’t expect them to post a reacceleration in revenue growth.

But that’s exactly what Nvidia posted across its Q3 earnings yesterday, with its reported $51.22 billion in data center revenue giving a huge lift to risk sentiment.

With revenue and adjusted earnings per share beating by ~3%, and guidance that was even rosier, Jensen Huang and co. have given a shot in the arm to America’s entire AI complex. Aside from Nvidia itself, here are a few of the winners this morning:

  • Companies directly in the semiconductor supply chain are catching a bid, with giants like Broadcom and AMD making the most notable moves in early trading, up 2.8% and 4.4%, respectively, as of 5 a.m. ET.

  • Upstart data center players are making even bigger gains, with IREN and Cipher Mining up between 8% and 10%.

  • Palantir is up 3.4%, aided by a mention from Nvidia’s CFO, who said that the company was “supercharging the incredibly popular Ontology platform with NVIDIA CUDA-X libraries and AI models for the first time,” having previously run the software on CPUs.

  • The neoclouds are also soaring: CoreWeave is up more than 9%, as its business model is tightly wound with Nvidia’s own. Future demand to rent the ~250,000 Nvidia GPUs that CoreWeave owns feels more assured than ever. Nebius, which offers a full-stack solution for this “surge demand” for AI compute, is gaining as well, up about 7%.

  • A number of stocks in the “powering the AI boom” theme are also springboarding from the chip designer’s results. Behind-the-meter energy play Bloom Energy is up more than 5%, while nuclear-geared stocks such as Constellation Energy, Oklo, and Nuscale have all caught a bid in premarket trading — with gains of 2.7%, 4.4%, and 7%, respectively, as of 4:40 a.m. ET.

  • In the AI server space, Super Micro Computer is the standout, up nearly 6%. Data storage names like Western Digital and Seagate Technology Holdings are also trading modestly higher, paring some of their initial rise to be 2.5% and 3.4% higher, respectively, at the time of writing.

  • Other speculative stocks and risk assets have also turned green since the results. Bitcoin reclaimed the $92,000 milestone and quantum stocks were up modestly, while equity markets in Europe and Asia traded higher and S&P 500 futures climbed 1%.

On the earnings call, Huang said that “AI is going everywhere, doing everything, all at once” — that is certainly the case in premarket trading.

Latest Stories

Sherwood Media, LLC produces fresh and unique perspectives on topical financial news and is a fully owned subsidiary of Robinhood Markets, Inc., and any views expressed here do not necessarily reflect the views of any other Robinhood affiliate, including Robinhood Markets, Inc., Robinhood Financial LLC, Robinhood Securities, LLC, Robinhood Crypto, LLC, or Robinhood Money, LLC.