Jim Cramer's New Stock Picks (September 2026) — and the Catch

This page is a dated record of what Jim Cramer has said on air, not a shopping list. He makes calls nearly every weeknight, so any list of his “new picks” starts ageing the moment it is written — which is precisely why the useful thing to keep is a log with dates on it, including the names he turned down. The newest entries run to 18 August 2026; the earlier batches are still below, unchanged and in date order, so you can check how the older ones read now.

Two things to hold alongside it. His record is one of the market’s favourite arguments — there are strategies built to bet against him, and a fund that tried it. And the further a call sits from his disclosed trust portfolio, the less accountability sits behind it.

This article does not forecast any share price and carries no price target, his or anyone else’s. It records what was said, when, and — where the market has since given an answer — what happened afterwards. It is not investment advice.

What are Jim Cramer’s newest stock picks?

These are his most recent verifiable calls, made on his nightly CNBC show Mad Money, its rapid-fire “Lightning Round,” and the morning notes of his subscription investing club, which manages a charitable trust portfolio. The freshest batch runs from the lightning rounds of 13, 17 and 18 August 2026. Older batches are kept below, in date order, so they can be checked.

Eighteen names in three nights, and only five of them got a clean yes. Hess Midstream, Carpenter Technology, Uber, Tanger and Neurocrine are the buys. Five were turned down: DXC Technology, AeroVironment, Coupang, Archer Aviation and United Microelectronics. The other eight are neither — a speculation he “blesses” rather than recommends, a stock he says he cannot form an opinion on, a “good idea” with an instruction not to buy it all at once, and a company he praises and then explains the problem with in the same breath.

That eight is the number worth keeping. On these three nights, the most common answer was not a verdict at all, and it is the half of the segment that roundups drop.

One redirection, in the same pattern as earlier in the month: the Coupang caller was pointed at MercadoLibre, and then at Amazon ahead of that. The ticker asked about was not the ticker named.

What did he say between 5 and 12 August?

Three lightning rounds — 5, 11 and 12 August — and roughly half of the names in them were ones he talked callers out of. These are CNBC’s own transcriptions of the segment, logged here by date. Stances are summarised; the fragments in quotation marks are his words.

Wednesday 12 August

Tuesday 11 August

Wednesday 5 August

Seventeen names in three nights, and the split is the part worth noticing. He was positive on seven, negative or dismissive on five, and told callers to wait, start small or buy something else instead on the remaining five. A record that only logged the buys would misrepresent the segment — the lightning round is as much a filter as a source of ideas, and most “Cramer picks” roundups keep only the half that sounds like a tip.

Several of those answers are redirections rather than verdicts, which is a category most write-ups lose entirely: he sent a Cheniere caller to Cheniere Energy Partners, a Stellantis caller to General Motors, a Rocket Lab caller to Voyager, and a Fastly caller to Cloudflare. In each case the ticker that was asked about is not the one he ended up naming.

What did he say in late July and early August?

The batch this page led with when it was first published, kept here unchanged. These are the lightning rounds of 29 and 30 July and 3 August.

He passed on plenty in the same rounds, which is the part most roundups drop: NVE Corp (“I need to know this company more because it just moved up so much”), Rigetti Computing (“that’s one I don’t really care for”), Establishment Labs (“I do not know this stock”), and AMC Entertainment waved through only as a pure speculation. Earlier in July he preferred KeyCorp to Truist (Jul 21), called Vertiv “in speculative hands right now” (Jul 16), and put buys on Bloom Energy (Jul 15) and Amprius Technologies (Jul 8).

Three of the August names have run a long way already, which is worth knowing before you act on a rapid-fire call: as of the 1 September close, Astera Labs is up about 68% for 2026 and Howmet about 24%, while Carvana — the one he framed as an opportunity — is down about 15% on the year.

What were July’s picks, and how have they done?

July’s batch is below, unchanged from when we first published it, because the point of dating calls is that you can check them later.

His lightning rounds over recent months produced a rapid-fire batch too: buys on DoorDash and Leidos (May 26) — while defending Uber and Reddit as unfairly caught in the rotation into chips — a “buy very slowly” on Reddit (early June), a thumbs-up on Equinix for data-center exposure (May 20), praise for RTX (May 1) and Goldman Sachs (May 6), a not-bullish-on-gold-right-now (May 7), and a money-losing quantum name he’d swap for IBM (early June). One reversal worth noting: he called Joby Aviation “way too risky” back in April, then warmed to it as “a terrific spec” by June. Treat that list as scattered snapshots of the tape, not a portfolio.

The conviction call: GE Vernova

If one pick in this batch carries extra weight, it’s GE Vernova. Asked about it in the June 30 lightning round, he didn’t hedge: it’s his favorite of the power names and one he says to keep buying — and unlike most TV calls, this one has money behind it, because the charitable trust his investing club manages holds what he describes as a very big position, which the club discloses under CNBC’s rules.

The backdrop is a monster run — with a fresh reminder of how fast it can wobble. The power-infrastructure maker — gas turbines, wind, grid electrification — closed at $1,113 on 2 July, up roughly 70% for the year to that point and having roughly doubled over twelve months on the surge in electricity demand from AI data centers, before pulling back sharply on 7 July, closing down 6.5% after trading as much as 11% below the previous close. Its most recent quarter showed revenue up 16%, orders up 71% organically, and a backlog around $163 billion. The obvious tension is that none of this is a secret: the stock isn’t cheap by any traditional measure, and its next earnings report on 22 July is the kind of event that resolves a hot stock’s direction one way or the other.

What happened at the 22 July earnings

The test landed, and it split. GE Vernova reported second-quarter results on 22 July 2026 and the two halves of the print pointed in opposite directions.

The demand side was better than the article above assumed. Orders came in at $24.2 billion, up 88% year over year, and the backlog set a record at about $176 billion — up from the ~$163 billion quoted in July. Gigawatts under contract in Gas Power moved from 100 to 116 in a single quarter, against a company target of at least 125 by year end. Revenue was $11.1 billion, ahead of the ~$10.73 billion analysts expected, and the company raised full-year guidance to revenue of $45.5–46.5 billion and free cash flow of $11.5–12.5 billion.

The profit side missed. Adjusted earnings of $2.47 a share came in well short of the $3.04 Wall Street was looking for — a wide miss for a stock that had been rewarded specifically for execution. Shares closed down 8.7% on the print, at $985.03.

The “After Q2” column is as of the 4 August 2026 close. As of the 1 September 2026 close the stock is $898.53 — about +37% for 2026, and about 23% below where it stood when he made the 30 June call. The backlog figure is a record; orders came in at $24.2bn; the raised guidance is revenue of $45.5–46.5bn and free cash flow of $11.5–12.5bn; and the $2.47 adjusted EPS compares with the $3.04 Wall Street had expected.

So the thesis that drew him to it — AI-driven power demand filling the order book — got stronger, and the stock got cheaper anyway. That is the honest scoreboard on a conviction call five weeks later: the story held, the quarter did not, and anyone who bought on the 30 June call is underwater on it. He has since revisited it. On 24 August 2026 he said he now feels “the market won’t pay up for its order book because maybe it’s gotten soft,” pointing at political and community pushback over data-centre power demand — while noting his trust still owns it and that he is reluctant to take the loss. That is a marked cooling from the 30 June call, and it is his own dated words rather than our inference.

Where do Cramer’s picks come from?

Knowing the source helps you weigh each call. His picks arrive through three channels of very different depth. The nightly show delivers researched segments where he argues a thesis at length — the Moderna turn was one of these. The lightning round is the opposite: callers fire tickers at him and he answers in seconds, which is where quick buys like Medline and AST SpaceMobile come from, and it’s the segment even fans treat as entertainment first. The third channel is his subscription investing club, which publishes the reasoning behind the charitable trust’s actual portfolio — the only channel where his opinions are tied to disclosed positions, GE Vernova being the current showcase. A useful rule of thumb: the further a pick sits from that disclosed portfolio, the more lightly to hold it.

The catch: does following Cramer actually work?

Here’s the part most “Cramer picks” articles skip, and it cuts both ways.

The skeptics’ case is loud and partly institutionalized. His high-conviction misses over the years became famous enough that “Inverse Cramer” turned into a genuine strategy: trackers exist that systematically bet against his most-recommended names, and in 2023 an actual Inverse Cramer exchange-traded fund launched to short his picks — though, in a twist the skeptics don’t always mention, it shut down within about a year after attracting only around $2.4 million in assets and losing roughly 15% along the way. The broader, more serious critique comes from the evidence on stock-picking itself: decades of scorecard data show most professional pickers trail simple index funds over long periods, and critics — including fiduciary advisers — argue that a nightly stream of buy calls encourages exactly the overtrading that erodes returns. On this view, his show is financial entertainment, and confusing it with a plan is the real risk.

The fairer reading includes his own framing. His long-stated manifesto is that tips are for waiters — that the show’s job is to teach viewers how to think about markets, not to hand out overnight winners — and the picks with real accountability live in the disclosed trust portfolio, not the lightning round. Plenty of his calls have also been right, sometimes spectacularly so; GE Vernova has been a huge winner over the period he has backed it publicly, though not since this batch of calls: it is down roughly 23% from the 30 June close, and by 24 August he was himself sounding cautious on the order book. The honest conclusion is that his value is as a fire-hose of ideas and market color from someone watching more closely than almost anyone — not as a signal to be traded blindly, in either direction.

This is a roundup and an explainer, not investment advice — Drawpie isn’t a financial adviser, and Jim Cramer’s opinions here are his, dated as shown, and subject to change on his next show. Do your own research and consider a licensed professional before acting on any pick, his or anyone’s. For the market theme behind several of these calls, see the rotation into AI-hardware stocks .