Sep 17, 2026 · 44 min
Cramer weighs policy risk, Fed tightening and AI oversight
Mad Money w/ Jim Cramer 9/17/26
The episode connects a harder investing environment with the practical promise of enterprise AI and the case for external regulation.
- 1Policy uncertainty and a new Federal Reserve tightening cycle make valuation and buying discipline more consequential.
- 2Databricks and Snowflake executives describe enterprise AI as a tool for faster analysis, lower costs, and broader access to data.
- 3Cramer argues AI companies cannot be trusted to police themselves without meaningful external oversight.
Don't miss
Cramer shifts from discussing AI’s commercial promise to arguing that the technology requires external regulation because self-policing is insufficient.
The brief
Jim Cramer opens with an unpredictable market shaped by Donald Trump’s policy uncertainty and a new Federal Reserve tightening cycle, then turns to valuation and buying strategy.
Cramer explains how investors can pyramid into a declining position by narrowing the gaps between purchases, while weighing Reddit’s valuation and Uber against a tougher rate backdrop.
Databricks CEO Ali Ghodsi and Snowflake CEO Sridhar Ramaswamy present enterprise AI as practical infrastructure for organizing data, accelerating analysis and helping non-coders make decisions.
The Lightning Round favors Casey’s, BlackBerry, GameStop, Credo and Iron Mountain, while Snap draws caution because it remains unprofitable.
The episode’s sharpest turn comes after the stock calls, when Cramer argues that AI needs meaningful external regulation rather than relying on companies to police themselves.
What was said on this episode
42 statements · 31 positive · 8 negative · 3 neutral
Donald Trump can make unpredictable statements at any time.
“We know he's capable of saying anything at any time”
Listen at 1:49
Donald Trump is impossible for markets to anticipate reliably.
“So he's impossible to game.”
Listen at 2:02
Declining bonds tend to coincide with declining stocks.
“When bonds go down, stocks tend to go down, too.”
Listen at 2:31
A 6% ten-year Treasury yield could be reasonable.
“The 10-year at 6%, I don't know, it could be. Could be reasonable.”
Listen at 3:09
Kevin Warsh will likely make the Federal Reserve opaque.
“I expect something similar from Kevin Warsh.”
Listen at 3:41
Markets currently favor a hawkish Federal Reserve stance.
“The market apparently wants a hawk”
Listen at 4:39
Investors should continue favoring cybersecurity stocks, including Okta.
“I continue to recommend the cybersecurity stocks, and I'd add Okta to the list”
Listen at 5:35
Intel is Cramer’s top stock pick, with Micron ranked second.
“I still think that Intel, as I tell club members, is the best stock in show, and Micron, number two”
Listen at 5:44
Investors should buy near Federal Reserve meetings to capture rebounds.
“buy as close to the meeting as possible”
Listen at 6:19
Reddit is valued at $29 billion.
“is a $29 billion company”
Listen at 7:00
Some companies would pay $40 billion for Reddit’s user and business base.
“there are companies that would pay $40 billion for just the book of business they have”
Listen at 7:02
Investors should narrow purchase intervals as a sharply declining stock becomes cheaper.
“we start shortening and lowering our gaps between each buy because it's too cheap”
Listen at 7:51
Databricks is valued at $190 billion.
“this private company is now worth $190 billion”
Listen at 9:31
Crisis Text Line’s AI intervention can save lives.
“it actually literally saves lives”
Listen at 11:22
Databricks Genie can automatically build an information graph that improves AI decision accuracy.
“it automatically gleans information and then creates this graph of small little snippets of information that helps the AI make accurate decisions”
Listen at 12:17
Databricks and Merck’s TEDDY model may shorten drug-development timelines.
“The hope is to really shorten the time it takes to develop drugs.”
Listen at 13:46
AI currently poses no existential risk.
“there is no existential risk”
Listen at 15:00
Buying stocks during Federal Reserve tightening usually causes investors to lose money.
“if you buy stocks when the Fed's tightening, it means you're trying to fight the Fed. And there's usually a good way to lose money”
Listen at 16:58
Investors should not remain bearish on technology too long during tightening cycles.
“Please do not stay too bearish on that sector for too long because it tends to bounce back and when it does, bounce back hard.”
Listen at 21:51
Investors should raise cash during market strength before potential declines.
“Raise some cash in the strength when we get these kinds of moves like we had today so you are ready if we go down hard.”
Listen at 24:09
Cramer continues to recommend Uber stock.
“I am still recommending Uber.”
Listen at 24:51
Snowflake’s AI products help businesses extract value and complete tasks faster.
“get real value from the data, get things done faster”
Listen at 26:52
AI has created unusually large opportunities for Snowflake’s data platform.
“AI has created more opportunity for a data platform like Snowflake than anything that we could have thought of.”
Listen at 29:32
Snowflake and Salesforce are complementary rather than direct substitutes.
“we are actually quite complementary”
Listen at 31:09
AI makes Snowflake optimization easier and improves customer efficiency.
“AI, in fact, makes things like optimizations easier so that our customers can use Snowflake more efficiently.”
Listen at 32:23
Snowflake’s sales-team production use cases increased over 40% year over year.
“Those numbers are up 40 plus percent year on year.”
Listen at 33:45
AI can increase human leverage by shifting work toward creation and judgment.
“what AI can do is... elevate all of us to focus on either building more or spending more time on the judgment aspects of what we do”
Listen at 34:36
Investors were wise to repurchase Casey’s stock after its conference call.
“I think you're wise to go back into it”
Listen at 36:53
Cramer agrees BlackBerry stock appears attractive.
“I agree with you.”
Listen at 37:18
Cramer requires Snap to become profitable before supporting it.
“I think that we need profitability before I can get behind it.”
Listen at 37:49
Cramer recommends buying GameStop stock.
“I'm willing to say that that stock is a buy.”
Listen at 38:21
Cramer recommends buying Credo over a 12-month horizon.
“I think you can buy the stock. I'm going to endorse it.”
Listen at 38:50
Cramer recommends buying Iron Mountain.
“I think it's a buy.”
Listen at 39:14
Allowing AI to self-regulate may be the wrong policy path.
“We may be going down the wrong path when it comes to letting AI regulate itself.”
Listen at 39:59
External AI regulation is necessary to stop harmful corporate behavior.
“At the end of the day, you need real regulation because it's the only thing that stops this kind of bad behavior when business threatens society and the welfare of its citizens.”
Listen at 40:45
NIST is a suitable starting model for AI regulatory standards.
“I kind of like this NIST”
Listen at 41:06
Major AI labs currently underinvest in protection against rogue AI agents.
“I think these massive AI labs spend far too little to protect us from the rogue agents”
Listen at 42:06
OpenAI and Anthropic should hire CrowdStrike, Palo Alto, and Okta for AI security.
“they would do well to bring in CrowdStrike. Palo Alto and Okta for a belt and suspenders approach.”
Listen at 42:13
CrowdStrike and Palo Alto can shut down rogue AI agents.
“Then we need CrowdStrike and Palo Alto to shut down those rogue agents.”
Listen at 42:27
Stronger AI regulation is needed to protect the industry from public backlash.
“I think we need to be more heavy handed to protect the AI industry from the backlash.”
Listen at 42:52
AI governance should use an informed regulator in a public-private partnership.
“We want a public-private partnership with the public represented by an informed regulator”
Listen at 43:09
A future rogue AI swarm could cause substantial financial and social damage.
“the next rogue AI swarm could cost us a lot more money and wreak a lot more havoc”
Listen at 43:26
Statements are attributed to the speaker as said on the episode and reflect their view at the time, not PodLume's. They are not advice.
Books & mentions
Some links are affiliate links — PodLume may earn a commission if you buy.
