Why The World Is Spending More On War

Summary

Prof G Markets — "Why The World Is Spending More On War"

Date: July 8, 2026
Hosts: Ed Elson, Scott Galloway
Guests: Steve Feldstein (Carnegie Endowment), Zed Francis (Convexitas), Ara Kharazian (Ramp)

Overview

This episode unpacks the dramatic escalation in global defense spending, focusing on the drivers behind the new global arms race—heightened by conflicts, shifting US military priorities, and the surge of tech industry involvement in defense. In addition to the geopolitical analysis, the episode covers recent turbulence in semiconductor markets and presents new research challenging dire predictions about AI-driven job losses.

---

Key Discussion Points & Insights

1. The New Global Arms Race

Guest: Steve Feldstein, Carnegie Endowment (03:38–13:12)

• Surge in Defense Spending:

  • NATO allies have agreed to over $90 billion in new defense deals at the Turkey summit.
  • The US is proposing a record $1.5 trillion defense budget for 2027.
  • Silicon Valley has already raised $19B+ for defense tech in 2026, outpacing all of 2025.

• Drivers Behind Increases:

  • More Global Conflict:
    • 250,000 people killed in battlefields last year, with rising insecurity worldwide.
    • "War is prevalent and around. Countries recognize that there's growing levels of insecurity." — Steve Feldstein (03:46)
  • Russian Revanchism:
    • Russia’s ongoing escalation in Ukraine and boosting of its drone and weapons capacity triggers NATO concerns.
  • Rapid Tech Advancement:
    • New technologies like AI-based targeting systems are influencing how wars are fought and what is required for future military preparedness.

• Tech Companies Entering the Fray:

  • Silicon Valley’s defense ambitions now complement traditional arms makers rather than replace them.
  • Citing Palantir’s "Maven" targeting system (utilized in the Iran war) as a paradigm for the future — combining enterprise tech innovation with conventional munitions.
  • "It doesn't mean that you don't need traditional ballistic missiles. What it does mean is… you need a more efficient way in which to generate the coordinates for how you would target and identify… military assets." — Steve Feldstein (05:38)

• Conflicts of Interest & Ethical Dilemmas:

  • Dual roles of big tech giants (e.g., Microsoft, Google) in civilian and military spheres are questioned.
  • "At some point, these issues are going to become front and center where countries will say, look, either you're providing military equipment… or you're selling to consumers… not for military use." — Steve Feldstein (07:30)

• Geopolitical Instability:

  • 2026 is more unstable than 2025, given the Iran conflict, escalation in Ukraine, U.S. policy reversals, and a "volatile president."
  • Limitation of US influence, with Trump focusing on the Middle East/Indo-Pacific and moving assets out of Europe.
  • "All these ingredients coming together make for lots of unpredictability… in the coming months." — Steve Feldstein (09:27)

• Forecast for Defense Markets:

  • Increasing spending likely, especially in counter-drone technology.
  • But actual manufacturing capacity must catch up to allocated budgets.
  • "That's a growth area… you'll see continuing demand and investments for more of these types of systems." — Steve Feldstein (11:13)

• Trump’s Greenland "Threat":

  • Trump revives interest in U.S. control of Greenland; Feldstein doesn’t expect imminent action, but “we can’t just take it off the agenda completely either because he's the president of the United States and he says it. And so therefore it matters.” (12:29)

---

2. Semiconductor Stocks: From Rally to Rout

Guest: Zed Francis, CIO, Convexitas (15:27–23:17)

• Chip Market Pullback:

  • Despite record earnings from Samsung (revenues doubled, profits up 1,800%), the stock fell 7%, dragging major indices down.
  • "Chips have just gotten crushed… and it seems like it's coming to an end." — Ed Elson (16:34)

• Why Are Chip Stocks Falling?

  • Valuations are stretched; P/E multiples are now well above three-year averages, raising performance expectations.
  • "I would expect… if there's anything, any small item that feels like a miss to have these kind of nasty downward moves… a very volatile earnings season for the sector." — Zed Francis (17:20)

• Short Bets & Volatility:

  • Michael Burry is shorting the semiconductor sector, expecting sharp declines (e.g., betting Micron will fall 30%).
  • "It's a common theme that folks are using options to bet against the space… daily volumes are about 4 to 1 puts to calls." — Zed Francis (19:22)

• Hedging as a Theme:

  • Heavy use of put options indicates widespread hedging—trading behavior itself now fueling volatility.
  • “A significant amount of hedging and potentially over-hedging in that space… the options are pricing it in. I mean it’s really expensive to have that view.” — Zed Francis (20:39)

• Hyperscalers vs. Infra Players:

  • Hyperscalers (Microsoft, Meta, Oracle) have lagged as semiconductor and memory companies boomed, but dynamics may be shifting.
  • Meta’s move to lease out excess compute capacity signals a possible business model pivot.
  • "They're just kind of doing a… grown up type of play where they're using their balance sheet to go ahead and build out capacity that other players cannot… capturing that spread over the long term." — Zed Francis (22:16)
  • Potential for hyperscalers to profit as intermediaries; chip momentum may persist if infrastructure buildout continues.

---

3. Is AI Really Destroying Jobs?

Guest: Ara Kharazian, Lead Economist at Ramp (24:42–33:25)

• Main Finding:

  • Contrary to predictions of mass AI-induced layoffs, new firm-level data shows companies adopting AI are actually expanding headcount.
  • "The firms using AI heavily grew their headcount 10% over the two years following adoption… outsized growth for entry level hires. So 12% over two years." — Ara Kharazian (24:54)

• Nature of AI Adoption:

  • Most companies adopting AI do so in superficial ways (e.g., basic ChatGPT), which offers limited productivity gains.
  • High-intensity adopters (top third) use multiple models and dedicated tools for business operations, driving real growth.
  • Even outside tech, high-intensity users aren’t cutting headcount.
  • "If they're not necessarily hiring more, their headcount is not necessarily changing because of AI adoption." — Ara Kharazian (30:51)

• Job Growth or Causation?

  • The research controls for company growth trajectories by comparing firms that adopt AI early to those that eventually do.
  • Growth accelerates after adoption, especially among high-intensity users—robust across tech and non-tech groups.
  • Skepticism remains warranted about attributing all layoffs to AI.
  • "Most layoffs and job losses in general are due to a number of reasons… There's a side finding… that the tech sector… reversed [job shedding] following AI adoption." — Ara Kharazian (32:03)

• Who’s Being Hired?

  • Companies looking for AI-savvy talent, especially among recent graduates.
  • Change in rhetoric from tech leaders: big layoffs blamed on AI last year are now being walked back as hiring trends reverse.

---

Notable Quotes & Memorable Moments

On the world’s increased investment in defense:

• "War is prevalent and around. Countries recognize that there's growing levels of insecurity." — Steve Feldstein (03:46)

On Silicon Valley in the defense sector:

• "At some point, these issues are going to become front and center…" — Steve Feldstein (07:30)

On market dynamics in chip stocks:

• "I would expect… if there's anything, any small item that feels like a miss to have these kind of nasty downward moves." — Zed Francis (17:20)

On the AI job loss panic:

• "The firms using AI heavily grew their headcount 10% over the two years following adoption… outsized growth for entry level hires." — Ara Kharazian (24:54)

On the ongoing debate over AI’s impact:

• "If you're someone on the job market and… companies [say] we have to do all these layoffs because of AI, I would say that you should be skeptical…" — Ara Kharazian (30:51)

---

Important Segment Timestamps

• 03:38 — Steve Feldstein: Why the world is spending more on war; implications of the NATO summit.
• 07:15 — Ethical conflicts of tech companies entering defense.
• 09:27 — Assessment of 2026's geopolitical instability and US unpredictability.
• 11:13 — Future defense tech growth areas (counter-drone tech).
• 15:27 — Zed Francis: Chip stock volatility, sector valuation changes.
• 17:20 — The case for increased chip sector volatility and risk of corrections.
• 19:22 — Michael Burry’s chip sector shorts and the prevalence of bearish bets.
• 22:16 — Hyperscalers leveraging their capital for AI infrastructure leasing.
• 24:42 — Ara Kharazian: AI’s real (not projected) impact on hiring.
• 29:53 — Methodological discussion on measuring AI impact; robustness of results.
• 30:51 — Rebuttal to sensational headlines about AI-driven layoffs.
• 33:25 — Summary and end of content portions.

---

Episode Tone & Takeaways

• Tone: Direct, analytic, slightly skeptical, and jargon-free.
• Takeaways:
  • Defense spending is surging due to new and old conflicts, tech integration, and shifting US commitments.
  • The semiconductor rally is showing cracks as valuations catch up with expectations, inviting both volatility and short-selling.
  • Widespread job loss from AI is not (yet) evident in data; the real story is nuanced and specific to sector and intensity of use.

---

This episode effectively weaves together global security, market reactions, and technology’s evolving role in both warfare and the workforce—anchored by sharp analysis and grounded by clear, real-world data.

Transcript

[Commercial Announcer]

Exchanges on the M and A and IPO landscape. Exchanges on the dynamics affecting global trade. For the sharpest analysis on finance, business and the economy, count on exchanges. The Goldman Sachs Podcast Listen now.

[Ara Kharazian]

A lot of companies are investing in AI, but most only have a small group actually using it and an even smaller group who are really seeing the value. Superhuman fixes that from the makers of Grammarly Superhuman Superhuman AI lives in every tool your team already uses for instant help with drafts, summaries and more. So habits and proficiency form. Naturally, when every person on your team works at their best, that's when your AI investment starts to compound. See what Superhuman can [e-mail]

[Steve Feldstein]

this episode is brought to you by Google Chrome. You think you know a browser, but Gemini and Chrome?

[Commercial Announcer 2]

That's new.

[Steve Feldstein]

It can help you with practically anything on the web, like restoring a vintage motorcycle from a 50 page restoration block. Or finally break down that long article you've had open for weeks. Gemini and Chrome is here for it, ready to make anything online make sense. There's no place like Chrome. Check responses set up required compatibility and availability. Various 18.

[Zed Francis]

Thousand money markets matter. If money is evil, then that building is hell. Show goes on the right generic wants to sell Sell.

[Ed Elson]

Welcome to Profge Markets. I'm Ed elson. It is July 8th. Let's check in on yesterday's market vitals. The major indices declined as chip stocks tumbled once again. More on that in a minute. Oil rose after three ships were attacked in the Strait of Hormuz. The US Rescinded its authorization of Iranian oil sales following those strikes. Microsoft stock rose on news that it is replacing OpenAI and Anthropic with its own models to bring down costs. And finally, SpaceX declined nearly 7% on its first day in the NASDAQ 100. Okay, what else is happening? The NATO summit kicked off yesterday in Turkey, and the big takeaway so far is that a new global arms race is underway. Trump is pulling US Military assets out of Europe and shifting focus to the Middle east and Indo Pacific. That is forcing Europe to shoulder more of its own defense now. NATO allies have agreed to at least $50 billion in defense industry deals and $40 billion in counter drone capabilities. But America isn't slowing down either. Trump's 2027 budget proposes $1.5 trillion for defense, which is the largest request in history. And Silicon Valley wants in defense Tech companies have raised more than $19 billion so far this year, already blowing past last year's record in 2025. In short, an enormous amount of capital is flowing into defense. Right now, the question is, what exactly are we gearing up for? Joining us to discuss this, we are speaking with Steve Feldstein, senior fellow at the Carnegie Endowment for International Peace. His new book, Bites and Bullets, Global Rivalry, Private Tech, and the New Shape of Modern Warfare, comes out in September. Steve, thank you so much for joining us on the show today. I'd love to get your initial reactions to what we saw at the NATO summit. It seems like a big conclusion or a big takeaway is that we're going to spend a lot more money on weapons, a lot more money on defense. Is that the right takeaway?

[Steve Feldstein]

I think it is. Thanks for having me on. And I really think that there's an overall context that's worth bearing in mind. And I think there are three factors that I've been looking at over the past couple years that are really relevant. So I think the first one is that we are just seeing far more conflict occurring around the world today than even in the last five years. Last year, something like 250,000 people were killed alone on battlefields. So it means that war is prevalent and around. And countries recognize that there's growing levels of insecurity, and so that matters. I think the second thing is that we're seeing a growing menace from Russia as it continues to get on a wartime footing because of the Ukraine war, as it continues to increase its drone capacity and manufacturing of other types of weapons. Countries in Europe, in NATO in particular, are concerned, and they know they need to keep up and they need to retool accordingly. And then finally, I think especially coming from the Iran war, we're seeing the use of new technologies, particularly AI targeting systems and so forth in play. And I think it's also giving a recognition to countries like Germany, France and others that they also need to really think carefully about innovation and what types of weapon systems they need to incorporate so that they are able to fight wars effectively in the future.

[Ed Elson]

Yeah, I mean, one of the stats that blew me away that we just learned is that defense tech has raised more than $19 billion this year, and that's in six months. So we already beat last year's record. And it seems as though Silicon Valley and technology is playing more and more of a role on the battlefield. What do you make of that dynamic and that shift? How should we feel about the fact that tech companies are participating in warfare at this point?

[Steve Feldstein]

It's a trend that's, that's, you know, really begun to. To ramp up. And I see it as not necessarily a substitute to traditional arms manufacturers, but more as a compliment. So I'll give you a good case in point. You know, a lot of people have focused on the Maven targeting system from Palantir that's paired with Claude and how it was used in the Iran war. And so a lot of people look at BON systems like that and they say, okay, we need something similar. We need a better way to have more efficient targeting and to match our missiles to potentially potential places, military targets to strike. So it doesn't mean that you don't need traditional ballistic missiles. What it does mean is that you need that plus you need a more efficient way in which to generate the coordinates for how you would target and identify, you know, different military assets. And so to that end, a lot of it is sort of creating other types of weaponry in addition to what's already needed in terms of munitions, based on the fact that countries see that there's more war and more insecurity all around, whether it's from China prospectively, Russia in the, in the near region or other places.

[Ed Elson]

When you look at some of the tech companies that are participating, I mean, some of the companies you've written about, Google, Meta, Cloudflare, obviously Palantir, one thing that always strikes me is these are companies that are offering enterprise software services and solutions and then also playing a role in, in the Ukraine war, in the war in Iran. I mean, that's quite a striking dynamic. And I wonder if that's something that we should, I don't know, maybe be worried or investigate in some capacity. The fact that, you know, the CEO of an enterprise company is also directing the outcomes of what happens on the battlefield. I'd be curious to get your views on what you make of that. And if that is worth thinking about.

[Steve Feldstein]

I think you're right. It is a new role. And I think there's a few different conflicts of interest that arise from that. I mean, I think one of the great examples is, you know, if you have a company like Microsoft, and it is both providing data infrastructure that is being used by militaries like Israel's military or other militaries, but it's also selling consumer goods to a wide variety of clients. Well, how do those two things work, work together? And I don't think there's an easy answer to that, but so far they've been able to skirt around the issue. At some point, these issues are going to become front and center where countries will say, look, either you're providing military equipment for the security objectives of a particular country, or you're selling to consumers across the board, neutral items meant for civilian applications and not for military use. And so these are, I think, questions that are coming to bear for a variety of companies, whether it's OpenAI, Anthropic, Microsoft, Google, AWS or others, which are so far somewhat having it both ways.

[Ed Elson]

You mentioned earlier that what all of this spending tells us is how unstable things have become from a geopolitical perspective. I remember last year, Jamie Dimon saying that this was sort of the most unstable, or maybe he said the most uncertain time since World War II. I mean, I think that was almost a year ago. That was certainly last year. Where are we now? Like, when you compare July 8, 2026 to July 8, 2025, is this a more unstable time and is it going to continue to get more unstable?

[Steve Feldstein]

It is. I mean, certainly last year we hadn't yet encountered the Iran war and all the spillover effects from that. So that's a really big thing that everyone is sort of grappling with. And it's not just about Iran and immediate adversaries in the Middle East. It's about a wider security architecture and questions there. We also see a continuing amount of fighting when it comes to the Ukraine war. And if anything, we're seeing even more escalation between the two sides with Ukraine striking Crimea, Russia coming back and striking Kyiv and making bigger threats in terms of other potential adversaries. And then that's not to mention the kind of prospective questions related to China, which frankly at this point has been quieter because of these other situations that taking place. So there are a lot of different geopolitical issues that are flaring up simultaneously. And we also are seeing, I think, a bit of an ebbing of power, or at least the limitations to American power at this point, not to mention a volatile president who seems to change on a dime that very quickly. The geopolitical objectives of the United States, all these ingredients coming together make for lots of unpredictability when it comes to security and where things, what things will look like in the coming months.

[Ed Elson]

I think the question for investors is, I mean, the main question is will the spending keep going up? Clearly, it's gone up last year, coming into this year. Will defense tech companies keep on getting massive contracts? Will they get larger contracts? Will some of the larger, more traditional players get larger contracts? Will government spend more and more? Do you think that defense spending will go even higher than it is today?

[Steve Feldstein]

So one issue is that we have seen a ramp up in NATO spending, as you mentioned, and I think that will continue to increase. But what's interesting right now is that there's a gap between the amount of money that's there and the ability of current manufacturers to be able to produce according to the money that's been allocated. So I think the first thing we're going to need to see is greater capacity that's built up to match the money and the allocated contracts that are there for munitions. I think the second thing that we're going to see is a real ramp up when it comes to countering drones. So drones are proliferating. We're seeing that Iran, we're seeing that in Ukraine and so forth. And we are also seeing a dire need by countries like uae, Saudi, the United States, for that matter, countries in Europe when it comes to actually countering these drones and coming up with the right systems. That's a growth area. That's an area where you just are starting to see products at scale come to the market. And I think you'll see continuing demand and investments for more of these types of systems, especially as we see drone use proliferate, which I'm certain will continue to be the case.

[Ed Elson]

Final question. Trump is now renewing his threats thoughts about Greenland. He said the territory, quote, should be controlled by the United States. Is the Greenland story not over? Is that something that we should be worth, that is worth us paying attention to like we did at the beginning of the year?

[Steve Feldstein]

Well, it's certainly not over in his head. And because he is president of the country, I don't think we can completely dismiss it. It's not something that I tend to focus on on a day to day basis. Trump says a lot of things. He has a lot of grievances. He has a lot of different issues that are top of mind for him. They tend to move from point place to place depending on context and where he is. So it's that I think is due for an imminent action. But we can't just take it off the agenda completely either because he's the president of the United States and he says it. And so therefore it matters.

[Ed Elson]

Steve Felting, a senior fellow at the Carnegie Endowment for International Peace. Steve, we appreciate your time.

[Steve Feldstein]

Thanks for having me.

[Ed Elson]

After the break, Chip stocks take a tumble. And for even more markets insights, you can subscribe to my weekly newsletter, Simply put@simply put. Prof.gmedia.com.

[Commercial Announcer 2]

Support for the show comes from gusto. When you run a small business, you don't just do the job itself. You're also the hiring manager, the payroll department and the benefits team. And you're usually switching between all those hats before lunch. Gusto takes a few of those items off your plate quickly and seamlessly. Gusto is an online payroll and benefits software built for small businesses. It's all in one remote, friendly and incredibly easy to use so you can pay, hire, onboard and support your team from anywhere. Automatic payroll tax filings, simple direct deposits, health benefits, commuter benefits, workers comp 401k, you name it. Gusto makes it simple and has options for nearly every budget. Unlimited payroll runs for one monthly price. That means no hidden fees and no surprises. You can save time with built in automated tools. Offer letters, onboarding docs, direct deposit, and more. It's quick and simple to switch to Gusto. Just transfer your existing data to get up and running fast. Plus you don't have to pay a cent until you run your first payroll. Try Gusto [e-mail] markets and get three months free when you run your first payroll. That's three months of free [e-mail] markets. Again, that's Gusto.com markets

[Commercial Announcer]

when you need to build up your team to handle the growing chaos at work, use Indeed Sponsored Jobs. It gives your job post the boost it needs to be seen and helps reach people with the right skills, certifications and more. Spend less time searching and more time actually interviewing candidates who check all your boxes. Listeners of this show will get a $75 sponsored job credit@ Indeed.com podcast. That's Indeed.com podcast. Terms and conditions apply. Need a hiring hero? This is a job for Indeed Sponsored Jobs

[Commercial Announcer 2]

this summer, soccer is here and so are the watch parties. And Uber Eats has your game day essentials covered. Everything you need to keep your crowd happy delivered straight to your door. Get wings, pizza, chips, dips, fresh ingredients and more. Keep the soccer going and all your game day essentials flowing. Order in so you can lock in for the games only on Uber Eats for a limited time.

[Ed Elson]

We're back with Prof. G Markets. After a meteoric rise, the semiconductor rally might finally be slowing down. Samsung, which is up nearly 400% over the past year, posted blowout earnings yesterday. The company saw revenues more than double and profits increased by 1,800%. And still, that wasn't enough to satisfy investors. The stock fell 7%, sending nearly the entire market down with it. The Philadelphia Semiconductor Index closed the day down 6%, the NASDAQ closed over 1%, and the S&P 500 closed over a half percent. So for more on how Samsung's beat somehow spooked the market and what's going on with chip stocks? We are speaking with Zed Francis, CIO and co founder at Convexitas. Zed, thanks for joining us. Chips have just gotten crushed over the course of the day and if actually if we look at the semiconductor index over the past five days, it's down 11%. This was supposed to be the leader. It has been the leader in the market throughout the year and it seems like it's coming to an end. Your initial reactions to what we're seeing?

[Zed Francis]

Yeah, this quarter earnings is going to be very different than Q1 earnings. And it's as simple as valuations have changed dramatically back in Q1. So think back to, you know, April and early May when all these names were reporting. They basically all kind of did the beat and raise. And mostly the stocks had very favorable performance on the day after those earnings. But spot PEs were below 3 year average back then. Today they're about 10 handles above their 3 year average. And forward PEs back then were below the S&P 500. Today they're above the S&P 500. Now the PEG ratio for the space is still reasonably attractive. But all that means is the market is pricing in much differently this time. The amount of growth embedded into these companies and so the hurdle is that much higher. So I would expect this entire earnings to, if there's anything, any small item that feels like a miss to have these kind of nasty downward moves. And I expect it to just in general be a very volatile earnings season for the sector.

[Ed Elson]

Is there any reason why it's happening now? I mean it just seems strange. Samsung reports this crazy earnings report. It's a beat on, on pretty much all fronts. Unless I'm wrong about that. I'm pretty sure it was. And yet now's the time that investors say we're out. This has gone too far again.

[Zed Francis]

I think you're going to need Micron level of blowout to really impress the market at this point with the current kind of valuation metrics in the space.

[Ed Elson]

Just looking at what Michael Burry has done, he's kind of made a name for himself in this semiconductor boom because he is shorting it. He shorted the semiconductor index. He shorted Micron. He also shorted Nvidia. He's betting that Micron's going to fall 30%. What do you make of Michael Burry's moves? Do you think that he's ultimately making the right choice?

[Zed Francis]

My expertise is in volatility. And when looking at volatility, he's not in a unique Camp implied volatility, the actual price embedded within the options you can actually trade is off the charts. You know, semiconductor, you know, using SMH, we can do the same thing with SOX. The spread of volatility in the semiconductor ETFs in comparison to QQQ, again, massive overlap of constituents. That spread is at its 100th percentile. SMH volatility on its own is near its all time 100th percentile. This is a sector that's obviously performed very well. Yes, it's 10ish percent off of those highs from a couple weeks back. But it's a common theme that folks are using options to bet against the space. And why you can say with a little bit more conviction is when you look at the indices again, sox, smh, things along that basket of style of trading instruments, daily volumes are about 4 to 1 puts to calls. Open interest is similarly about 4 to 1 puts to calls. And during Q2, implied volatility just kept on climbing during all that trading activity that that was heavily put slanted suggesting that it's mostly people buying puts as volatility is expanding. Well that's mostly trading. And additionally as the semiconductor space has faded, as you said over the last 5ish trading days, volatility in those ETFs is actually fading a little bit with it. So we're seeing that stocks up, volume up, stocks down, volume down. Which again suggests to me a significant amount of hedging and potentially over hedging in that space. Why? When it's profitable to be hedging the space, there's a little bit of that monetization actually pushing volatility lower. So I think it's, we'll call it not a unique view by a famous short participant. And frankly the options are pricing it in. I mean it's really expensive to have that view.

[Ed Elson]

Just when we look at the comparison between I guess the chips and so the companies that are making money off of the AI infrastructure build out boom right now versus the hyperscalers, that is the companies that are paying to build out the AI infrastructure. So Microsoft, Oracle, meta, et cetera. It's quite interesting because the hyperscalers have gotten punished so far this year. The infra players, the chip stocks, the memory stocks, those are crushing, which makes sense. Those are the companies that are really making all the money. But it seems like it's starting to reverse now and it seems like the two are now starting to converge. I'd be curious to get your views on what this all means to what we're Seeing for the hyperscalers also for a lot of the software names which are also bouncing back this week. Are those two dynamics related, do you think?

[Zed Francis]

Meta's announcement last week, I think definitely we'll call it, spooked some folks in terms of the rotation now that they're going to lease out some of their excess capacity. Maybe it's not excess capacity, but I think that's how the market's currently reading it. But what I think is interesting for the long term where both might be successful is again it's an asset heavy build out, it's a capital intensive build out and these hyperscalers are investment grade, can borrow essentially as cheaply as anybody out there. So if you're somebody like meta, borrowing at 30 year, you know, term at six, six and a half percent to be able to facilitate leasing out this compute to other players that cannot essentially lock in any sort of term in terms of financing, especially not at those rates. They're just kind of doing a, we'll call it grown up type of play where they're using their balance sheet to go ahead and build out capacity that other players cannot and hopefully capturing that spread over the long term. So you know, I view it as this is a next phase for the hyperscalers. They're building things out for themselves now. They're coming up with additional ways to increase free cash flow from this spend on infrastructure build. But it doesn't necessarily mean the semiconductor wave is over. In fact it might elongate it depending on how successful they are, essentially becoming that intermediary between the end client and building out that infrastructure.

[Ed Elson]

All right, Zed Francis, CIO and co founder at Convexitas. Zed, thank you so much.

[Zed Francis]

Thanks for having me.

[Ed Elson]

Over the past couple of years we've heard many projections about what AI may do to the labor market. According to Dario Amadei, AI could eliminate, quote, half of all entry level white collar jobs. And according to Sam Altman, entire job categories would be, quote, totally, totally gone. But lately business leaders have been changing their tune. Sam Altman recently told cnbc quote, our industry underestimated how much we're going to be able to keep people at the center of everything. And Amade said that companies, quote, can now do more with the same amount of resources. A recent paper from economists at Revelia Labs and RAMP seem to back that up. They found that the biggest AI spenders are hiring more than non adopters, not less. So we wanted to speak to one of the authors of that study and ask the question, is AI coming for Our jobs or not. Joining us is Ara Kharazian, lead economist at Ramp. Ara, thank you so much for joining us on the show. So the change of tune has been notable from these AI leaders. But you've been studying this. You wrote a paper on AI job loss. Your findings seem to be that AI is not killing jobs. What did you find?

[Ara Kharazian]

You know, this is a market where there's been a lot of mixed messages both as far as what businesses are hearing about how they should be using AI and what people on the job market are hearing as far as how AI is going to affect them. So this was the first ever study to use firm level data on what happens at firms that adopt AI. So instead of using surveys or AI exposure scores, we could look at these firms bought AI, these firms didn't. What happened to their hiring going forward? So what we found is the firms using AI heavily grew their headcount 10% over the two years following adoption. The firms that used AI lightly only grew at about the same rate as the control group. So we can talk exactly about what that means, high intensity versus low intensity. Beyond that, we also found outsized growth for entry level hires. So 12% over two years.

[Ed Elson]

Yeah. Could you explain, break down those, the difference in those control groups there, what it means to sort of be adopting it heavily versus lightly?

[Ara Kharazian]

Well, what we found in our research is that the vast majority of firms that are using AI are using it in a fairly simple and basic way. You might have a ChatGPT subscription across your entire employee base, but I think if you've used that technology, you will notice very quickly. It's nice to have, but it's not particularly productivity enhancing. It's certainly not the technology that's going to vault the United States into the next generation of economic growth and development. However, the top third of firms in our data set are using AI a little bit differently. They're more likely to use multiple models, they're more likely to use the coding agents, they're more likely to use dedicated software for a specific task like customer service operations. So those businesses tend to be assigned to our high intensity group. And it's there that we see these concentrated gains. Now, there are caveats with that too. Those firms tend to be in high growth sectors already now. Their growth accelerates after AI adoption. And while we have not seen any declines in hiring outside of these high growth firms, we're also not necessarily seeing a lot of gains from AI adoption. So that's, I think, the open question for research going forward whether or not the Gains to AI adoption will be experienced by firms outside of tech.

[Ed Elson]

That seemed to be one of the interesting points is that we're seeing this specifically in the tech sector. We don't quite know how it works for the rest of these sectors and how generalizable it is. One critique that I'd be interested to get your views on is that, I mean, it doesn't seem that this is proving causality here, meaning it could be that if you're a company that is heavily adopting AI, maybe your headcount is growing for other reasons, such as maybe you're a venture backed company and you're forced to continue to grow. I know a lot of AI startups that are growing significantly and in that sense I'm not sure it necessarily surprises me or disproves the long term thesis, which is that AI could ultimately reduce jobs, at least in the short, to maybe medium term. How do you respond to those critiques and do you think that they are valid?

[Ara Kharazian]

It's valid in that we do actually control for it. So we run the analysis on two different control groups. One, we run it against firms that use AI versus firms that never use AI at all. And there we see a very similar dynamic to what you're describing. Firms that use AI are notably different than firms that don't even try it. They're already faster growing, they're more likely to be VC backed or in tech. So instead we run this analysis against a different control group, firms that have not adopted AI yet in that period, but eventually do. And there we're able to compare firms that are on otherwise similar growth paths. So in that version of the analysis, which is how we primarily report our results, we do see that firms that adopt AI are already faster growing, but that growth accelerates after adoption. Now, to the extent that we can or cannot prove causality, I do think the step for further research is to now identify what are the mechanisms by which firms see this growth. One of the very complicated facts about measuring AI adoption is that the firms that are using it well have no incentive to publish their playbook. They have no incentive to tell us what they're doing that is different and allows them to grow faster than their competitors. Now, we do have some signs of that starting to emerge in our data. Specifically that they're using some of the more advanced tools and that they're using multiple models and they're experimenting heavily. But that's the question for future research.

[Ed Elson]

For my own understanding too, the control group that, as you explain, adopted AI later compared to the earlier AI adopters who are presumably VC backed and probably their whole group thing is to they're probably AI startups really that are growing really significantly. And you're saying, well, we found a batch of companies that adopted it later. To what extent does that still control for the problem, I guess would be my question, because could it not also be possible that many of those companies are also VC backed and they simply started adopting AI later? Is the assumption that companies that adopt it later are late stage companies, companies more similar to like, you know, a large enterprise like a Microsoft or an Oracle or whatever?

[Ara Kharazian]

Well, not necessarily. I mean, when we're measuring the growth in the high intensity adopters group, we are comparing them to firms that otherwise adopt with high intensity in the post adoption period. And so we are comparing firms that are like for like both in terms of their likelihood of being VC backed, in terms of what they sell and whether or not they're tech companies already and whether or not they're startups. So I do think our analysis is robust to that. But even so, if you compare, if you take out all of the tech companies in our data set, you still find high intensity users of AI. And while they're not necessarily adding a lot of headcount, they're also not cutting headcount. And so I think that's the finding that is informative. As far as if you're someone on the job market and you are seeing these statements from CEOs saying that we have to do all these layoffs because of AI, I would say that you should be skeptical because our data shows that if they're not necessarily hiring more, their headcount is not necessarily changing because of AI adoption. Now the mix of jobs may change going forward, but overall the companies that adopt AI tend to grow faster following adoption.

[Ed Elson]

So when you see some of these headlines that we've seen, obviously now we're seeing sort of a change in the messaging from these AI leaders, presumably because it's gotten so unpopular, I think, but I mean, jury's still out on that. But we have seen a lot of headlines from companies. They say, you know, we're going to lay off 8,000, 10,000 people and we're doing it because of AI. In fact, 100,000 layoffs so far this year have been attributed to AI. Do you think that they are kind of misattributing what the layoffs are really all about and just saying something to, I don't know, please, shareholders. I mean, what do you make of those headlines?

[Ara Kharazian]

Look, I can't surmise exactly why companies say it. I mean, I think the fact is that most layoffs and job losses in general are due to a number of reasons. I mean, there's a side finding in this paper as well that the tech sector in general was shedding jobs in the post pandemic over hiring period and that that trend reversed following AI adoption at these firms. We didn't spend a lot of time on that in the paper, but there's a side finding there that is AI that actually allowed the tech sector to start hiring again and that powered that boom. So there's that. And not to say that there won't be people who will be affected by AI in terms of their job prospects. Let's remember this paper only focused on white collar work. So we're not looking at the impact of self driving cars on drivers or manufacturing automation on factory workers. And then second, even within white collar work, there are still workers that may be affected by AI in the workplace and whether or not they're able to learn it quickly enough. Now, after we found this broad growth in hiring following adoption, the next test of the research is, well, who are they hiring and what are they hiring differently? And although I think it's too soon to say, on many metrics and job functions, we are seeing an outsized growth in entry level hiring. And you can imagine that that is, first of all, that's being powered by the firms that are high intensity adopters of AI. It's our first sign that they're looking for different types of employees, likely people who already know how to use AI and use it well. And what better place to look than recent grads, people in college and people who know how to use AI and new technologies well, as has been the case for all of the other technological developments over the last 20, 30, 40, 50 years.

[Ed Elson]

All right, Ara Kharazian, lead economist at Ramp. Ara, we really appreciate your time. Thank you.

[Ara Kharazian]

Thanks for having me.

[Ed Elson]

Okay, that's it for today. This episode was produced by Claire Miller and Alison Weiss and engineered by Benjamin Spencer. Our video editor is Brad Williams. Our research team is Dan Shalon, Kristin o' Donoghue and Mia Silverio. And our social producer is Jake McPherson. Thank you for listening to ProfG Markets from ProfGu. If you liked what you heard, give us a follow. I'm Ed Elson. I will see you tomorrow.

[Steve Feldstein]

Hi, Ryan Reynolds here for Mint Mobile. Are you looking for a beach read this summer? May I suggest your big wireless bill? It's got suspense, mystery, a slightly flat emotional arc, and a shocking twist where you realize you've been overpaying the entire time. Fortunately, though, Mint's story is better. Every plan $15 a month, even unlimited. That's it. Happy ending.

[Ed Elson]

Zero tears.

[Steve Feldstein]

Give it a [e-mail] Switch upfront payment

[Commercial Announcer]

of $45 for three months, $90 for six months, or $180 for a 12 month plan required $15 per month equivalent taxes and fees. Extra initial plan term only greater than 50. Gigs me slow when network is busy.

[Steve Feldstein]

See terms Best thing that's ever happened to you.

[Commercial Announcer]

Financially.

[Zed Francis]

Go easy. Sold my car on Carvana. Amazing offer, really. I hit 200 on a scratcher. Did the scratcher come to your house and hand you a check?

[Ara Kharazian]

No.

[Zed Francis]

How many scratchers did you hit to get that? I hit a button on Carvana.com once. Okay, that's fair. It's like the lottery, except you always win. Not like the lottery at all, actually. Exactly. Inexplicably good offers worth bragging about.

[Ed Elson]

Sell your car today on Carvana. Pickup fees may apply.

[Commercial Announcer]

The grill is shot. The chairs are held together by optimism. And what happened to the rug? Sounds like your outdoor setup is not ready for patio season. Fix it all with Wayfair. Shop Wayfair for grills, rugs, furniture, and more. With 20 million 5 star reviews, Room of choice delivery and expert setup on qualifying orders, it's never been easier to do more for less. Get 10% off your first eligible purchase. Hurry to Wayfair.com or download the app now.

[Ed Elson]

Wayfair Every style, Every home.