Frazis Capital discusses AI trade unwinding, semiconductor valuation dislocations, and liquidity risks from equity issuance.
Summary
Speakers attribute the recent market turbulence to a massive hedge fund deleveraging event triggered by rising credit spreads and financing costs, rather than fundamental demand destruction. They highlight that while neo-clouds face structural fragility, hyperscalers like Google continue to generate strong returns on AI capital expenditure. The episode notes significant valuation compression in semiconductors and Australian healthcare, presenting potential opportunities despite ongoing liquidity drains from large-scale equity issuances by companies like Anthropic.
Episode sections
- 0:00 – 1:26
Market Turbulence and AI Trade Unwind
Discussion of market turbulence triggered by AI firm IPI failure, leading to hedge fund deleveraging and index drops of 30%.
- 1:26 – 2:23
Real-Time Data vs. Market Sentiment
Analysis of strong real-time data like GPU yields and Anthropic's cash retention despite the sell-off.
- 2:23 – 4:30
Potential Oversupply and Financing Risks
Risks of GPU oversupply if financing fails for pre-booked orders and VC spending pulls back.
- 4:30 – 9:13
Hyperscaler Strength and Valuation Dislocation
Bullish case based on hyperscaler cash funding and Google Cloud growth, contrasting with beaten-down memory stocks.
- 9:13 – 16:08
Hedge Fund Deleveraging and Market Structure
Structural shift where hedge funds cleared two years of AI positions in weeks, with retail as counterparty.
- 16:08 – 18:42
Valuation Opportunities in Semiconductors and Tech
Opportunities in semiconductors trading at 10x P/E and memory stocks at 3.7x forward P/E after deleveraging.
- 18:42 – 22:42
Political Risks and Regulatory Delays
Risks from political shifts in US and Australia potentially delaying data center construction and impacting demand.
- 22:42 – 25:15
Australian Market and Healthcare Opportunities
Broad sell-off in ASX tech and healthcare creating opportunities in beaten-down compounders like Intuitive Surgical.
- 25:15 – 26:17
Semiconductor Selloff and Historical Parallel to Software
Comparison to 'SaaSpocalypse,' arguing profitable software companies rallied 2-3x from lows despite narrative collapse.
- 26:17 – 27:02
Hyperscaler Funding and Semiconductor Recovery
Bet on semiconductor recovery driven by sustained GPU demand and Google's continued high CapEx spend.
- 27:02 – 27:34
Portfolio Shift to Healthcare and ASX
Interim focus on 'smashed' healthcare compounders and ASX opportunities while AI trade unwinds.
- 27:34 – 27:50
Neo-Cloud Profitability Challenges
Neo-clouds must prove growth and move from negative margins to profitability to gain credit market credibility.
- 27:50 – 28:21
Liquidity Crunch from Equity Issuance
Risk of liquidity drain from massive equity issuance by hyperscalers and neo-clouds like Anthropic.
- 28:21 – 30:43
Market Impact of Capital Raising
Hundreds of billions in equity issuance forcing asset liquidation to fund investments, creating supply glut.
Assets discussed
| Asset | Kind | Stance | Context |
|---|---|---|---|
| NVIDIANVDA | equity | neutral | For many months now, credit spreads for the neo clouds and the capital-intensive parts of the industry were rising significantly, but also big tech, hyperscalers, even NVIDIA also rising. |
| Meta PlatformsMETA | equity | neutral | Meta said that they're still improving their ads. |
| AlphabetGOOGL | equity | neutral | Google had some results. |
| INTUITIVE SURGICAL INCISRG | equity | bullish | But we bought intuitive surgical recently. |
| BOSTON SCIENTIFIC CORPBSX | equity | bearish | whether it's your Nikes or your Boston Scientific, like those kinds of things, like credit scoring companies, all those things that like a whole species of fund would specialize in, they're all down significantly, like 60-70%. |
“Stance” is how the speakers talked about the asset, not a recommendation.
Claims worth verifying
The failure of firm IPI triggered the market sell-off and marked the end of equity market risk tolerance.
Unidentified speaker ·0:17
How to verify: Search financial news archives for 'IPI failure' and 'market sell-off trigger' around the date of the podcast (2026-11-06) or recent prior months.
Credit spreads and the cost of finance debt had risen to prohibitive levels, over 15% for some issuers.
Unidentified speaker ·0:31
How to verify: Check credit spread data for high-yield issuers or specific neo-cloud companies in the months leading up to November 2026.
Many companies are down 60-70%, and the index is down 30% due to hedge fund deleveraging.
Unidentified speaker ·0:45
How to verify: Compare current market indices and individual stock prices to their peaks from earlier in 2026 to verify the percentage drawdowns.
GPU rental yields and DRAM prices were rising in real-time during the sell-off.
Unidentified speaker ·1:42
How to verify: Check real-time cloud computing pricing data (e.g., from cloud providers or rental platforms) and DRAM spot price indices (e.g., from TrendForce) for the period.
Anthropic raised over $140 billion and reported having $130 billion in cash remaining.
Unidentified speaker ·2:03
How to verify: Review Anthropic's public financial disclosures, press releases, or credible news reports regarding their funding rounds and cash position.
Credit spreads for neo-clouds, capital-intensive industry parts, big tech, and NVIDIA were rising for many months.
Unidentified speaker ·2:34
How to verify: Analyze credit default swap (CDS) spreads or bond yields for these specific entities over the 6-12 months prior to the podcast.
Memory providers have locked up substantial commitments for many years, and optical company Momentum is booked out for two years.
Unidentified speaker ·2:56
How to verify: Check annual reports or press releases from major memory providers and Momentum (or its parent/related entity) regarding long-term supply agreements.
Google Cloud grew over 100% year-on-year, and Google spent $50 billion on CapEx in the quarter.
Unidentified speaker ·5:08
How to verify: Review Google's (Alphabet) quarterly earnings reports for the quarter preceding November 2026.
SK Hynix and Samsung are down 50-60%+, while Micron is down 30-35%.
Unidentified speaker ·6:10
How to verify: Check historical stock price data for SK Hynix, Samsung Electronics, and Micron Technology to calculate percentage declines from recent highs.
Hedge funds moved from a massive overweight to net negative position in AI-related investments, clearing two years of buying in weeks.
Unidentified speaker ·
How to verify: Review hedge fund positioning data from sources like 13F filings, fund manager surveys, or quantitative hedge fund exposure reports.
During the 'SaaSpocalypse,' good software companies went up 2-3x from their lows despite negative narratives.
Unidentified speaker ·25:15
How to verify: Check historical stock performance data for major SaaS companies (e.g., via Bloomberg or Yahoo Finance) between the peak of the SaaS selloff and subsequent recovery highs.
Software company profitability increased significantly due to efficiencies from AI (AR).
Unidentified speaker ·25:45
How to verify: Review quarterly earnings reports and margin data for major software companies during the period referenced to identify profit growth trends.
Google spends approximately $50 billion per quarter on AI infrastructure.
Unidentified speaker ·26:25
How to verify: Check Google's (Alphabet) quarterly capital expenditure disclosures in SEC 10-Q filings.
Neo-cloud companies have margins of -24% and need to become profitable to access credit markets.
Unidentified speaker ·27:38
How to verify: Review financial statements and credit rating reports for neo-cloud companies like CoreWeaver or Nebius to verify current margin structures and credit access conditions.
Anthropic is raising $100 billion.
Unidentified speaker ·28:10
How to verify: Check press releases or financial news reports regarding Anthropic's fundraising rounds and total capital raised.
Hyperscalers like Google and Meta have shifted from stock buybacks to equity issuance.
Unidentified speaker ·28:06
How to verify: Compare historical share repurchase data with new equity issuance data in SEC filings (e.g., 10-K, 10-Q) for Google and Meta.
Interest rates have peaked based on bond yields in the US and Australia.
Unidentified speaker ·29:55
How to verify: Analyze US Treasury yield curves and Australian Government Bond yields for trend reversals or peak levels in late 2026.
Do your own research
- Verify current credit spreads for neo-clouds and hyperscalers using Bloomberg or Credit Suisse data to assess financing risk.
- Check Google's latest quarterly earnings report for Cloud revenue growth rates and CapEx guidance.
- Review Anthropic's financial filings to confirm cash balances and burn rate relative to the $140 billion raise.
- Analyze SK Hynix and Samsung's forward P/E ratios and contractual obligations with hyperscalers to evaluate valuation dislocation.
- Monitor US and Australian bond yields to assess if interest rates have peaked as claimed.
- Investigate recent equity issuance pipelines from major tech firms to quantify liquidity drain.
- Review political developments in Texas and Australia regarding data center construction regulations and delays.
Full transcriptexpand
- 0:00
- Unidentified speaker: Hey Liana, how are you doing?
- 0:01
- Unidentified speaker: I'm both thanks.
- 0:02
- Unidentified speaker: How are you, Mike?
- 0:03
- Unidentified speaker: Good.
- 0:04
- Unidentified speaker: Obviously, been uh even more turbulent than usual in financial markets lately.
- 0:08
- Unidentified speaker: Absolutely.
- 0:09
- Unidentified speaker: And why do you think that's been?
- 0:11
- Unidentified speaker: Why do you think that this is coming on this time of the year?
- 0:14
- Unidentified speaker: Do you think the market knows something and it's catching up?
- 0:16
- Unidentified speaker: Yeah, it's a hard one.
- 0:17
- Unidentified speaker: I mean, all triggered that that firm has failed IPI.
- 0:19
- Unidentified speaker: So I remember we're talking, half joking, like could that really derail the entire global AI trade?
- 0:25
- Unidentified speaker: It seems like it has.
- 0:26
- Unidentified speaker: Um I mean, it basically marked the equity market shut for that kind of risk.
- 0:31
- Unidentified speaker: And separately, CDS and credit spreads and the cost of finance debt had been rising basically to prohibitive levels, you know, 15% plus for some issuers.
- 0:41
- Unidentified speaker: And for whatever reason, that that triggered a full tailspin.
- 0:44
- Unidentified speaker: There's been a real liquidation.
- 0:45
- Unidentified speaker: You know, a lot of companies are down 60, 70%, index is down 30%.
- 0:50
- Unidentified speaker: There's some like major companies that reported massive beats and raises and long-term contracts that still dropped, you know, substantially.
- 0:59
- Unidentified speaker: So it was a massive hedge fund deleveraging and liquidation.
- 1:05
- Unidentified speaker: But the real question is, is this that one in two, one in three chance where the market panics like it did in Liberation Day, where there was a big semiconductor sell-off for six, seven months before earlier this year?
- 1:17
- Unidentified speaker: Like, is this one of those kind of flash crashes?
- 1:20
- Unidentified speaker: Or has the market kind of sniffed something out?
- 1:22
- Unidentified speaker: You know, has it really sensed the change?
- 1:25
- Unidentified speaker: And what do you think that would look like?
- 1:26
- Unidentified speaker: Well, at the moment you can't see it anywhere.
- 1:28
- Unidentified speaker: I'll say that.
- 1:29
- Unidentified speaker: And I think that's what's made it so challenging.
- 1:31
- Unidentified speaker: So we thought we knew all the companies were going to beat and raise based on estimates and expectations and guidance.
- 1:38
- Unidentified speaker: That was clear.
- 1:39
- Unidentified speaker: We've also been tracking things that are real time.
- 1:42
- Unidentified speaker: So this entire sell-off GPU rental yields have been rising, lifting, DRAM has been rising, all the real-time measures have been increasing.
- 1:49
- Unidentified speaker: There was a lot of shakiness around Anthropic when they released their numbers.
- 1:53
- Unidentified speaker: But actually, those numbers were pretty good.
- 1:54
- Unidentified speaker: You know, Anthropic accelerated after that quarter end, the step quarter end, and turned out that Anthropic only had it only used like eight or nine billion of their cash.
- 2:03
- Unidentified speaker: So they'd raised like over 140 billion.
- 2:05
- Unidentified speaker: Everybody, like we all assumed that they'd spent at least a good chunk, if not most of it, and that's why they're coming back to the market.
- 2:12
- Unidentified speaker: But in reality, they still have almost all that cash, like 130 billion plus they're reported.
- 2:17
- Unidentified speaker: Um so it was it's much more cash efficient than anybody was thinking.
- 2:23
- Unidentified speaker: But what would it actually look like?
- 2:24
- Unidentified speaker: It would probably, if this if this turns into real sell-off, it'll be one of those ones where it was not really possible to see in the data.
- 2:30
- Unidentified speaker: I'd say the only clue was that credit spreads were rising.
- 2:34
- Unidentified speaker: For many months now, credit spreads for the neo clouds and the capital-intensive parts of the industry were rising significantly, but also big tech, hyperscalers, even NVIDIA also rising.
- 2:44
- Unidentified speaker: So if you look back on this, I'd say the company's performance weren't able to see it there.
- 2:50
- Unidentified speaker: In fact, we see a lot of companies that are 25% bigger quarter on quarter, so just over three months.
- 2:56
- Unidentified speaker: For the first time, memory providers have locked up substantial amounts of their commitments over many, many years.
- 3:02
- Unidentified speaker: In opticals, companies like Momentum have said they're booked out for the next two years, they're fully at capacity.
- 3:08
- Unidentified speaker: There was no clue there.
- 3:09
- Unidentified speaker: The only clue was really credit spreads.
- 3:11
- Unidentified speaker: So, how could it actually unwind?
- 3:12
- Unidentified speaker: It would probably be something like what happened in crypto, where at some point of the next three to six months, it turns out that there is more supply than people thought.
- 3:21
- Unidentified speaker: And the people that were buying aren't actually able to finance those purchases.
- 3:25
- Unidentified speaker: So, to use firms as an example, let's say they obviously had NVIDIA e-marked a huge amount of GPUs for them for the next generation for next year.
- 3:34
- Unidentified speaker: If they're unable to get financing, and it looks like they won't be able to get financing because the IPO is the route to that, that's what unlocked all the debt and unlocked all their plans.
- 3:43
- Unidentified speaker: Maybe that supply comes back to the market, and maybe the market is more oversupplied than people think.
- 3:47
- Unidentified speaker: Maybe there's a pullback in VC spending.
- 3:51
- Unidentified speaker: So VC companies are very well funded, raising large amounts, sometimes even billions of dollars, to basically spend on AI and GPUs and things like that.
- 3:59
- Unidentified speaker: If there was a pullback from the VCs and a pullback from the startups that they're funding, it could all happen at once and there'd be an oversupply, and people wouldn't be able to finance the purchases that they want.
- 4:10
- Unidentified speaker: And you'd go from this period where you've got massive oversubscribed order books to, you know, people trying to get rid of their excess GPUs.
- 4:18
- Unidentified speaker: But I would say even that, like you can see how it happened, and there's many examples of that in GPUs over the last five years, certainly seven or eight years, with those multiple crypto booms where the same thing kind of happened.
- 4:30
- Unidentified speaker: It still seems like a stretch.
- 4:31
- Unidentified speaker: And the thing that kept us bullish is most of the vast bulk of these GPUs in the CapEx is done by hyperscalers who are mostly funding it by cash and at the very least have very profitable businesses.
- 4:42
- Unidentified speaker: Like it's much better for the GPUs instead of sitting in a very fragile brand new company in Australia like Fermos, it's much better if that capacity sits on the balance sheet of a hyperscaler that has customers and actual use of that cash flow, actual use of those GPUs in their internal business.
- 4:57
- Unidentified speaker: Meta said that they're still improving their ads.
- 5:00
- Unidentified speaker: The loss curves are still improving when they're trained on serving ads.
- 5:03
- Unidentified speaker: They're still making money from just applying GPUs to their core business.
- 5:06
- Unidentified speaker: Google had some results.
- 5:08
- Unidentified speaker: You know, Google Cloud's grown over 100% year on year.
- 5:10
- Unidentified speaker: They're getting a real return from that.
- 5:12
- Unidentified speaker: And, you know, Google spent 50 billion on CapEx in the quarter.
- 5:15
- Unidentified speaker: So maybe 200 of the 600 billion plus is Google.
- 5:20
- Unidentified speaker: Also, Anthropic is funded.
- 5:22
- Unidentified speaker: You know, they've got over 100 billion cash now.
- 5:23
- Unidentified speaker: And if they raise another 100 billion, they're funded for their CapEx plans as well.
- 5:28
- Unidentified speaker: So the funding, if you step outside from the neo clouds, the funding, and ultimately it was the hyperscalers that were the main customers of the neo clouds.
- 5:37
- Unidentified speaker: The money was there.
- 5:38
- Unidentified speaker: There's real businesses behind this, behind these purchase orders.
- 5:41
- Unidentified speaker: But that's not what the market's saying.
- 5:44
- Unidentified speaker: The market is pricing now a major crash.
- 5:46
- Unidentified speaker: A good example of that is saying memory, where SK Heinex is a 4P of two types.
- 5:51
- Unidentified speaker: Micron in the United States is probably gonna do 50 billion of gap profit quarter.
- 5:57
- Unidentified speaker: So that's 200 billion over the year, and probably for the next two years.
- 6:00
- Unidentified speaker: And they're saying they're booked out and contractually these obligations and they're with hyperscalers.
- 6:05
- Unidentified speaker: So with good credits, not bad credits.
- 6:07
- Unidentified speaker: But nevertheless, you know, those those stocks are down significantly.
- 6:10
- Unidentified speaker: I think Micron's only down 30-35%.
- 6:12
- Unidentified speaker: But SK Heinick, Samsung, they're down 50-60% plus, probably more now.
- 6:18
- Unidentified speaker: So the memory providers are definitely pricing in a downturn, and that's something that I guess could play out.
- 6:25
- Unidentified speaker: Maybe this time next year, actually memory prices are falling.
- 6:27
- Unidentified speaker: Maybe Chinese supply hits the market and there's like a pullback in VC funding.
- 6:32
- Unidentified speaker: Actually, in the VC funding, what I think actually is happening is it feels like the market is out for blood.
- 6:36
- Unidentified speaker: It needs a scalp.
- 6:37
- Unidentified speaker: So in the first sell-off in June, there was obviously Leopold Ashenbrenner and situational awareness.
- 6:43
- Unidentified speaker: And once that was sorted, once that blow-up had taken place and that risk had been cleared, then markets rallied.
- 6:48
- Unidentified speaker: It almost feels like the market is trying to sniff out and trying to force somebody like Anthropic to delay their IPO.
- 6:54
- Unidentified speaker: Anthropic is very different to Fermos.
- 6:56
- Unidentified speaker: You know, this is a company with real technology, already 100 billion plus on their balance sheet.
- 7:00
- Unidentified speaker: They invented clawed code and agentic engineering.
- 7:03
- Unidentified speaker: It's a real business with real customers.
- 7:06
- Unidentified speaker: And once you start using this stuff, you're never going back.
- 7:08
- Unidentified speaker: I think that's the other thing that's been hard, is obviously we're relatively AI forward.
- 7:12
- Unidentified speaker: We're in a small business, so it's easy for us to use these tools.
- 7:15
- Unidentified speaker: Once you get it all set up and you can really see how these models can use your own computer to do certain things, the power is immense and there's no going back.
- 7:23
- Unidentified speaker: And the amount of things that we do use it for, every time we do a request, there's GPUs spinning up in some data center somewhere.
- 7:30
- Unidentified speaker: And you know, we basically need claw code for our software.
- 7:34
- Unidentified speaker: I'm using GrockBot almost every day.
- 7:36
- Unidentified speaker: I feel like that's a huge unlock in terms of what that can do.
- 7:39
- Unidentified speaker: If you get that really humming, it's huge efficiencies.
- 7:43
- Unidentified speaker: And there's no going back on that spend.
- 7:45
- Unidentified speaker: That's real money that I'll probably be paying for the rest of my life.
- 7:48
- Unidentified speaker: Yeah, it's a challenge.
- 7:50
- Unidentified speaker: It feels really surprising because I feel like all businesses and even a lot of individuals on a personal level have adopted using AI into their daily lives.
- 7:58
- Unidentified speaker: There's been a huge difference between companies who have streamlined things in a productive way over the last couple of years.
- 8:05
- Unidentified speaker: So it still feels very surprising that the downturn would be happening for it in the market.
- 8:11
- Unidentified speaker: And I guess I'm wondering where do you see some sort of reversal happening and when and why.
- 8:17
- Unidentified speaker: Yeah, I mean, it ultimately comes down to are people actually paying for it?
- 8:20
- Unidentified speaker: Like, is that end user there?
- 8:21
- Unidentified speaker: You know, effectively the hyperscale is investing huge amounts and getting very good returns.
- 8:27
- Unidentified speaker: Like the cloud businesses, I mentioned Google up 100%, you know, they've already accelerated.
- 8:32
- Unidentified speaker: But you still need somebody to use that compute, and it's the large language models that are the main users of that.
- 8:37
- Unidentified speaker: And even if they have in customers that are paying them, which are real businesses and real consumers, mostly enterprise.
- 8:42
- Unidentified speaker: And even if they stumble and people start using open source more, which is definitely a major trend, that still needs to be hosted somewhere, and you're going to host that on one of the hyperscalers.
- 8:52
- Unidentified speaker: It also feels like AI adoption, it feels like people are using it, but the vast majority of people are not.
- 8:57
- Unidentified speaker: The vast majority of enterprises are not using it.
- 8:59
- Unidentified speaker: Like the amount of people that I see who are quite ticked forward aren't quite aware of what the latest models in the last few weeks even are capable of and how much they've improved and how much value they can add.
- 9:09
- Unidentified speaker: There's probably an element where the market's got ahead of themselves.
- 9:13
- Unidentified speaker: I think there was a really good Ken Griffin, who's my favorite hedge fund manager.
- 9:18
- Unidentified speaker: He had a really good piece where it was kind of like he was saying you kind of got to respect the market and you can't just use hedge fund deleveraging as an excuse every time your sector gets blown up or sector you're invested blown up.
- 9:28
- Unidentified speaker: But the reality is there is a massive hedge fund deleveraging event at the moment.
- 9:32
- Unidentified speaker: It may be well justified, but there's no doubt that on the day-to-day, week to week of the last few weeks, it's gone from one of the biggest overweights to net negative.
- 9:41
- Unidentified speaker: So hedge funds as a group, obviously there's other market participants, but hedge funds have gone from having a massive overweight to AI-related investments to actual underweight.
- 9:50
- Unidentified speaker: They cleared almost two years of buying in a few weeks.
- 9:54
- Unidentified speaker: So that's actually what's been driving the price.
- 9:56
- Unidentified speaker: It looks like retail has been net buying, taking the other side, but the hedge funds are definitely delivering.
- 10:02
- Unidentified speaker: And I think this is the first time I found like with our models, the last three years, like leading up into this, we obviously got caught in this one because we had a lot of AI exposure.
- 10:11
- Unidentified speaker: But in the last three years, our models and our approach has generally been a step ahead of the deleveraging events.
- 10:17
- Unidentified speaker: It's kind of happened, and then the market would come.
- 10:19
- Unidentified speaker: I'd say this one seemed to be much, much faster.
- 10:21
- Unidentified speaker: And I feel like that's even interesting.
- 10:23
- Unidentified speaker: I wonder if like maybe the success of that approach.
- 10:26
- Unidentified speaker: Effectively, we're ahead of the CTAs and ahead of the Bridgewaters and the Vold Targeters.
- 10:30
- Unidentified speaker: They were moving much slower.
- 10:31
- Unidentified speaker: Our signals were picking up moves before theirs.
- 10:34
- Unidentified speaker: This one moved faster.
- 10:35
- Unidentified speaker: And there's obviously been a massive, I guess, extinction event in funds.
- 10:40
- Unidentified speaker: Firstly, Australia, but we'll come back to that, but also in terms of single managers.
- 10:43
- Unidentified speaker: Where single managers often picked a bunch of stocks.
- 10:46
- Unidentified speaker: When it's a very narrow market, if let's say you've got a bunch of managers who randomly pick their top five to 15 stocks, the vast majority of them can simply underperform if it's only a handful of companies winning.
- 10:56
- Unidentified speaker: And basically for the last year and a half, only AI and semiconductors have been winning.
- 11:00
- Unidentified speaker: Everything else has been significantly underperforming, whether it's quality, whether it's Aussie companies, even healthcare was pretty challenged at various periods.
- 11:07
- Unidentified speaker: Your healthcare compounders all down significant substantial amounts, pull up a chart of Boston Scientific or something like that, and see huge moves in otherwise stable companies.
- 11:17
- Unidentified speaker: Oh yeah.
- 11:17
- Unidentified speaker: So this event, basically, what's happened is more and more capital is gone passive, more and more capital has gone to pot shops like your millenniums, your citadels.
- 11:26
- Unidentified speaker: And it's more than it's a huge amount of capital that those firms are running.
- 11:32
- Unidentified speaker: But in terms of like volume in the market, it's way higher than that because they're extremely active.
- 11:37
- Unidentified speaker: They trade every quarter.
- 11:38
- Unidentified speaker: They typically trade around earnings because that's when there's sufficient liquidity to make hundreds of millions of dollars if that that's the size of their position.
- 11:46
- Unidentified speaker: And it feels like that has been a big feature.
- 11:48
- Unidentified speaker: So once these companies were down 10, 20, 30 percent, anybody in those pods has to effectively close it.
- 11:55
- Unidentified speaker: And the market was effectively kind of long semis and short everything else.
- 12:00
- Unidentified speaker: And then so as that unwind has happened, you've literally seen semis collapse and everything else kind of lift.
- 12:06
- Unidentified speaker: Yeah, I'd say both probably one in three times, one in two to one in three times, these things happen, and then it's actually a huge buying opportunity if you take the other side.
- 12:17
- Unidentified speaker: I'd say in my career, in 15 years, only 2022 is the only example where it was actually the start of something rather than the end.
- 12:24
- Unidentified speaker: And it does feel like there's a bit of meat on the bone here.
- 12:27
- Unidentified speaker: And I think the meat on the bone here comes back to that one clue that we could have paid attention to, or shouldn't or we noticed it, but we're weighting dings like GPU rental prices, like earning speeds, like long-term contracts, overfilled order books, capacity booked out for years in advance.
- 12:43
- Unidentified speaker: We thought those things were more important.
- 12:45
- Unidentified speaker: But really, the CDS and the credit markets were showing distress long before anything else.
- 12:50
- Unidentified speaker: And credit's like more complicated.
- 12:52
- Unidentified speaker: If you have a half-built data center and there's a credit issue and they can't get financing, that takes a long time to sort out.
- 12:59
- Unidentified speaker: That's not something that can be easily fixed.
- 13:01
- Unidentified speaker: It's just a mess.
- 13:02
- Unidentified speaker: You know, it's a mess that might take years to play out.
- 13:05
- Unidentified speaker: And if GPUs are being used as security, the data center's being used as security, but it's half-built, and you've got multiple, you know, these are all in SPVs with different lenders, obviously, it's credits, every contract's different.
- 13:15
- Unidentified speaker: It just takes a long time to unwind these things.
- 13:17
- Unidentified speaker: And so if the end demand really isn't there, or is somewhat less than what people are expecting, or even if it's there, but the financing isn't there for whatever reason, interest rates are very high.
- 13:28
- Unidentified speaker: Maybe it's just reached a point where the economics don't work for the people that want to put the money in.
- 13:33
- Unidentified speaker: The base rate's 5%.
- 13:34
- Unidentified speaker: And these are risky projects, a lot of execution risk, regulatory risk, like people are protesting, there's huge political risk, which obviously there were just elections in the United States that we could chat about.
- 13:44
- Unidentified speaker: All that could actually push interest rates to a point where it doesn't make sense to do these projects, even though there is demand.
- 13:50
- Unidentified speaker: There's always been a catch to these.
- 13:52
- Unidentified speaker: There's always been a flaw in the argument, these neoclouds.
- 13:54
- Unidentified speaker: They're always saying how profitable they are.
- 13:55
- Unidentified speaker: But if you're that profitable, why don't you find yourself with equity?
- 13:58
- Unidentified speaker: You know?
- 13:59
- Unidentified speaker: And I feel like that's the key.
- 14:00
- Unidentified speaker: If you're going to pay back a GPU in a year, why are you borrowing to do that?
- 14:03
- Unidentified speaker: Just raise the equity, get your 100% return, and then you've got a GPU spitting out tokens for years and years afterwards at whatever price.
- 14:11
- Unidentified speaker: The marginal price of electricity, like a capital cost is gone.
- 14:14
- Unidentified speaker: These should be equity funded.
- 14:15
- Unidentified speaker: And actually the reason they're not is because the returns, all inclusive returns, aren't actually that high.
- 14:21
- Unidentified speaker: And what they need to do is they need to raise equity and debt.
- 14:24
- Unidentified speaker: But specifically, equity, they need to promise a dream and a future growth.
- 14:28
- Unidentified speaker: And to get that future growth, they can't just use equity, they need to use debt.
- 14:31
- Unidentified speaker: Kind of like Fermits.
- 14:32
- Unidentified speaker: Like Fermits were like, we got seven billion, but they wanted to buy build tens of billions of dollars, if not hundreds of billions of data centers.
- 14:38
- Unidentified speaker: They had to promise that growth.
- 14:40
- Unidentified speaker: And the only way they could do that was to do it with debt.
- 14:43
- Unidentified speaker: It would be an interesting thought experiment if they just did a seven billion dollar project, purely equity financed, if that would have got away.
- 14:50
- Unidentified speaker: But then if you're doing that, then where's your equity?
- 14:52
- Unidentified speaker: You know, they're basically saying we're gonna raise seven or five billion USA of cash, and actually we're worth 30 or 40 billion.
- 14:58
- Unidentified speaker: Like the 5 billion equity cash, 7 billion Aussie roughly, like that's the rule, that's the real value.
- 15:03
- Unidentified speaker: The difference between that and the extra 30, 40 billion dollars of equity value, that's the Fugazi.
- 15:08
- Unidentified speaker: That's real stuff.
- 15:09
- Unidentified speaker: You can only justify that if you say there's gonna be immense growth and on top of what we're gonna get from the equity we put into this project.
- 15:16
- Unidentified speaker: And that was a step ultimately the market just couldn't take, and rightfully so.
- 15:20
- Unidentified speaker: And so if you did equity fund it, maybe you're doing a five bill equity project at a six or seven bill vow, and that six or seven bill is that extra one to two bill is is what goes to the people that pull the project together.
- 15:30
- Unidentified speaker: I think that's the kind of thing that would have could have worked, but yeah, it's a it's an interesting one.
- 15:35
- Unidentified speaker: It's also been a such a challenging market.
- 15:36
- Unidentified speaker: Like we had two stocks and one of them bought another.
- 15:39
- Unidentified speaker: So we own Marvel and Lementum, both sitting on like pretty substantial profits at the time.
- 15:45
- Unidentified speaker: Both are down.
- 15:45
- Unidentified speaker: Usually what happens when one of your portfolio companies buys another doesn't happen often, but when they do, you know, the one that gets bought is up 20, 30 percent or more, and the one that buys it drops by some similar amount.
- 15:54
- Unidentified speaker: The one that buys it is often bigger, so maybe the one that's being bought goes up 20, 30 percent, and then the one that's buying it drops seven or eight percent.
- 16:01
- Unidentified speaker: In this case, they both drop like 30, 40 percent.
- 16:04
- Unidentified speaker: So it's like how can you structure a deal so badly market?
- 16:08
- Unidentified speaker: And the funny thing is now this stuff looked cheap next year.
- 16:12
- Unidentified speaker: Like the forward multiples are always cheap, but now this stuff actually is starting to look genuinely cheap.
- 16:16
- Unidentified speaker: Like we're looking at companies.
- 16:18
- Unidentified speaker: Obviously, we talked about memory that's like two, three times Micron is on a current P of 10 and a forward P of like 3.7.
- 16:26
- Unidentified speaker: Like these are low.
- 16:27
- Unidentified speaker: And there's like the really exciting stuff that we like.
- 16:29
- Unidentified speaker: Like one thing, it's been good and bad for us, all up at times like this, it's very bad, but we're always drawn to the faster growing smaller companies.
- 16:36
- Unidentified speaker: So when you have some really like rock star companies that are growing 100% plus in the semiconductor space that are doing stuff that they're like the only companies doing what they're doing, not just the general so-called bottlenecks, but like people with actual technology that's unique to them, or they have their own particular advantages.
- 16:53
- Unidentified speaker: Those companies are trading now at 10 times Ps.
- 16:56
- Unidentified speaker: The valuations have really turned down, and again, I think that's been the challenge.
- 17:01
- Unidentified speaker: So valuations are cheap, order books are locked out, live GPU prices have been rising.
- 17:07
- Unidentified speaker: Everyone was hoping anthropic would be doing better than it was, but it's still doing very well.
- 17:12
- Unidentified speaker: There's still no reason not to think they're gonna reach 100-120 bill of ARR.
- 17:16
- Unidentified speaker: I used to think, like, and I've written in notes and we've talked about in private podcasts that open source is a real threat.
- 17:23
- Unidentified speaker: But for what?
- 17:24
- Unidentified speaker: If you're actually on the cutting edge of AI, and I mean in that corporate sense rather than a research sense, but if you're properly using these tools, you're not gonna trust, you need to use state of the art.
- 17:34
- Unidentified speaker: Like, there's no way I would let a Chinese company have access to a computer or something like that.
- 17:40
- Unidentified speaker: You're probably taking some level of risk with open AI and Anthropic, but they have so much to lose and they're so highly regulated that if you're gonna trust anyone, it's them.
- 17:47
- Unidentified speaker: I actually think they do if they get 200, 250 bill revenue, it's extremely high margin.
- 17:54
- Unidentified speaker: Most of their expense base is for training, which is good and bad.
- 17:57
- Unidentified speaker: On the one hand, they're gonna have to keep doing that.
- 18:00
- Unidentified speaker: On the other hand, they could tone that back at any moment.
- 18:03
- Unidentified speaker: And as these technologies diffuse out and revenue grows, perhaps the training costs can come down as a percentage while still increasing.
- 18:10
- Unidentified speaker: If that's the case, those companies are well and truly funded for this.
- 18:14
- Unidentified speaker: So you can see the trap.
- 18:16
- Unidentified speaker: You can see the trap.
- 18:17
- Unidentified speaker: Because then these same companies are down 40%, 50%, 60%, index is down 30%, and obviously that's held up by I don't know, maybe the leading companies like Taiwan Semi down 20%, 19, 20%, SML down 25%.
- 18:31
- Unidentified speaker: Anything less, they're obviously like one-of-a-kind companies which are on their own part of the supply chain.
- 18:37
- Unidentified speaker: Anything smaller than that, faster growing, is down twice that.
- 18:40
- Unidentified speaker: It's an interesting one.
- 18:42
- Unidentified speaker: But I think in these moments, you just kind of have to respect the market a little bit, because it could go either way.
- 18:47
- Unidentified speaker: And I'm no US politics expert, but for whatever reason, there's some left-wing, right-wing split.
- 18:54
- Unidentified speaker: I guess it was like, you know, Trump parading around all the AI leaders, being AI forward naturally.
- 19:00
- Unidentified speaker: And I think already I'd say the left is like less interested in progress, and obviously they're energy intensive.
- 19:07
- Unidentified speaker: Apparently they don't use much water at all, but there's like environmental concerns.
- 19:11
- Unidentified speaker: Basically, yeah, the left wing does not like AI, and they'll probably do what they can to make that a political issue.
- 19:18
- Unidentified speaker: Because I think the person on the street is generally skeptical.
- 19:21
- Unidentified speaker: And yeah, it's interesting because it's one of those ones where I don't know if it's just like the battle lines were drawn and then everybody like formed their battle lines.
- 19:32
- Unidentified speaker: You know, if Trump's pro-AI, well, we're gonna do the opposite of that, and hence that caused a split, or if it's something in that where the the conservatives are more interested in progress, big investment projects, and like winning in defense, in have in having the best possible capabilities.
- 19:47
- Unidentified speaker: Whereas there's definitely part of the left that's like degrowth, doesn't want to develop, would prefer if like if GDP went down and things like that.
- 19:56
- Unidentified speaker: Either way, the left wing just won significantly more power than they had a week ago, and one of their election issues was slowing down data centers.
- 20:04
- Unidentified speaker: And there's probably a lot of ways they can do that.
- 20:06
- Unidentified speaker: Even if they don't have the power to do it directly, you can drown things in red tape or green tape, you can just slow things down.
- 20:13
- Unidentified speaker: And then let's say, to go back to the question, how would it play out?
- 20:16
- Unidentified speaker: Let's say you have a data center that is just going to be delayed due to regulatory things, and then there's protests and they've got to do a bunch of new things to satisfy them.
- 20:24
- Unidentified speaker: I think Texas is a big one where there's big projects there, and that just swung from conservative to left, from red to blue.
- 20:33
- Unidentified speaker: Maybe all the equipment in that that if those data centers delayed a year, all the equipment that's been purchased with full intention to use will not be used.
- 20:42
- Unidentified speaker: Will either need to be reallocated or we'll just sit there waiting for it to be built and installed and approved.
- 20:48
- Unidentified speaker: I mean look at Aussie Smallcats, look at what's happened there.
- 20:52
- Unidentified speaker: I wonder if that's what's in store for semis.
- 20:54
- Unidentified speaker: So we're looking at a bunch of these, something like Life 360, it's kind of grown steadily, it's been down, it's down from roughly $50 to $20.
- 21:03
- Unidentified speaker: So a 60% draw, far more, high to low was much bigger, but now it's kind of like slowly rebounding after a year and two months.
- 21:11
- Unidentified speaker: This started in June, so already six months into this.
- 21:14
- Unidentified speaker: And so maybe the low is ahead of us, and then there's a really choppy period where the neo clouds that can't get finance put their stock back on the market or the sites get taken over by somebody else, where there's just like a long, drawn-out reworking process where all those headlines are negative, where projects are delayed, valuations stay low.
- 21:34
- Unidentified speaker: Meanwhile, the companies continue to grow, they continue to buy back shares.
- 21:37
- Unidentified speaker: Maybe the growth expectations come down when they announce that there's another leg down three to six months, like mid to end of next year.
- 21:47
- Unidentified speaker: That's kind of where you get like a very strong recovery, and you get the next two, three, four X out of these companies, which is still on the table.
- 21:53
- Unidentified speaker: We knew it was going to be choppy, but there's no doubt that every year of our lives, demand for computers has gone up.
- 21:58
- Unidentified speaker: You think about a data center, like it was kind of useful before, like obviously you store your photos, like think of us.
- 22:03
- Unidentified speaker: You store your photos in a data center, you have your Gmail, you have your worker outlook with Microsoft.
- 22:08
- Unidentified speaker: There was definitely a need for data centers, but me and you didn't need like GPUs spinning up answering questions in the same way that we do now.
- 22:15
- Unidentified speaker: And that is not going to change.
- 22:18
- Unidentified speaker: So the fact that the global economy is reorganizing so that more money is spent on AI, I think that is almost certainly here to stay.
- 22:26
- Unidentified speaker: And even in the bear market view, like we could paint a scenario where there's a real bear market downturn here, which we just went through.
- 22:33
- Unidentified speaker: But even in that case, spend on AI and compute is gone staying high.
- 22:38
- Unidentified speaker: That's conviction view, which of course could be wrong.
- 22:42
- Unidentified speaker: Which should you invest then?
- 22:43
- Unidentified speaker: Uh there's a few things we like.
- 22:44
- Unidentified speaker: I think like now's actually a time to look at all the stuff that got absolutely pollaxed during that AI burn.
- 22:50
- Unidentified speaker: So there's some compounders that the fintech ones that I don't think will come back.
- 22:54
- Unidentified speaker: But we bought intuitive surgical recently.
- 22:57
- Unidentified speaker: I think there's like some interesting med tech plays that have massively compressed in terms of multiples.
- 23:03
- Unidentified speaker: The biggest drawdown in quality that I can remember, we generally avoided those companies anyway, whether it's your Nikes or your Boston Scientific, like those kinds of things, like credit scoring companies, all those things that like a whole species of fund would specialize in, they're all down significantly, like 60-70%.
- 23:21
- Unidentified speaker: So I think some of those that are doing well and growing, you're not talking 100%, you're talking about 8 to 15% top line, and the good ones maybe 10 to 25% of the bottom line, they're looking really compelling, like a really rare entry opportunity into those high-quality businesses.
- 23:35
- Unidentified speaker: Usually they're just too expensive and they run really hard and it's very hard to buy.
- 23:39
- Unidentified speaker: The other thing is I actually think in the local market, it's possibly time.
- 23:42
- Unidentified speaker: There's been a big fun cycle here as well.
- 23:45
- Unidentified speaker: So basically every tech stock went down.
- 23:47
- Unidentified speaker: Tech index here dropped, must have been over 60%.
- 23:50
- Unidentified speaker: Whether it was REA, Wise Tech, Zero, Car Sales, all the local companies, Life E60, it's hard to think.
- 23:57
- Unidentified speaker: I can't really think of an example of a company that isn't significantly down.
- 24:00
- Unidentified speaker: There's obviously been a lot of fund closures here.
- 24:02
- Unidentified speaker: So when a fund closes, they have to sell.
- 24:04
- Unidentified speaker: And like at the moment in Australia, there's been no marginal buyer than that, really.
- 24:08
- Unidentified speaker: Everyone's kind of suffering.
- 24:10
- Unidentified speaker: Tax rates are up, real estate's down, interest rates are up.
- 24:14
- Unidentified speaker: Like it's a really toxic environment for Australians.
- 24:16
- Unidentified speaker: But that's also created an opportunity where the last 12 to 15 months, this stuff is basically down in a straight line and it's just been bouncing along the bottom for the last few months.
- 24:25
- Unidentified speaker: So I wonder if there is an opportunity there.
- 24:27
- Unidentified speaker: Particularly as we get towards the end of that, what I I'm guessing is the end of that redemption cycle where you have all these funds closing and presumably have finished transferring or closing their positions.
- 24:36
- Unidentified speaker: And that actually suits us pretty nicely.
- 24:38
- Unidentified speaker: If it's healthcare and ASX Tech, beaten down ASX Tech, that seems that's in our wheelhouse.
- 24:45
- Unidentified speaker: So we'll definitely look at those opportunities.
- 24:47
- Unidentified speaker: And there will be a monster semiconductor rally on the back end of this now that exposures have gone down.
- 24:52
- Unidentified speaker: Is this your chart that was like fund manager exposure to software and semiconductors?
- 24:57
- Unidentified speaker: And it kind of went like software went up to 15, 20%, then down to zero, and that was basically the low.
- 25:02
- Unidentified speaker: And semiconductors went from 2-3% up to 20% and has come smashing down now.
- 25:08
- Unidentified speaker: So we're kind of at that moment.
- 25:10
- Unidentified speaker: And you remember like March April this year.
- 25:12
- Unidentified speaker: Wow.
- 25:13
- Unidentified speaker: I don't know if that's fast or slow.
- 25:14
- Unidentified speaker: Feels like a long time ago.
- 25:15
- Unidentified speaker: But remember there was like SaaSpocalypse, and everybody was saying it's all over, and people were writing thought pieces on Twitter and about how everything's done, and then the good companies all went up 2-3x from the lows.
- 25:25
- Unidentified speaker: I feel like we're really there now, or will be there soon in semiconductors where everybody's painting scenarios about how there's gonna be a big collapse and blah blah blah.
- 25:34
- Unidentified speaker: And it was kind of the same way, and you know what's similar with software?
- 25:39
- Unidentified speaker: The software move was all narrative.
- 25:40
- Unidentified speaker: Like go look at the quarterly results of most of these companies, you just see this going up like this.
- 25:45
- Unidentified speaker: And actually, for the first time, they actually took profitability seriously.
- 25:47
- Unidentified speaker: And with the help of AR, we have to get efficiencies as well.
- 25:50
- Unidentified speaker: So you saw profits, saw gap profits increasing massively as well.
- 25:54
- Unidentified speaker: You can't see the sell-off.
- 25:55
- Unidentified speaker: It's not like earnings, you go to GFC and you'll see like earnings go up and then down and then up again.
- 26:00
- Unidentified speaker: It's not that it just went up every quarter, and in the meantime, the stocks massively up, then massively down, then massively up, then massively down in that last little tip.
- 26:08
- Unidentified speaker: But because they because the good companies continue to grow, the narrative didn't really hold.
- 26:12
- Unidentified speaker: And then there was a huge rally off those lows.
- 26:14
- Unidentified speaker: And I think that's what happens with semis here, to be honest.
- 26:16
- Unidentified speaker: That's my bet.
- 26:17
- Unidentified speaker: I feel like I know we're all gonna use computers more and these GPUs, and I can see like the returns, and I know that the hyperscalers are funded, and maybe the growth stops, that 50 bill a quarter from Google and similar amounts from others.
- 26:29
- Unidentified speaker: I don't think that's gonna change.
- 26:31
- Unidentified speaker: I don't know that semiconductors are gonna continue to grow and massively beat and do 10, 20% a quarter, obviously on the right thematics and the right companies.
- 26:40
- Unidentified speaker: But if they continue to do that, they're gonna come roaring back.
- 26:43
- Unidentified speaker: And there's like heavy short interest now.
- 26:45
- Unidentified speaker: The problem is that when funds kind of exit these stocks and short interest goes up, they get really choppy.
- 26:50
- Unidentified speaker: At the moment, it's just been a slide down, but it usually takes a bit of time for that to wash out, and then you get a nice clean rally afterwards.
- 26:57
- Unidentified speaker: I still think that's the base case.
- 27:02
- Unidentified speaker: Yeah.
- 27:02
- Unidentified speaker: So in the meantime, we're gonna focus on healthcare compounders that are smashed.
- 27:06
- Unidentified speaker: We'll try and find some interesting things in Australia on the ASX.
- 27:10
- Unidentified speaker: Good thing about the ASX companies, we know them pretty well.
- 27:13
- Unidentified speaker: And on the other side, we will be very focused on the right companies.
- 27:19
- Unidentified speaker: And so I don't think they all recover.
- 27:22
- Unidentified speaker: Like the software companies, some of them didn't recover.
- 27:24
- Unidentified speaker: The ones that actually were disrupted by AI, like outsourcing software or something like that, didn't recover.
- 27:32
- Unidentified speaker: The ones that continue to grow did.
- 27:34
- Unidentified speaker: And I think it's semiconductors, it's gonna be like the neo clouds, it'll be very hard for them.
- 27:38
- Unidentified speaker: No one's gonna believe their growth plans, they're gonna have to go from minus 24% margins to profitable margins if they're gonna turn things around, if they want credit markets to take them seriously, they're gonna have to show actual profits.
- 27:50
- Unidentified speaker: Yeah.
- 27:50
- Unidentified speaker: I guess the one risk, the biggest risk now is actually that there's just a liquidation issue.
- 27:55
- Unidentified speaker: Sorry, not liquidation.
- 27:58
- Unidentified speaker: The biggest risk now is actually a liquidity issue.
- 28:01
- Unidentified speaker: So you've got Google and Meta issuing stock.
- 28:06
- Unidentified speaker: Used to be massive buyers of stock doing buybacks.
- 28:10
- Unidentified speaker: Anthropic raising a hundred bill.
- 28:13
- Unidentified speaker: You've got every neo cloud doing it at the market.
- 28:15
- Unidentified speaker: Offerings are just selling stock into the market.
- 28:18
- Unidentified speaker: And so I actually think what's happened is all this is hitting the market at once.
- 28:21
- Unidentified speaker: And Anthropic has an IPO, but if you want to put 100 billion into that, you have to get it ready, you have to get the cash ready.
- 28:27
- Unidentified speaker: So across the market, participants have liquidated 100 billion there.
- 28:31
- Unidentified speaker: The offerings from Google and Meta are substantial.
- 28:34
- Unidentified speaker: I don't think they're quite 100 billion, but you're getting up there, especially when you add them all up.
- 28:37
- Unidentified speaker: So you probably have hundreds of billions of dollars of equity issuance in the pipeline.
- 28:41
- Unidentified speaker: And if semis are down as much as they are, I don't know who takes the other side of that.
- 28:46
- Unidentified speaker: We're basically mostly out now, having taken a pretty big drawdown to get there just because it happened so fast.
- 28:52
- Unidentified speaker: People like us are not really going to be certainly not piling into.
- 28:57
- Unidentified speaker: I mean, we never own neo clouds.
- 28:58
- Unidentified speaker: But I doubt there's a bunch of people willing to put their hand up to buy core weaver Nebius and things like that.
- 29:03
- Unidentified speaker: Yeah, there's a real liquidity crunch going on.
- 29:07
- Unidentified speaker: And that could continue for a while.
- 29:08
- Unidentified speaker: And so you could get into a situation where everything is under pressure.
- 29:11
- Unidentified speaker: But again, that's why it's still kind of confusing, because there used to be two parts of the AI trade.
- 29:17
- Unidentified speaker: It was the money spenders and the money receivers.
- 29:19
- Unidentified speaker: And we always basically positioned ourselves on the money receivers, but now both of them have gone down.
- 29:25
- Unidentified speaker: I guess now, more recently, the hyperscalers, the big spenders, they've kind of steady, probably because they're free cash for generative.
- 29:32
- Unidentified speaker: And if AI data centers are slowed down through real entry action, their existing data centers are far more valuable than if every Tom Dick and Harry was able to raise money and build a data center.
- 29:42
- Unidentified speaker: It makes the actual hyperscalers more valuable.
- 29:44
- Unidentified speaker: But yeah, that correlation has broken down.
- 29:46
- Unidentified speaker: Like more recently, the spenders and the receivers, particularly receivers, have been smashed.
- 29:51
- Unidentified speaker: So I wonder what happens.
- 29:55
- Unidentified speaker: Maybe there's one positive note to leave on, and that is it does look decisively that interest rates have peaked based on bond yields in the United States and more locally even in Australia.
- 30:06
- Unidentified speaker: So maybe there's just one more rate rise or two more rate rises here.
- 30:11
- Unidentified speaker: And it's all tied up together, right?
- 30:12
- Unidentified speaker: So that pressure from data centers and every company being funded to build data centers, and that was highly inflationary.
- 30:19
- Unidentified speaker: If that comes off, it's very financially restrictive at the moment.
- 30:22
- Unidentified speaker: Just financial conditions are highly restrictive based on high interest rates.
- 30:26
- Unidentified speaker: If that damage has been done and you can see it obviously the market sell off, then yeah, that's like that's like the next step towards recovery.
- 30:36
- Unidentified speaker: Why don't we wrap it up then?
- 30:38
- Unidentified speaker: Sounds good.
- 30:39
- Unidentified speaker: Thank you so much, Michael.
- 30:40
- Unidentified speaker: See you next time.
- 30:42
- Unidentified speaker: Thanks, Al.
- 30:42
- Unidentified speaker: See you next time.