GUY: Good morning... it is Tuesday, May 6th, 2026, and this is Morning Signal. I’m Guy, caffeinated, mildly offended by what the stock market thinks is a reasonable reaction to a full-blown oil shock...

AVA: And I’m Ava... also caffeinated, but more curious than offended, because honestly the weird part is that the market might have a logic here, even if it’s a dangerous one. We’ve got to unpack it.

GUY: Yeah, today is one of those mornings where every smart person seems to be describing the same movie from different camera angles. Stocks are acting like liquidity still rules. Oil is acting like geopolitics still matters. And the economy underneath feels way narrower than the index headline suggests.

AVA: Exactly. And the theme across the podcasts was basically this... the melt-up can continue, but it’s being powered by fewer engines, with more fragility, and a lot more inflation risk than the market wants to admit.

GUY: So let’s start there... markets and macro first, because that’s the center of gravity today.

GUY: So over on Macro Voices, Erik Townsend laid out the cleanest version of the current disconnect. During that recording, front-month WTI was above 110 dollars, spot Brent was above 120, and on the weekly scoreboard June WTI was around 106.88, up almost 15% week over week. At the same time, the S and P 500 was sitting around 7135 and making fresh highs.

AVA: Right... and over on Macro Voices, Townsend’s analogy was February 2020, which is a pretty loaded comparison. Not because he’s saying we’re getting the exact same crash, but because he’s saying you can sometimes watch the real-world shock first, and then watch equities spend a weird lag period pretending none of it counts.

GUY: Exactly. And I think that’s the key. Markets are not pricing the first-order shock anymore, they’re pricing the second-order assumption that central banks will look through it, that consumers can absorb it, and that AI capex will keep earnings momentum alive. That’s the bet.

AVA: OK but here’s what’s interesting... over on RenMac Off-Script, Neil Dutta kind of explained why that bet can work for a while and still be deeply misleading. He said if you strip out some of the federal-government noise, the last two quarters look more like 1.5% real GDP growth, not some booming broad-based economy. Real goods consumption is weak. Residential investment is weak. Structures are weak. Transportation equipment is weak. His phrase was the economy is “running on one engine.”

GUY: Which is AI.

AVA: Which is AI. Information-processing equipment and software capex. That’s the engine. Not the median household. Not housing. Not trucks. Not broad manufacturing strength.

GUY: And that matters, because the index can look healthy when the economy is basically one giant corporate capex program wearing a trench coat. If Nvidia-adjacent spending, hyperscaler buildout, networking gear, and software infrastructure are booming, the S and P can absolutely look fine while the real economy feels lumpy and narrow.

AVA: But hold on though... over on Forward Guidance, Danny Dayan would say that’s not just an earnings-composition story. He thinks policy itself is still adding fuel. His line was blunt... every single day the Fed is not hiking, it’s easing.

GUY: I loved how provocative that was, because it sounds wrong until you unpack what he means. He’s not saying the fed funds rate mechanically drops every day. He’s saying financial conditions are a flow variable. If credit growth is reaccelerating, if money growth is picking up, if the market keeps moving cuts forward in its imagination, if households and businesses adapt to higher rates better than expected... then a static policy rate can become less restrictive over time.

AVA: Over on Forward Guidance, Dayan’s data was pretty specific too. He pointed to core inflation running about 4.4% on a three-month annualized basis even before the war shock. He pointed to M2 growth accelerating. He said commercial loan growth was roughly 12% at regional banks and the fastest in 15 years at large banks. So his argument is... don’t look only at the policy rate. Look at the transmission.

GUY: Right. And his bigger point may be the most important macro argument of the week. Over on Forward Guidance, Dayan said the Fed’s estimate of neutral, around 3.0% to 3.1% in the long run, is just too low. He thinks the real neutral rate is more like 4.3% to 4.5%.

AVA: Which is a huge difference.

GUY: Massive. Because if he’s right, then the Fed isn’t mildly restrictive. It’s basically near neutral... maybe even a little easy. And that explains why housing won’t fully crack, why manufacturing hasn’t imploded, why durable goods are hanging around, why risk assets can levitate.

AVA: I buy part of that, but not all of it. I think he’s right that the economy tolerated higher rates better than the 2010s framework suggested. But over on Excess Returns, Rich Bernstein made a slightly different argument that I actually think is more politically realistic. He said the 2% inflation target is basically antiquated in practice. Not officially... but in practice. The real tolerance looks more like 3% to 3.5%, maybe even 3% to 4%.

GUY: And that’s a credibility-gap argument, not a model-gap argument.

AVA: Exactly. Bernstein’s point is not “the Fed’s equations are wrong,” it’s “the Fed cannot force 2% without unacceptable political and economic pain, so it behaves as if the target is looser than the slogan.”

GUY: Which I think is right. Central banks can say 2% all day long, but if you’ve got heavier defense spending, tax support, sticky services inflation, and now an oil shock... the actual policy reaction function tells you what the real target is. The market has figured that out before the Fed has admitted it.

AVA: Over on Excess Returns, Bernstein tied that to the late 1960s “guns and butter” analogy, which is pretty smart. He’s saying this isn’t exactly the 1970s yet, where oil instantly detonates demand. It’s more like a slower inflation seedbed... defense outlays, fiscal support, and a central bank that doesn’t really want to slam the brakes.

GUY: He cited that proposed 1.5 trillion dollar defense budget and a big tax package as part of that. So you’ve got fiscal policy adding nominal growth at the same time supply conditions are more constrained. That’s inflationary even if it doesn’t cause an instant recession.

AVA: And that’s where the oil debate gets really nuanced. Over on Forward Guidance, Dayan said oil in the 90s to low 100s is more inflationary than demand-destructive. He doesn’t think you get real macro breakage until something like 150 to 200 dollar oil. Over on Excess Returns, Bernstein said something similar in a different language... oil as a percent of wages is still much lower than in the 1970s, so gasoline can crush sentiment before it crushes spending.

GUY: Right, and that’s an important distinction. If gas jumps, consumers feel poorer immediately. They complain immediately. Politicians panic immediately. But actual spending destruction can lag because energy is a smaller share of the household budget than it used to be. So the first effect is inflation and a savings hit, not necessarily an instant recession.

AVA: But over on Macro Voices, Townsend’s pushback was basically... fine, maybe it doesn’t crush the consumer overnight, but markets are still not pricing the physical lag effects. Shipping delays. Inventory drawdowns. Jet fuel issues. Margin compression. Those show up later.

GUY: That’s the point people miss. Oil shocks are not just “how much does gasoline cost Friday?” They’re logistics shocks. They’re working-capital shocks. They’re delivery-time shocks. They hit airlines, chemicals, freight, importers, retailers, industrials. It’s like throwing sand in the gears, and the first market reaction often underestimates how much equipment is downstream of that gearbox.

AVA: Wait really... because I think this is where investors can talk themselves into complacency. If payrolls don’t immediately collapse and the consumer doesn’t instantly stop spending, they conclude the shock is manageable. But over on Macro Voices, Townsend is saying the lags are the story, not the rebuttal.

GUY: Exactly. It’s the difference between being not-dead-yet and being healthy. The market is interpreting “no immediate recession” as “no serious problem.” Those are wildly different things.

AVA: And then over on Macro Voices, Daniel Lacalle added the liquidity explanation for why that gap can persist. He said global money-supply growth is the fastest since 2021, led by China and supported by the U.S. and the U.K., but velocity is weak because war and uncertainty reduce productive turnover. So excess liquidity goes into financial assets before it shows up in stronger real activity.

GUY: That was a really useful frame. If velocity is soft, money doesn’t chase as many real transactions. It chases whatever can absorb it quickly... and financial assets are perfect for that. So you can get a bad macro backdrop and a good tape at the same time. In fact that combination almost defines late-cycle melt-up behavior.

AVA: OK but I don’t want people hearing that and thinking it’s bullish in a healthy way. Over on Macro Voices, Lacalle wasn’t saying the oil shock is harmless. He was saying liquidity can delay recognition. That’s different.

GUY: Completely. It’s like adding helium to a house with a cracked foundation. It can keep floating... that doesn’t mean the crack disappears.

AVA: Over on RenMac Off-Script, Jeff deGraaf brought in the technical angle, and honestly it lines up eerily well with the macro story. He said breadth is narrow. Utilities and energy look clearly strong in their sector work. Outside that, a lot of areas are just middle-of-the-road, and communication services and tech are being carried by a few giant names.

GUY: Which means the index is telling a prettier story than the average stock. That’s always dangerous because concentration looks safe right up until it doesn’t. If five or seven names are doing the lifting, every investor thinks they own the market, but really they own one narrative.

AVA: Over on RenMac Off-Script, deGraaf’s concrete warning was the SOX. His bubble heuristic is that when an index doubles in two years, it enters a danger zone where marginal buyers may be exhausted and the next three to six months get a lot more vulnerable. He said the semiconductor index had just tripped that threshold.

GUY: And semis are not some side quest here. Semis are the market’s spinal cord right now. If the SOX wobbles, it’s not just chip stocks. It’s capex confidence, AI confidence, cloud confidence, index confidence.

AVA: Over on Excess Returns, Chris Davis attacked the same issue from investor psychology. He said those catchy labels like MAG7, FANG, Nifty Fifty... they’re often signs that investors have stopped treating concentration as risk. They start treating it as destiny.

GUY: That’s one of the oldest tells in markets. When the acronym becomes the argument, you’re late. Nobody in history ever got safer because they found a better marketing name for crowding.

AVA: And over on Excess Returns, Davis had that great story about his mother’s portfolio. Roughly 3 million dollars, originally funded with 9% to 10% muni bonds, spending around 200,000 dollars a year, then slowly shifted into equities as the bonds matured. By basically not selling, that portfolio outperformed his own fund by about 500 basis points a year for nearly 20 years, because winners like Amazon and Google were allowed to compound into huge weights.

GUY: Right... and that’s such a useful distinction. Concentration created by compounding is not the same as concentration created by narrative chasing. One is earned over time by business performance. The other is rented by people who show up because the chart looks like a launch ramp.

AVA: Exactly. And Davis said his firm improved after raising the position cap from 5% to 10%. But he still trims when the forward expected return compresses. He mentioned Meta and Applied Materials as names they reduced when modeled IRRs fell toward 4% to 5% from around 14%.

GUY: Which is the discipline piece. Even a great company can be a mediocre stock if you’ve moved all the returns into the past tense.

AVA: Over on Excess Returns, Bernstein made that portfolio point sharper. In a more inflationary regime, he prefers front-end cash flows to distant-duration stories. He said people paying more for gasoline and groceries care about cash in 15 days, not promises in 15 years.

GUY: That’s why his dividend payer argument matters. In the post-GFC world, people got rewarded for owning long-duration growth basically forever. If the real regime is now more 3% to 4% inflation with occasional spikes, then near-term cash generation should re-rate up and pure fantasy-duration stuff should re-rate down. That’s a real style shift.

AVA: And that’s where I want to push you a little. Because if the market is still being fed by AI capex and passive easing, maybe duration can keep working longer than you think.

GUY: It can... and that’s the trap. Over on Forward Guidance, Dayan was explicit: buy every dip until one of three conditions breaks the regime. Oil around 150, the 10-year Treasury around 5.50%, or a genuinely hawkish Fed response. He even pairs that with hedges in oil upside, bond shorts, and VIX upside, with the VIX around 17.

AVA: That’s a pretty elegant framework, honestly. Participate in the melt-up, but name your exit ramps.

GUY: It is elegant. I just think the timing is brutal. Over on Macro Voices, Townsend said he doubled down on his S and P hedge at 7200 while staying long crude after buying the dip around 79 WTI earlier. That tells you how hard this tape is... you can be bullish oil and bearish index complacency at the same time.

AVA: And over on Macro Voices, Patrick Ceresna basically split the difference. He agreed the macro logic is deteriorating underneath, but said systematic flows, liquidity, and AI leadership can keep equities pinned high for weeks or longer. Which, honestly, feels like the most annoying answer and maybe the most realistic one.

GUY: The market’s favorite trick is staying irrational for precisely long enough to make disciplined people look stupid and undisciplined people feel brilliant.

AVA: That’s wild... but yes.

GUY: One last macro point before we move to tech. Over on Macro Voices, Townsend also said he sold more than half his gold position above 4730. That’s notable because geopolitics usually screams “buy gold,” but his argument is that oil-driven inflation reduces the odds of Fed easing, pushes real rates risk back into the picture, and could drag gold below the 200-day and maybe even below the prior 4100 low before the secular bull resumes.

AVA: I actually think that’s an important reminder. Geopolitical stress is not automatically gold-positive if the dominant transmission is “higher inflation plus less easing.” Sometimes the thing people call a safe haven is just another duration trade with a mythology budget.

GUY: Beautifully said. All right... let’s pivot to tech and AI, because this is the one engine everybody keeps circling back to.

AVA: So over on the Dwarkesh Podcast, Reiner Pope gave one of the clearest explanations I’ve heard for why AI economics increasingly favor scale. His central claim was that batch size is the dominant variable in inference economics. If you don’t have enough simultaneous demand, the serving costs can be, his phrase, a thousand times worse than at scale.

GUY: That “thousand times worse” line should make every investor sit up. Because a lot of people still talk about AI like software margins are automatic. Pope is saying no... the economics depend on traffic density, queueing, memory bandwidth, and how efficiently you amortize weight loading across users.

AVA: Exactly. Over on the Dwarkesh Podcast, Pope formalized runtime as basically the max of compute time and memory time. In low-batch settings, you keep paying to fetch the model weights over and over. In higher-batch settings, you amortize that, then eventually hit another floor tied to pure compute throughput and KV-cache traffic.

GUY: Which is why scale matters so much. The giant providers with huge traffic pools can batch requests, spread fixed costs, and squeeze the unit economics. A smaller provider with spiky or thin demand can be structurally uncompetitive even if the model quality looks similar.

AVA: Over on the Dwarkesh Podcast, Pope also explained why “fast mode” products are so expensive. If you pay a big premium for lower latency, what you’re really buying is a smaller batch. Less waiting, but worse hardware amortization. He gave that example where you might pay 6x more for only about 2.5x faster responses.

GUY: I loved that because it turns pricing tiers into architecture reveals. It’s not arbitrary. It’s physics and queueing theory wearing a product manager costume.

AVA: Exactly. And over on the Dwarkesh Podcast, Pope had another really actionable rule of thumb... optimal batch size needs to exceed roughly 300 times sparsity. For a DeepSeek-style mixture-of-experts model with 32 active experts out of 256 total, he got to something like 2,000 to 3,000 concurrent sequences or tokens in flight per decode step, which could translate to roughly 128,000 tokens per second per rack if batches leave every 15 to 20 milliseconds.

GUY: Which means efficient inference needs serious traffic, but not infinite traffic. It supports the idea that the economics favor large platforms, but maybe not only three firms on earth. There’s room for scale-up domains that are big enough to be efficient without total monopoly.

AVA: Right. But there is still a real centralization force. Over on the Dwarkesh Podcast, Pope said public API pricing increasingly leaks where the internal cost curve bends. He used Gemini’s roughly 50% pricing jump above 200,000 context as evidence that Google’s architecture likely enters a more KV-cache-dominated region there. He even backed into around 1.7 kilobytes of KV storage per token.

GUY: Which is catnip for developers and investors. If pricing exposes hidden constraints, then pricing pages become industrial maps. You can infer where memory bites, where batching breaks, where context length gets expensive. That’s strategic information.

AVA: And it links directly to the capex boom. Over on the Big Technology Podcast, Bill Gates said AI is not tulips... it’s more like the internet bubble. Meaning the technology is real, the long-term value is real, “cheap intelligence” is a genuine platform shift, but a lot of the first-wave infrastructure spending will still be wasted.

GUY: That’s the right analogy. The internet bubble produced enormous value and enormous overbuilding. Both were true. Fiber was overbuilt. Plenty of companies died. The internet still won. Same thing here... bad data center locations, wrong chip generations, projects that never earn their cost of capital... all very possible. AI can still be transformational.

AVA: Over on the Big Technology Podcast, Gates was blunt that big tech doesn’t get to sit this one out. His line was basically... if you want to be a tech company, you don’t get to say no. The strategic penalty for underinvesting may be bigger than the financial penalty for some overbuilding.

GUY: And that is exactly why the capex keeps going even if return-on-investment math looks fuzzy in the near term. Nobody wants to be the executive team that saved shareholders money in year one and accidentally forfeited the next decade.

AVA: But over on Diary of a CEO, Scott Galloway gave the cleanest valuation pushback. He said the current AI buildout implies one of two things over the next three to five years... either roughly 1 trillion dollars of incremental revenue gets created from new AI-enabled products, or labor displacement delivers huge margin expansion. Otherwise, current valuations may need to compress by 50% to 70%.

GUY: That’s the math test. And I think he’s directionally right, even if the exact percentages are debatable. Market cap expansion has outrun realized monetization. At some point, investors stop rewarding “strategic necessity” and ask “show me the cash.”

AVA: I slightly disagree on the timing. Over on Diary of a CEO, Galloway also pointed out that labor-market apocalypse hasn’t shown up in the aggregates the way the most dramatic AI rhetoric would imply. U.S. unemployment around 4.5%. Youth unemployment around 8.8%. Radiologist job postings rising. Coding job listings up about 11% year over year. New business permits per capita doubling over 10 years.

GUY: Right, and that’s important. AI disruption can be real without producing instant macro unemployment carnage. It may hit distribution first... customer service, junior legal work, trucking, entry-level office work... before it shows up as “national joblessness.”

AVA: Exactly. Which means the market may be overpricing near-term monetization and some commentators may be overpricing near-term labor collapse. Both can be true.

GUY: Over on The Compound and Friends, Chris Ankrom added a contrarian angle I thought was refreshing. He thinks the market is too quick to assume AI wipes out every incumbent software vendor. A lot of system-of-record businesses in high-consequence workflows are more defensible than the market is giving them credit for.

AVA: That’s a big point. AI can disrupt interfaces and workflows without instantly dislodging the database of truth. If you run payroll, ERP, compliance, claims processing, medical records, logistics, industrial planning... the incumbent system still owns the permission set, the historical data, the audit trail, the embedded process.

GUY: Yeah. The sexy demo is not the same as a replacement cycle. Investors keep conflating “better assistant” with “core software vendor dies.” Those are different claims.

AVA: Over on The Compound and Friends, Ankrom’s broader investing idea was the “coffee can” mindset... buy durable compounders the market temporarily misjudges, then tolerate the ugly drawdowns required to harvest long-run compounding. That’s especially relevant in this AI panic because anything not labeled “winner” gets treated as roadkill.

GUY: And sometimes the best trade is buying the business everyone assumes will be disrupted immediately... when in reality the displacement takes eight years, not two. Markets discount on headlines, enterprises migrate on compliance calendars.

AVA: Exactly. Over on Business Breakdowns, Cloudflare was a good case study in a less-hyped but still essential layer. Management framed 2020 as proof it could “make sure the internet worked” under extreme stress, and the real signal wasn’t just branding. Customers with more than 100,000 dollars in ARR grew 57%, and they added more than 10,000 new paying customers in the quarter discussed.

GUY: That matters because enterprise security and zero-trust products get sticky fast. Once a large customer standardizes on that layer, switching is painful. It’s not as flashy as frontier-model demos, but it’s exactly the kind of infrastructure spend that survives narrative rotations.

AVA: And it’s also why I’m not in the camp that says “avoid tech if inflation stays sticky.” The better frame is... own the tech that remains necessary when money gets more expensive.

GUY: Right. Necessary beats aspirational in late-cycle regimes.

AVA: Over on Invest Like the Best, Brian Chesky gave a very different but related AI insight. He said AI is forcing companies to redesign management itself. His argument was that the age of AI eliminates “pure people managers” and rewards leaders who stay hands-on and have direct access to ground truth.

GUY: I buy that. In a faster product cycle, every extra layer is a latency tax. If models improve monthly and consumer expectations shift in real time, you cannot run a five-meeting approval chain like it’s 2016.

AVA: Over on Capital Allocators, WCM made the same point from a culture angle. They talked about trust, generosity, behavior-based talent management, and just building internal AI tools quickly and killing them if they don’t work. Less bureaucracy, more experimentation.

GUY: That’s actually a bigger deal than it sounds. AI doesn’t just reward technical capability. It rewards organizations that can shorten the distance between signal and action. The companies that still need three committees to approve a pilot are going to lose to companies that just ship, test, and iterate.

AVA: But hold on though... that organizational flattening has a cost too. Fewer layers can mean more speed, but also fewer shock absorbers. More founder mode, less institutional memory.

GUY: True. Faster systems are usually less forgiving. Same in markets, same in companies. We’ll come back to that when we talk prediction markets.

AVA: Before we leave AI, I want to hit one more subtle thing from the Dwarkesh Podcast. Over on the Dwarkesh Podcast, Pope said some frontier models may now be optimized less for pretraining purity and more for lifetime deployment economics. He floated the idea that some systems could be around 100x overtrained relative to Chinchilla-optimal pretraining once you account for reinforcement learning and downstream inference economics.

GUY: That’s huge. It means the optimization target is shifting from “win the benchmark” to “earn the lifetime margin.” Smaller active models, more efficient serving, architecture shaped by deployment economics... that has investment implications across chips, memory, power, and software stacks.

AVA: Exactly. The AI race is not just about bigger models. It’s about cheaper useful intelligence at scale.

GUY: All right... let’s go to geopolitics, because this oil move is not some abstract macro factor. It has a live geographic center, and that matters.

AVA: So over on Macro Voices, the live center of gravity was obviously Iran and Hormuz. The immediate market fact pattern was clear... front-month WTI above 110 during the discussion, spot Brent above 120, gasoline products ripping higher too. Townsend’s whole point was that markets are still treating this as a headline event rather than a sustained physical disruption.

GUY: And that distinction matters because headlines are discountable, physical bottlenecks are not. You can ignore a cable-news scare. You cannot magically create shipping capacity, refinery flexibility, or extra inventories once the chain gets tight.

AVA: Over on Macro Voices, Lacalle gave the best geography map. He said the U.S. is comparatively resilient because it’s the world’s largest oil-and-gas producer and a net exporter of petroleum products. China has large stockpiles, a ban on refined-product exports, and continuing access to Russian supply. In his words, the big problem is Europe.

GUY: Europe again. And not just because Europe pays more for energy, but because Europe never truly rebuilt its energy shock absorbers after 2022. So this isn’t simply “higher oil.” It’s higher working capital, higher input costs, weaker industrial margins, more pressure on consumer sentiment, and potentially worsening credit quality.

AVA: Right. Over on Macro Voices, Lacalle walked through that chain pretty clearly. If energy costs rise sharply, firms need to finance more expensive inventories and fuel bills before they can pass those costs on. That ties up capital, compresses margins, and can eventually show up as credit deterioration.

GUY: That’s the part equity markets almost always underprice in the first phase. They price earnings revisions eventually, but they don’t initially price how much balance-sheet stress comes from higher working-capital needs. Especially in Europe, where margins were already less cushioned.

AVA: And this matters for politics too. Over on Macro Voices, Lacalle noted that roughly 25% of Iran’s GDP and 60% of government revenues move through Hormuz-related trade flows, which would normally make you think Iran can’t sustain prolonged disruption. But he also said regime survival thresholds are different from market-economy pain thresholds.

GUY: In plain English... “it hurts them too” is not a timing model. Governments under sanctions or authoritarian pressure can absorb economic pain longer than democratic consumer societies can absorb fuel inflation. That’s why betting on quick rational de-escalation is dangerous.

AVA: Exactly. Over on Macro Voices, Townsend framed it as a contest of staying power. The U.S. thinks Iran eventually has to shut in production as storage fills. Iran thinks it can create enough pain through Hormuz disruption to outlast its adversaries politically.

GUY: And when you layer in China’s stockpiles and Russian supply access, plus Europe’s vulnerability, you get a lopsided endurance map. The same oil price does not mean the same thing to every region.

AVA: Over on Macro Voices, Lacalle also highlighted emerging-market dispersion. India, Colombia, and Mexico as more vulnerable in different ways, while Brazil and Argentina could partially benefit from stronger commodity prices.

GUY: That wartime dispersion trade is important. Commodity exporters can get a terms-of-trade boost while importers suffer inflation and external-balance stress. So “EM” as a bucket becomes less useful. You need to know who sells molecules and who buys them.

AVA: And this also loops back to the AI story in a way I think investors are underestimating. Over on the Big Technology Podcast, Gates warned that some AI infrastructure bets will fail because electricity is too expensive. Over on Macro Voices, Lacalle is basically saying Europe is exposed to exactly that problem again. So energy security is becoming an AI competitiveness issue, not just an inflation issue.

GUY: Completely. If your region has unstable or expensive power, you lose twice. Households suffer and data-center economics deteriorate. The next industrial map may be drawn as much by power cost and grid reliability as by talent or tax policy.

AVA: That’s wild, but yes. We spent years talking about data as the new oil. Now it turns out actual oil, natural gas, and electricity are still very much the old oil.

GUY: And still undefeated.

AVA: So let’s bring the threads together, because today’s podcasts really converged in the cross-currents.

GUY: Over on RenMac Off-Script, Dutta said the economy is running on one engine... AI capex. Over on Forward Guidance, Dayan said passive easing is still feeding nominal demand. Over on the Big Technology Podcast, Gates said major tech platforms basically have no choice but to keep spending. Put that together and you get a pretty clear mechanism for why markets can levitate. AI capex props up semis, networking, data centers, software, cloud, and index earnings... even while the household economy looks softer.

AVA: Right, but over on RenMac Off-Script and Excess Returns, deGraaf and Davis both warned that concentration is exactly what makes this feel safer than it is. Narrow breadth plus acronym worship is not resilience. It’s dependency.

GUY: And over on Diary of a CEO, Galloway gives you the monetization test sitting underneath that dependency. If we don’t see something like 1 trillion dollars of new AI revenue or giant labor-cost savings over three to five years, then a lot of this current valuation structure is air. Not fraud... just overcapitalized hope.

AVA: I think the really important synthesis is this... the AI complex is both the macro stabilizer and the market’s single biggest point of fragility. It’s the thing keeping nominal growth up, and it’s the thing where expectations are most stretched.

GUY: Exactly. That’s the paradox. The hero and the vulnerability are the same basket.

AVA: Over on Forward Guidance, Excess Returns, and Macro Voices, there was also an interesting sequencing agreement on oil even though the tone differed. Dayan says oil in the 90s to low 100s is inflationary, not recessionary. Bernstein says gasoline pain is more about sentiment than immediate spending collapse because oil as a share of wages is still low by historical standards. Townsend says fine, but the lagged supply-chain and margin effects still matter. Those aren’t actually contradictory.

GUY: No, they fit together almost perfectly. Phase one... inflation without destruction. Phase two... earnings and credit pain with a lag. The market is correctly saying “not 1974 tomorrow morning.” It may be incorrectly saying “therefore not a problem.”

AVA: Over on Excess Returns and Forward Guidance, Bernstein and Dayan also agree, from different angles, that the Fed is not meaningfully targeting 2% in the world as it exists. Bernstein says the practical tolerance is 3% to 4%. Dayan says the neutral rate is higher and policy is already too easy. Either way, the implication is the same... inflation volatility is the real threat.

GUY: And inflation volatility is deadly for long-duration assets because it destabilizes the discount rate. Markets can live with high inflation if they understand the regime. They struggle when the inflation path is jagged and policy credibility is fuzzy.

AVA: Over on Odd Lots, Thomas Peterffy brought in a market-structure angle that actually fits this moment better than I expected. He argued prediction markets are moving from novelty toward financial infrastructure. The example they discussed was a Kalshi contract asking whether there will be a recession before 2027, trading around 33%.

GUY: I love that conceptually. Instead of inferencing recession odds through equities, rates, credit, and ten economists yelling on television, you just look at a contract pricing the actual question.

AVA: Over on Odd Lots, Peterffy’s point was that these markets can reduce messy debates to one plain number backed by real capital, kind of like fed funds futures became the shorthand for policy expectations. He also thinks Interactive Brokers has an edge because it already has serious users, whereas a platform like Kalshi may need sports or entertainment flow to bootstrap liquidity.

GUY: The distribution point is smart. Infrastructure often accrues to the place that already has the account relationship, not necessarily the most elegant standalone product.

AVA: But over on Odd Lots, Peterffy also flagged the plumbing risks... standardization and leverage. Best execution across venues only really works if the contracts are fungible. And as soon as firms start offering leverage on event contracts, you know somebody somewhere is going to blow themselves up.

GUY: That’s the eternal story of market innovation. Cleaner price discovery arrives hand-in-hand with a new way for people to do dumb things with leverage.

AVA: And his most controversial view was on insider trading. Over on Odd Lots, Peterffy argued that because information leaks anyway, insider-trading bans can actually preserve the advantage window for sharks rather than eliminate it. He thinks faster dissemination would compress the edge to a second or two.

GUY: I don’t buy that as policy... but I do think it reveals something bigger. We are moving toward systems that want less interpretation and more direct pricing. Prediction markets do that. API pricing does that in AI. Founder mode does that inside companies. Everyone is trying to remove the middle layer.

AVA: Exactly. Over on Invest Like the Best, Chesky wants fewer pure managers and more operator-managers close to the facts. Over on Capital Allocators, WCM wants fewer permission layers. Over on Odd Lots, Peterffy wants experts forced to back views with money instead of just commentary. Different domains, same instinct... flatten the stack.

GUY: Which improves speed and accountability... but also reduces damping. Less bureaucracy, less insulation, fewer buffers. Faster systems can be smarter and more brittle at the same time.

AVA: And that’s probably the best way to describe this whole market. Faster, smarter, more liquid, more concentrated... and more brittle.

GUY: So let’s finish with what we’re watching, because this is one of those tapes where the catalysts actually matter. The regime has specific tripwires now.

AVA: Absolutely. So first... over on Forward Guidance, RenMac Off-Script, and Excess Returns, the immediate macro catalyst is the mid-May 2026 CPI and PPI data. Dayan’s whole case depends on inflation already having reaccelerated before the war shock. If CPI comes in hot, that strengthens the argument that the Fed’s de facto tolerance is being tested and rate-cut fantasies get pushed further out.

GUY: I’d add gasoline pass-through there. Watch not just headline CPI, but how fast energy filters into household expectations. If inflation expectations rise while wage growth stays around that roughly 4% nominal pace Dutta talked about on RenMac Off-Script, the squeeze gets more visible.

AVA: Second... over on Macro Voices, the weekly EIA inventory data and actual Hormuz shipping conditions matter a lot over the next two to six weeks. Townsend’s point is that the market still hasn’t seen the full physical shortage because transport times and inventory effects lag. So watch inventories, refined-product tightness, freight stress, and whether disruptions persist even on quieter headlines.

GUY: Exactly. If the headlines cool off but inventories keep drawing and product markets stay stressed, that’s how you know this is moving from geopolitical narrative into physical economic reality.

AVA: Third... over on Forward Guidance, the next FOMC in mid-June 2026 is huge. The question is whether the Fed stays passively easy, as Dayan argues, or starts laying the groundwork to get more serious. Watch the statement language, press conference tone, and especially whether any dissents emerge.

GUY: And don’t just watch for hikes. Watch for any admission that the labor market doesn’t require as much support as they previously assumed. If the Fed even subtly acknowledges that, the market will read it as a warning shot.

AVA: Fourth... over on Forward Guidance, Dayan named the market thresholds. WTI near 150 and the U.S. 10-year near 5.50%. If neither gets hit, he thinks the melt-up can continue. If one does, the regime gets much harder to maintain.

GUY: Those are the fault lines. Oil at 150 means consumers and margins probably can’t shrug it off. Ten-year at 5.50 means the bond market is no longer politely tolerating the inflation regime. At that point, equity multiples have a real problem.

AVA: Fifth... over on RenMac Off-Script, watch semiconductor breadth and relative performance over the next three to six months. If deGraaf is right that the SOX doubling within two years puts it in a danger zone, then weakness there is not just sector noise. It’s a signal that the market’s one obvious growth engine may be tiring.

GUY: Especially watch whether the leadership broadens or narrows further. If semis stall but the rest of tech doesn’t pick up the baton, that’s when narrow breadth turns from trivia into threat.

AVA: Sixth... over on the Big Technology Podcast and Diary of a CEO, watch hyperscaler capex guidance and enterprise AI monetization over the next one to two quarters. Gates says the spend race is mandatory. Galloway says the revenue math still has to show up. If spend stays huge but monetization remains fuzzy, the “internet bubble, not tulips” debate gets much more intense.

GUY: That’s the whole game. The market can forgive overbuilding if there’s evidence of a giant future profit pool. It gets less forgiving if capex becomes permanent and monetization stays PowerPoint-shaped.

AVA: Seventh... over on Business Breakdowns, Cloudflare’s next earnings matter more than people think. Watch whether customers above 100,000 dollars in ARR keep growing near that prior 57% pace and whether security and zero-trust demand stays resilient. In a market obsessed with AI glamour, durable infrastructure growth stories are going to be tested on actual enterprise traction.

GUY: Yeah, those are the names that tell you whether necessary tech can still compound in a tougher macro regime. I’d watch not just the growth rate, but net retention, enterprise deal sizes, and any commentary on AI-driven network traffic or security complexity.

AVA: Eighth... over on Odd Lots, prediction-market infrastructure rollout is worth monitoring over the coming quarters. If Interactive Brokers actually builds meaningful prediction-market integration with consolidated feeds and standardized contracts, that could change how macro and geopolitical probabilities get priced across asset classes.

GUY: And it could also become a really useful stress gauge. If recession odds, election odds, policy odds, and conflict odds all get cleaner market prices, traditional commentary gets disintermediated. But again... leverage and fragmentation are the risks.

AVA: Ninth... I’d add a softer but important metric from RenMac Off-Script and Excess Returns. Watch the ratio of gasoline prices to wage growth and watch consumer sentiment versus actual spending. Bernstein’s point was that sentiment can break before spending does. If that gap closes the wrong way, the market’s “no recession yet” story gets shakier.

GUY: Good one. The consumer doesn’t need to collapse to matter. They just need to stop offsetting everything else.

AVA: And tenth... over on Macro Voices, watch the dollar. Townsend flagged an unfilled DXY gap near 99.38. If that fills quickly on renewed escalation, it’s a sign global stress is tightening financial conditions faster than equity investors want to admit.

GUY: Stronger dollar, higher oil, sticky inflation... that combo is not a charity event for global risk assets.

AVA: Not at all.

GUY: So where do we land this morning? My version is... the melt-up is real, the logic behind it is real, and the danger underneath it is also very real. Liquidity, passive easing, and AI capex can keep this market up longer than the bears want. But the higher it floats on one engine, the more violent the repricing can be if that engine sputters.

AVA: My version is similar but maybe a little less dramatic. I think the market is not crazy to hold up here, because the spending impulse from AI is genuinely enormous and the oil shock is not instantly demand-destructive. But I do think investors are underestimating how narrow the support structure is... and how much energy, inflation, and power costs now matter to the AI story itself.

GUY: So... participate, but don’t hallucinate.

AVA: That might be the bumper sticker.

GUY: It’s going on a hat.

AVA: Please don’t make the hat.

GUY: No promises. That’s Morning Signal for Tuesday, May 6th, 2026.

AVA: Thanks for starting your morning with us. We’ll be back tomorrow to connect whatever the market decides to ignore next.

GUY: Stay sharp, watch the levels, and don’t confuse a crowded trade with a safe one.

AVA: See you tomorrow.