The market got two pieces of good-looking news this week, and they pull in opposite directions. The first is simple and genuinely positive: inflation is cooling, and not just in energy. The second is more unsettling. We got a much clearer picture of how much money the AI buildout now needs, and where that money is coming from. Cheaper borrowing costs are arriving just as the AI industry has begun to fund itself. The first of those is a straightforward tailwind for stocks. The second is the circularity risk we have warned about all year, and this week it grew rather than shrank. Our position is unchanged: encouraged by the economic data, increasingly cautious about how the AI boom is being paid for.

Start with the data, because it is the more important of the two. July CPI came in cool, with headline inflation at 3.4% year over year and core easing to 2.5%, its lowest reading since 2021. Both monthly prints landed in line. Then the producer side confirmed it: July PPI was flat at 0.0% for the month, below the 0.2% the street expected. That is the disinflation reaching the stickier, domestic side of the basket, not just cheaper energy doing the work. Prediction markets noticed: Polymarket’s odds of a September Fed hike have dropped to 24%. The Fed voted 9-3 to hold two weeks ago and now waits on the July PCE report, where core ran 3.3% through June, before its mid-September meeting. For a market braced all year for Kevin Warsh’s quantitative tightening, cooling inflation is doing more to keep this rally breathing than any single earnings beat.

Now the more unsettling news. This was the week Nvidia stopped being just a chip supplier and became a lender. Wall Street groups are assembling a $500 billion funding package, with Nvidia’s involvement, to bankroll the next wave of AI data centers. Financing at that scale means the chipmaker is helping fund the very customers buying its chips, which flatters growth on the way up and concentrates the risk in one place on the way down. It did not stop there.

On August 11, CME Group said it will launch Compute futures on October 5, contracts that let companies trade and hedge the hourly rental price of Nvidia’s H100 and B200 chips much the way they already trade oil or electricity. Turning raw computing power into a tradable financial product is genuinely clever, and it is also the kind of thing that makes the hair on our necks stand up. When a boom starts spinning off its own futures markets and financing vehicles this quickly, we start counting the ways it rhymes with 1999. The bank of AI is now open for business!

Alongside all of it, Anthropic’s investors are floating a $2 trillion valuation for a potential October IPO, double the original $1 trillion target. Price it near that and Anthropic would eclipse SpaceX’s record $1.77 trillion June debut, the largest listing in history. We hold none of these names and are not calling the top. But we said on May 29, with our SpaceX prediction, that when the market’s most exciting story is priced on narrative rather than earnings, the cycle is probably late. A two-trillion-dollar private company, a chipmaker turning into a lender, and an exchange selling futures on computing power itself all belong in that same ledger.

That is why the risk we flagged on August 7 still sits at the top of our watch list. A meaningful slice of the cloud acceleration we cheered from Amazon, Alphabet, and Microsoft is compute bought by a small handful of AI labs, some of it financed by the very companies reporting the growth. This week’s headlines did not soften that concern. They sharpened it.

There is also a reassuring side to this. A Bloomberg report this week put OpenAI’s annualized revenue run rate above $40 billion, roughly double its late-2025 pace. The revenue underneath these valuations is real and growing fast, not purely a story. But it cuts two ways: that $40 billion still leaves OpenAI behind Anthropic, which reported a $47 billion run rate back in May. The money is materializing, and the real revenue these companies are generating is the bull case.

In a counterweight to the entire AI story. Berkshire Hathaway posted a Q2 beat, with book value per share growing 12.5% year over year and operating revenue up 10%, helped by solid insurance underwriting and the OxyChem close. No flashy catalyst, no capex slide, just compounding. In a market obsessed with growth rates, Berkshire’s edge remains patience and a fortress balance sheet. We hold it as exactly that, a hedge for an environment where the exciting money is getting expensive.

Portfolio Update

We made no trades this week. That is now the fourth held-and-watched stretch of the year, and the 13% Treasury buffer we trimmed to on July 17 stayed exactly where it was. The best trade is sometimes no trade, and into a live inflation print with the AI-financing story running this hot, patience was the position.

Radical transparency on the number: from last Friday’s close to today’s close, we finished the week up +0.4%. We measure ourselves against the S&P 500, which rose about +0.36% over the same window, so we finished a hair ahead of the benchmark. We started the week drifting down about 0.6% into Tuesday as the tape turned cautious ahead of the inflation data, clawed back to flat on Wednesday’s cool CPI print, then pushed higher on Thursday and into Friday as the flat PPI sealed the disinflation read. The gains were modest, but they came, and the back half is where we edged ahead.

Our position in Blackstone climbed back toward breakeven at about -2% as the private-credit panic keeps unwinding. Netflix, our July entry, added roughly 5%. Accenture nudged all the way back to flat, a full validation of the June 18 double-down we made on the day of its guidance-cut drop, an add that is up about 36% since. And the Thursday-and-Friday rally, led by two names, is what pushed us from flat to a small win.

Reddit is joining the S&P 500 on August 18, and the shares jumped roughly 13% on the news. Nubank, a high-conviction hold all year, topped $1 billion in quarterly net profit for the first time, up 49% year over year on a record 33% return on equity, and rose about 10%.

One more note: we ran fresh valuation work across the entire book this week, name by name. The idea is simple: we only want to own good, growing companies when we are not overpaying for that growth. We will spare you the spreadsheet. What matters is the conclusion: the exercise left us more confident than ever in three holdings in particular, Reddit, Novo Nordisk, and Blackstone. Each came through looking like a company we are glad to own at today’s price rather than one we are merely tolerating, and all three are names the market has doubted at some point this year.

Looking Ahead

The next real test is July PCE, the Fed’s preferred inflation gauge, where core ran 3.3% through June. CPI and PPI both blinked cooler this week, but PCE is the print Warsh actually acts on before the mid-September meeting. If it confirms, the melt-up case gets a good deal cleaner. If it runs hot, the buffer we kept earns its keep again. Underneath the macro, we are watching one thing above all: whether the AI-financing story, Nvidia as lender and Anthropic at two trillion, keeps compounding smoothly or starts to show its seams. The whole cloud trade we are long now rests partly on that financing holding together. So here is the question we are sitting with. If the money powering the boom increasingly comes from inside the boom, is that durable scale, or is it the loudest late-cycle tell we have seen yet?

Have a good weekend. Stay hydrated. 💧

Follow our portfolio in real-time on eToro: here

*The views, thoughts, and opinions expressed in this newsletter are solely my own and do not reflect the official policy or position of VTEX or its affiliates. All analysis provided here is based on public information and is for educational and entertainment purposes only.

We are back. It has been three weeks since the last letter, our longest silence of the year, and we will be honest about the reason: we spent two of those weeks on a beach, doing a passable impression of someone who does not check their portfolio on holiday. We failed at that impression, obviously. But here is the humbling part. The portfolio climbed perfectly well without us hovering over it. If there is a lesson in a good sun lounger, it is that the best trade is sometimes no trade at all.

While we were away, the market did the opposite of what a nervous investor fears. It went up, and fairly broadly. Since our last letter on July 17, the S&P 500 rose 3.8% and the Nasdaq 4.2%, both pushing back toward record ground. Apple even passed Nvidia as the world’s most valuable company in mid-July, a small tell that the market was quietly rotating away from the pure AI-buildout names and toward proof that the spending pays. More on the proof shortly, because it turned out to be the whole story of these three weeks.

The macro backdrop cooperated, mostly. The Fed held rates at its July 29 to 30 meeting even with a live hike priced going in, which is exactly the outcome we kept a buffer for. Then this Friday handed us a twist. The July jobs report showed the US economy lost 23,000 jobs, the first decline since February, with prior months revised lower, and with this the September rate-hike odds promptly fell to 34% from 55%. The threat of a Fed hike is receding, which is friendly to our portfolio, but it eased because the labor market may be cooling, not because inflation was cleanly beaten. We are not declaring the inflation fight won. During these weeks, Eurozone inflation actually rose to 2.9% in July, a sitting Fed voter publicly backed a hike, and US borrowing costs hit a nineteen-year high.

Geopolitics stayed noisy and, as it has all year, resolved quietly. Oil round-tripped through the window, spiking toward $90 and $100 on renewed US-Iran strikes and tanker attacks in the Strait of Hormuz, then easing back as the physical barrels kept moving and Tehran and Oman opened talks in early August to reopen the Strait. Crude never sustained triple digits. In sum, the key number to watch now will be the July CPI numbers that will be released on Wednesday. They will give us a glimpse of how this seesaw in the price of oil is impacting inflation numbers.

Now the main event. The reason this letter runs long is partly the beach and partly the calendar: nearly every company we own reported earnings in these three weeks. So we are flipping the usual proportions. The portfolio section below is closer to an earnings teardown than a weekly update, because the single biggest thing that happened is that our holdings had to show their numbers. The short version is that the quarter we waited on all year delivered its proof, cloud growth accelerated across the board, and this time the market largely paid up for it. The condition it attached is the interesting part.

Portfolio Update

First the scoreline. Since July 17, the portfolio is up 3.9%, on an unchanged cost basis, so no deposit is flattering the number. That edged the S&P’s 3.8% and sat just behind the Nasdaq’s 4.2%. Year to date, the picture stays humbler, roughly +10% against the S&P at +13% and the Nasdaq at +14%. We will say it out loud: the real-economy basket of stocks and the Treasury buffer cost us ground earlier in the year, and this three-week climb is starting to earn some of it back.

The one move worth claiming outright predates the window. On July 17, we cut the IB01.L Treasury buffer from about 20% of the portfolio to about 13%, redeploying roughly 7% into the core index funds and mega-cap tech. We were being conservative going in, but the data pointed to the market being likely to go up, so we acted on it, even if perhaps not as much as we should have. We are still holding that 13% back specifically to defend against the kind of volatility we have seen this year. But the reality is that this trim is what let us slightly beat the S&P over this three-week window instead of slightly lagging it. We will need to get more aggressive if we want to genuinely claw back the nearly three points we are giving up to the S&P year to date. Still, this was proof of a good decision. Beyond that trim, these three weeks were held-and-watched, with no new trades. Which brings us to the earnings.

Start with the hyperscalers, because they are our largest theses and they set the tone.

Alphabet ( ): Beat on every headline number and got sold anyway. Revenue was $120B, up 24%, Google Cloud grew 82% to $25B, accelerating from 63% the quarter before, with backlog now at $514B, and the Gemini app reached 950 million monthly users. Then, two lines down: capex of $44.9B for the quarter, a full-year 2026 guide raised to $195 to $205B, and free cash flow that turned negative at -$5.9B. That last figure is why the stock fell despite a flawless top line. Over the following days, as the other hyperscalers announced results, the stock clawed back all its losses and is now sitting basically where it was before earnings.

Microsoft ( ) : Showed the flip side. Revenue was $90B, up 18%, Azure grew 43% and crossed $100B in annual revenue for the first time; adjusted EPS was $4.74 against $4.24 expected, and the shares jumped about 8%. The difference from Alphabet came down to one word: discipline. Microsoft’s CFO trimmed calendar-2026 capex to roughly $175B even while guiding Azure to another 45% next quarter.

Amazon ( ) : Closed the case. Net sales of $201B, up 20%, operating income up 43% to a 13.7% margin that landed right inside the 13 to 14% range we called back in January, and AWS growth of 37%, its fastest pace in eighteen quarters, now a $169B annualized run rate. The lazy old bear case, that Amazon spends too much to ever be a high-margin business, is dead on the page.

Put the three together, and the read is simple. Cloud is accelerating, not fading. AWS 37%, Google Cloud 82%, Azure 43%, all up sequentially. And the market is no longer punishing the growth; it is rewarding it, but only when the spending comes with a visible hand on the brake.

We do want to flag one caution here, because it complicates the tidiness of that read. A meaningful slice of this acceleration is compute being purchased by a small handful of AI labs, OpenAI and Anthropic chief among them, rather than broad-based enterprise demand. Some of that is close to circular: Alphabet is itself financing a large piece of Anthropic’s build-out, meaning part of the cloud growth it reports is, in effect, self-funded. That does not make the growth fake, but it does mean the durability of these numbers rests on a handful of anchor tenants staying flush and staying committed, rather than on a broad base of ordinary enterprise customers. If that concentration ever slows or gets restructured, the growth rates we are celebrating this quarter could prove less sticky than they look. We are not bearish on the trend, but the quality of this growth deserves more scrutiny than the headline percentages give it. This is the main reason we are still keeping 13% of the portfolio parked in bonds.

Uber ( ): The bet that still will not cooperate. The quarter was genuinely good: gross bookings of $58.0B, up 24%, trips up 18%, adjusted EBITDA up 33%, and non-GAAP EPS of $0.81, up 35%. The stock fell about 3% anyway, on a Q3 guide decelerating bookings to 18 to 22% and a headline profit flattered by a one-time equity gain.

Friday afternoon update: Uber climbed 6% to close the week around $75, a sign the market may be starting to come back around to the numbers underneath the guide. We keep being right on the business and wrong on the stock, and “less wrong” is the only honest update, because the position did climb this window even as it stayed underwater.

Away from tech, the read-throughs were reassuring.

Accenture ( ): Even though they did not report results, we have to mention Accenture. This is one of our most controversial picks, and on June 18 we doubled our position when Accenture dropped sharply after its last results. That trade is now up 36% in the last 45 days.

Visa ( ) : Posted net revenue of $12B, up 14%, with cross-border volume up 13% and a $0.67 dividend, a clean signal that consumer spending is still healthy.

Blackstone ( ) : Was the encouraging one. Distributable earnings of $2.0B, or $1.52 a share, up 26% and well ahead of expectations, on record assets under management of $1.35T. The fundamentals keep validating the predator-not-prey hold, and the loss is narrowing. The stock has climbed back 30% from its April lows; we are still -10% in our overall position, but we will keep holding.

Four more holdings reported, and they matter.

Reddit ( ) : Is compounding, with revenue of $805M, up 61%, its eighth straight quarter above 60%, ad revenue up 64%, and a Q3 guide comfortably above the Street.

Novo Nordisk ( ) : Gave us the most important read of the window for a struggling holding: revenue near $12B, an earnings beat, the strongest-ever GLP-1 US launch by volume for oral Wegovy at more than 5 million scripts, and a raised full-year guide. It is the first real answer to the moat worry, though management was candid about generic pressure ahead.

Airbus: Delivered 351 aircraft in the half on revenue up 12% and reiterated full-year guidance, so the turnaround stays on track.

Waste Management ( ) : Our quiet anchor, grew free cash flow 34.5% and raised its margin guide, doing exactly the unexciting job we own it for.

There is no doubt that the mega-cap core carried the book. The special-situation satellites are still the drag, but two of the three narrowed their losses this window. That is the update we would want if we were the ones reading it.

Looking Ahead

The season is not done with us. Berkshire Hathaway reports today, eToro on August 11, and Nubank in the middle of the month, and each is material to a position we hold. On the macro side, we are watching three gauges above all others: the Strait of Hormuz, where the reopening talks matter more than the headlines and where we watch shipping data rather than rhetoric, the labor market, which just handed us the first negative print of the cycle, and the inflation reading upcoming on Wednesday.

The market has told us plainly what it will reward: accelerating growth as long as the spending looks disciplined. The open question is whether a cooling jobs market stays a mild tailwind for rates or hardens into a genuine growth scare. We think the balance still tips constructive, and we are positioned that way.

And having caught up three weeks in one sitting, we promise the next letter will be shorter. Probably.

Have a good weekend. Stay hydrated. 💧

Follow our portfolio in real-time on eToro: here

*The views, thoughts, and opinions expressed in this newsletter are solely my own and do not reflect the official policy or position of VTEX or its affiliates. All analysis provided here is based on public information and is for educational and entertainment purposes only.

The most important thing that happened this week is easy to miss under the geopolitics: inflation started cooling in more than just energy. That matters more than any single headline, because it changes our base case for the rest of the year. Three things now point the same way. Inflation is easing into the stickier, services side, not just at the gas pump. The labor market still has not cracked. And the tech earnings now arriving should show hyperscaler revenue and capex still accelerating. Put together, they make another rate hike less likely and earnings-driven upside more likely, which is why we are raising our odds that this market keeps climbing through Q3. This week, we started to position ourselves for it.

Let’s start with the inflation numbers. June core CPI came in flat month over month, below the 0.2% the Street expected, and the annual rate eased to 2.6% from 2.9%. The number that matters most is underneath that: services inflation excluding energy was unchanged, its calmest reading in 5.5 years. That is the tell. This was not just cheaper gas flattering the headline. The stickiest, most domestic part of inflation, the part the Fed has struggled with all year, finally sat still. Cooling that reaches into services is the kind that changes a central banker’s calculus.

The obvious risk to that story restarted this week. US and Iranian forces traded strikes again, Iran targeted tankers in the Strait of Hormuz, and Donald Trump demanded a 20% fee on cargoes passing through. Oil, which had recently drifted back near $70, popped to $87. A sustained energy spike would slow the cooling we just described, and Fed governor Christopher Waller made a point of warning that a hike could still land at the late-July meeting. So we are not waving the risk away.

But here is why we think the cooling holds at these lower levels rather than reversing. Oil at $87 is still well below the peak this same conflict produced earlier in the year, when it ran toward triple digits. The market is already pricing this in. On Polymarket, the odds of a July rate increase collapsed this week from 30% to nearly zero, and the odds of a September hike fell from 53% to 33%. That repricing, in the space of a few days, is the clearest sign that the disinflation read is landing harder than the oil scare.

The next leg of the thesis is earnings season, and it has already started strong. The banks over-delivered out of the gate. JPMorgan posted a 41% jump in profit on blockbuster trading, and Wall Street is on track for its biggest investment-banking fee haul in four years as large deals return. The real test, though, is still ahead, with the hyperscalers. We expect that AWS and Google Cloud will show cloud revenue still accelerating and capex guidance still rising. If they confirm it, the last big question mark over the tech trade into Q3 comes off the board, and a market with cooling inflation, a steady labor market, and reaccelerating hyperscaler earnings is a market that grinds higher. That is the case we are increasingly willing to position for.

In the M&A space, we had major news this week. Uber agreed to a $15 billion takeover of Delivery Hero, up from the roughly $12.5 billion bid we flagged back in May, with Delivery Hero spinning off its Turkish and some European operations to leave Uber the core footprint it wanted. The strategic prize is scale: the deal would let Uber consolidate and own the market-leading delivery operation across dozens of new markets in a single move. It is the same logic we have applied to the autonomous-vehicle story all year: own the platform, not the assets.

Portfolio Update

We made a few moves this week, and it is worth keeping them in proportion.

The one worth naming: we used the cooler inflation read to trim our Treasury buffer, taking it from roughly 20% of the portfolio down to about 13%. That is a reduction of roughly a third, not a wholesale exit. We still hold a substantial buffer, and that is the entire point. Due to the inflation reading and the strong start to earnings, the risk and reward of sitting entirely in cash got a little worse, so we acted.

Where the freed-up capital went was deliberately unexciting. We fed the core index funds first, adding to QQQ, VOO, VXUS, and VGK, which is the discipline regardless of the tape. We added to Microsoft, on the view that the software fortress is a fair-value hold rather than a stretched one, and to Berkshire Hathaway, our low-conviction hedge for exactly this kind of Fed environment. We also topped up Nubank, on continued conviction in the neobank scaling story.

We also opened a small position in Netflix, and it is worth explaining why. The stock fell about 8.6% after hours because it guided third-quarter revenue to $12.86B, against the $13B the Street wanted. Read that again: the actual quarter met estimates, the $3B advertising goal for year-end is intact, and the entire selloff came from a guidance figure barely more than 1% shy of consensus. That is a rounding error dressed up as a disappointment. We think the market overreacted to a great business having one soft-looking line in an otherwise fine report, so we used the drop to start a position at roughly 24x earnings, a level we are happy to own it at. None of these moves is a big swing. They are measured additions funded by a modest buffer trim.

Looking Ahead

Two things decide how the next few weeks read. The first is the Strait of Hormuz, and we are watching shipping data, not headlines. If tanker transits stay near normal and the oil spike fades the way it has twice before this year, the disinflation trend reasserts, and the cooling holds. If the Strait genuinely closes, energy-driven inflation becomes a live problem again right as Warsh weighs QT, and the buffer we kept earns its keep.

The second is earnings season. The banks have set a strong tone, but the real proof point is our two largest long-term theses. We want to see whether AWS and Google Cloud hold their growth rates against tougher comparisons, and crucially, whether the capex guidance comes with matching revenue. If growth and margins hold together, it is the strongest data point yet for the thesis we have run all year, and the clearest green light for the Q3 melt-up case.

The banks got rewarded for delivering this week, and Netflix got docked for hedging. When the megacap tech names report, which side of that line do you expect them to land on? We will find out soon enough.

Have a good weekend! Stay hydrated. 💧

Follow our portfolio live on eToro: here

*The views, thoughts, and opinions expressed in this newsletter are solely my own and do not reflect the official policy or position of VTEX or its affiliates. All analysis provided here is based on public information and is for educational and entertainment purposes only.

No letter last Saturday. Life got in the way, so this one covers two weeks instead of one, and if it feels like a lot happened, that’s because it did.

Both weeks kept circling back to one question: who actually pays for the AI buildout, and what happens the moment the people footing the bill get nervous. Amazon answered by borrowing. Kevin Warsh’s Fed answered by arguing with itself in public for the first time. And Microsoft, Samsung, and SpaceX all gave us a preview of what happens once the market stops taking that bet on faith.

Amazon announced it is raising a fresh round of 25 billion in bonds to fund AI infrastructure, its largest debt issuance of the year, and management says it won’t issue more in 2026. That’s a company generating enormous annual free cash flow choosing to borrow anyway, which tells you AWS demand is outrunning what cash flow alone can fund. Days later, the minutes from Warsh’s first FOMC meeting landed, and the split inside the room was real: a few policymakers wanted to raise rates immediately, while many more said a hike would be “necessary soon” if inflation stays elevated. Layer in May’s PCE print at 4.1%, the hottest reading in three years, and the picture sharpens. The companies building the AI economy are borrowing at a record pace, exactly as the Fed debates tightening into it.

The market’s patience for that bet is thinning, though. Microsoft cut 4,800 jobs, resetting its Xbox division even as it keeps funding AI everywhere else, a reminder that “spend on AI” and “spend on everything” are not the same instruction. Samsung posted a record AI-driven quarter, and its shares still fell 10%, investors punishing capital intensity instead of rewarding it. SpaceX stumbled 6.8% on its Nasdaq-100 debut, tracking almost exactly the prediction we published on May 29 that the IPO pop would fade within weeks. Put together, the market is done rewarding AI spending on faith and is starting to demand proof that it converts into revenue or a defensible moat.

Then, on a separate track, our ceasefire call got tested directly. Trump declared the US-Iran ceasefire “over,” calling further negotiation “a waste of time,” after both sides traded strikes following attacks on tankers, and oil popped on the headline. That lands one week after Morgan Stanley cut its oil forecasts, citing the Strait of Hormuz reopening faster than anyone expected, with tanker transits already back near pre-conflict levels. Our published thesis all year has been “quiet de-escalation by attrition,” not a clean resolution, and this is the most direct test that call has faced. For now, headlines are moving faster than the shipping data. If that gap holds, this fades from the front pages the way it has twice before. If it doesn’t, energy-driven inflation becomes a live problem again, right as a hawkish-leaning Fed decides how hard to lean.

Portfolio Update

We’re up 2.2% over the two-week window, and the year-to-date return now sits at +7.5%. More importantly, since the last week of May, when we moved roughly 20% of the portfolio into the bond buffer, we’ve outperformed the S&P 500 by 0.5%, narrowing part of the gap against our main benchmark.

No positions were opened or closed this period, so this is a held-and-watched update, not a trading one. GOOG remains our largest unrealized gain in the portfolio; BX remains our largest unrealized loss. Both are held, not traded, this cycle.

The one place Core/Satellite discipline is doing real, active work right now is the cash sleeve. Our roughly 19% IB01.L allocation is a deliberate buffer built specifically for a scenario like this one, a divided Fed and a wobbling ceasefire, and it stayed untouched this period. It’s also the position most directly responsible for the benchmark-narrowing move described above.

Looking Ahead

Three things are on our radar for the next two weeks. Warsh just named his monetary policy review task force, including Marc Andreessen, former Walmart CEO Doug McMillon, and former Bank of England Governor Mervyn King, and we’re watching for early signals on how deep H2 quantitative tightening actually goes. We want to see real Strait of Hormuz shipping data, not just rhetoric, before treating “ceasefire is over” as more than a headline. And earnings season starts again in about two weeks, which means real numbers instead of position papers and speeches.

That last one matters most. AWS grew 28% last quarter, and Google Cloud grew 63%, and the first question we’re asking is whether those growth rates hold up against tougher comparisons or start rolling over. The second, bigger question is what management says about capex. Every hyperscaler has raised spending plans through this cycle on the assumption that demand keeps pace, and Amazon’s bond sale only makes sense if that assumption still holds. If AWS, Google Cloud, and Microsoft’s Azure all guide toward further acceleration in AI infrastructure spending without matching revenue acceleration to back it up, that is the exact faith-without-proof pattern we flagged earlier in this letter, playing out in the highest-conviction names in our own portfolio. If growth and margins hold together instead, that is the strongest data point yet for the thesis we’ve been running all year.

Have a good Sunday! Stay hydrated. 💧

Follow our portfolio live on eToro: here

*The views, thoughts, and opinions expressed in this newsletter are solely my own and do not reflect the official policy or position of VTEX or its affiliates. All analysis provided here is based on public information and is for educational and entertainment purposes only.

We got multiple reminders this week that the easy part of 2026 may already be behind us. PCE data for May came in at 4.1% year-over-year, the highest reading in three years, and Core PCE at 3.4% confirmed this was not a rounding error. The timing, however, matters: oil has only just dropped to around $70 a barrel, which means the deflationary tailwind from cheaper energy has not had time to flow through yet. The uncomfortable question the market spent the rest of Thursday wrestling with is whether inflation is this sticky even before we feel any relief from lower oil, or whether that relief is still coming. Regardless, whatever happens in Iran is what will likely define where this goes from here.

Last week, we said the Strait of Hormuz framework would struggle to hold. The situation has deteriorated faster than expected. Since the MOU was signed on June 17, Israel has conducted over 100 air strikes on Lebanon, killing at least 83 people. The ceasefire, for all practical purposes, is not holding. Iran’s foreign minister has called Lebanon “make or break” for nuclear talks -- a significant escalation in rhetoric. The US delegation, led by JD Vance, is unsuccessfully trying to hold things together. It is hard to escape the impression that Trump has handed Vance a problem without a solution and will likely blame him for the fallout. Our reading and prediction have not changed, so it’s worth repeating what we said last week:

“The most likely path is neither ratification nor relapse. It is a quiet de-escalation by attrition: the Strait stays functionally open, both sides let the formal deal fade from the front pages, nobody has any incentive to restart the fire, and the oil relief largely sticks even though no permanent deal ever gets signed.”

In the tech world, Google was in the headlines all week. Gemini 3.5 Flash continues to impress on efficiency, and the broader AI story at the company remains intact. There was some noise on Monday around Anthropic poaching John Jumper -- the researcher who won the Nobel Prize in Chemistry in 2024 for AlphaFold while at DeepMind; and OpenAI poaching Noam Shazeer, one of the authors of the historic paper on transformers. The stock dropped 6% on Monday, but don’t let this change your view on Google. DeepMind is arguably still the most developed and entrenched academic and scientific research bench in the world. The organization has over 1,000 peer-reviewed publications, including 13 papers in Nature and Science -- the two most prestigious scientific journals on the planet. The 2024 Nobel was shared by both Demis Hassabis and John Jumper, both of whom built that research culture from the ground up. Two departures do not dismantle an institution that took two decades to build. The moat here is not any single person. Google’s structural advantage is $180B+ in annual CapEx, custom TPU silicon, and a cloud platform that grew 63% last quarter -- none of that leaves with two people.

As tech steadied mid-week, Amazon gave us some good news: Prime Day’s first day clocked $8.3 billion in US online spending, beating forecasts and marking the single biggest e-commerce day of 2026. The advertising and logistics flywheel is turning.

SpaceX completed its first week as a public company, and the stock is already under pressure. closed around $151, well off its highs of $210 and even below its close on the first day of trading. One more pop may come when the stock enters a major index like the Nasdaq -- passive fund flows create a mechanical bid. But once lock-ups begin to expire, expect the selling pressure to build.

Portfolio Update

The portfolio was down roughly 1.5% on the week. Google was the main drag.

The standout on the upside was , which gained 7% -- the highlight of the week. The Uber story is working: pricing power, expanding margins, and a platform that continues to become harder to displace. We stay long.

Year to date, we are up 5% against the S&P at approximately 7.5%. We are tracking the gap and staying disciplined.

Looking Ahead

The story we are watching most closely going into the second half has nothing to do with any single name -- it is the broader AI valuation picture.

Anthropic and OpenAI are both moving toward public market moments. Anthropic’s last private raise valued it at roughly $965 billion. OpenAI’s implied valuation is in a similar range. Both numbers are predicated on exponential AI adoption curves continuing uninterrupted.

Here is what we think is more likely: companies are starting to optimize. The move from “throw tokens at everything” to “build efficient pipelines” is already underway. As AI budgets get consumed through the first half of the year and CFOs start asking hard questions about ROI, the volume growth story gets harder to tell. Models like Gemini 3.5 Flash are actually accelerating this shift -- efficiency gains mean you can do more with less, which is great for users but compresses the metrics that underpin the most aggressive valuations.

If that narrative shift hits during the Anthropic or OpenAI roadshow, the market reaction could be significant. That is one reason we are watching the broader tech trade carefully as we head into H2.

Have a good Sunday! Stay hydrated. 💧

Follow our portfolio live on eToro: here

*The views, thoughts, and opinions expressed in this newsletter are solely my own and do not reflect the official policy or position of VTEX or its affiliates. All analysis provided here is based on public information and is for educational and entertainment purposes only.

After one week off on vacation, we are back for a packed edition. For the past couple of months, we have been carrying the same two open questions forward: what would Kevin Warsh do at his first FOMC, and how would the Iran conflict be resolved? Both got answered over the last few days.

Let’s take them in order.

Warsh’s First FOMC: Hawkish, On Purpose

On June 17, Kevin Warsh chaired his first meeting as Fed Chair and came out swinging. A hawkish policy statement and nine members projecting a rate hike this year surprised markets. The S&P 500 fell 1.2%, the 2-year treasury yield jumped 11.4 basis points, and rate-sensitive sectors like homebuilders were hit hard. Warsh emphasized restoring price stability and dismissed short-term market reactions.

This is precisely the H2 liquidity drain we started preparing the portfolio for back at the end of May. There was no hike this meeting — but the message was blunt: price stability first, market tantrums second.

The Iran Deal

The second answer arrived the same week. The US and Iran reached an agreement to reopen the Strait of Hormuz, oil fell to multi-month lows, and equities rallied on the relief. There is also reportedly a $300bn reconstruction fund attached.

Here is where we part ways with the consensus. Read the terms honestly, and this is not a US win wrapped in a ceasefire bow — it is a strategic defeat. Iran outlasted the conflict, secured the Strait reopening on favorable terms, attracted a reconstruction windfall, and got the hardest question — its nuclear program — deferred rather than resolved. By any sober scoring, Tehran came out on top. The 14-point deal, as structured, looks politically unsignable for the United States. No US president puts his name to something that reads this much like capitulation. So the market is pricing a resolution that may never actually be ratified.

But — and this is the nuance that matters for the portfolio — unsignable does not mean a return to war. The most likely path is neither ratification nor relapse. It is a quiet de-escalation by attrition: the Strait stays functionally open, both sides let the formal deal fade from the front pages, nobody has any incentive to restart the fire, and the oil relief largely sticks even though no permanent deal ever gets signed. The Strait reopens, the subject quietly changes, and life goes on.

Is that so crazy? We don’t think so. Wars more often end in exhaustion and a changed subject than in a signing ceremony. And let’s be honest — there’s always some other headline waiting in the wings to capture the news cycle’s attention.

The market wants to net these against each other: hawkish Fed (bad) versus easing energy (good). We think that math is wrong.

The energy relief is real and probably durable. But it does not rescue tech multiples, because Warsh’s hawkishness was never about energy in the first place.

That is why we will keep our cash buffer on. Not as insurance against an oil spike we don’t expect, but as insurance against a Fed that stays restrictive even as inflation starts to cooperate.

Accenture: A Brutal Day, and a Quietly Important Pivot

One of our stock picks of the year had a terrible week. Accenture cut full-year revenue guidance to 3-4% growth from 3-5%, citing continued weakness in its US federal business, and the stock dropped roughly 18% in a single session — its worst day since 2016. We were already deep underwater. Radical Transparency: This is our worst-performing stock by percentage, and we own that.

On the same day, Accenture announced a $4.18B pivot into operational technology cybersecurity, acquiring a majority stake in Dragos, plus runZero and NetRise. The strategic rationale is twofold: the OT security market is estimated at $27.39 billion in 2026 and projected to nearly double by 2031, and the software-led, expertise-intensive nature of OT security is one of the segments where AI cannot easily automate away the work. This sits on top of the Whalar creator-economy acquisition from the prior week.

The fear is that AI cannibalizes Accenture’s billable-hours model. Buying software ARR in a domain AI can’t easily commoditize — on the same day the market punished the old model — is the right strategic instinct, even if the timing of the guidance cut buried the news.

So we added on the dip. Bringing our cost basis to roughly 13x forward earnings — about half its ten-year average. Let´s be clear about what this is: we are betting on a deep-value turnaround, not the clean compounding story we first wrote about. Our conviction on Accenture is now Medium. We bought weakness, and if the stock rebounds and bookings stabilize, we will take some profits.

Other portfolio-related news

Amazon is reportedly in talks to sell its custom Trainium AI chips to external data centers, with management saying demand is strong enough that it won’t cannibalize AWS.

Apple/Google: WWDC launched the revamped Siri running on Gemini. Every iPhone upgrade this fall becomes a Gemini activation on top of an already enormous base. Pure validation of the GOOG full-stack thesis, which remains our biggest winner.

Visa is plugging its payment network into ChatGPT to let AI agents transact within spending limits — Visa positioning as infrastructure for the agent economy rather than a casualty of it.

Blackstone, alongside Apollo, finalized a ~$35bn private credit deal to help finance Anthropic’s growth.

Portfolio Update

Over the last two weeks, the portfolio has been up +1.4%. YTD: approximately +6.5%

A genuinely constructive two weeks, with a shape worth understanding. The portfolio rode the Iran relief rally, then gave some back when Warsh went hawkish on the 17th. Netting out: we absorbed a hawkish Fed surprise and still finished up.

We made some minor moves this week.

Closed at +17%. A satellite position that had run its course. We rotated the proceeds rather than letting them sit — the discipline is to recycle a completed thesis into places with more road ahead.

Added to (the dip-add discussed above).

Added to . There is no doubt the stock has been a frustrating hold but we are still very confident in the health of Uber’s business and in the long run we expect this position to be a major winner.

Added to $BRK.B, building the defensive satellite anchor for exactly the QT environment Warsh just confirmed.

Looking Ahead

The live risk is the one the market is not watching: what happens to the Iran deal that can’t be signed. Our base case is that it quietly fades and the calm holds. But if the formal process collapses noisily — if Trump walks publicly, or Tehran overreaches — the headlines get ugly even if the Strait stays open. We’d treat any oil-driven selloff on that news as noise rather than a thesis change, which is precisely the scenario the buffer exists to exploit.

Have a good Sunday! Stay hydrated. 💧

Follow our portfolio live on eToro: here

*The views, thoughts, and opinions expressed in this newsletter are solely my own and do not reflect the official policy or position of VTEX or its affiliates. All analysis provided here is based on public information and is for educational and entertainment purposes only.

We came into this week having made a deliberate defensive move. Not panic. Not a crash call. Just a quiet acknowledgment that the risk-reward had shifted — and a decision to wait for the data to catch up with the narrative. This week, it started to.

The week opened with Jensen Huang on stage at Computex, telling the world that AI agents would consume more software tools, not fewer. The market took that as permission to rally everything. ServiceNow +10%. IBM +12%. Salesforce +6.5%. It was a big Monday. It was also the kind of move that tells you more about sentiment than fundamentals. A market that needs almost no excuse to run is not a market processing information carefully. It’s a market looking for any reason to stay long.

Then came Broadcom on Thursday. Revenue guidance came in slightly below the whisper number. Not below consensus, below the most optimistic expectation. The stock dropped 14% pre-market, wiping over $300 billion in market cap before the US open, dragging chips and $QQQ with it. When you are priced for perfection, the only direction surprises can go is down. This is not a Broadcom problem. This is a valuation problem. Cracks are starting to appear, and it is coming for other names in the space, too.

Friday was the week’s defining moment. The US economy added 172,000 jobs in May. Unemployment held steady at 4.3%. Let’s take a second to appreciate what that actually means: we are two-plus years into the most aggressive AI adoption cycle in history, and the labour market is still adding jobs at a healthy clip. The fears of mass AI-driven unemployment that dominated headlines in 2023 and 2024 look, right now, remarkably overblown. That is genuinely good news for people, for the economy, and for anyone who worried the AI transition would leave workers behind.

For markets, however, it was a different story. A strong labour market gives Kevin Warsh everything he needs to act. The Nasdaq closed down 4.18% — its biggest single-day drop since April 2025. The S&P 500 fell 2.64%, snapping nine consecutive weeks of gains. The VIX spiked nearly 40% in a single session, closing above its 200-day moving average for the first time since the Iran war began. Semiconductors were the epicentre: , , , 6.20%. Bitcoin fell toward $60,000. Rate hike odds on prediction markets crossed above 50% for the first time this cycle.

There was one more detail worth noting: Meta was reportedly weighing its own multi-billion dollar equity raise to fund AI infrastructure — following Google’s $80B offering earlier in the week. Two of the world’s most cash-generative companies, both tapping external markets to fund the AI arms race in the same week. That is not a sign of a cycle with room to run. That is a sign of a cycle consuming itself.

Our Portfolio

Week: -2% | YTD: ~+5%

This was the week the defensive positioning started paying off in relative terms.

The Nasdaq dropped 4.6% on the week. The S&P 500 fell 2.6%. Our portfolio came in at -2% — a full half-point better than the broader market, and more than two and a half points better than the index we are most exposed to. That gap did not come from stock picking this week. It came from the decision made the week before to hold a meaningful cash and bond buffer while valuations were stretched. Boring decisions, visible results.

On the individual holdings, Uber had a quietly productive week that the macro noise drowned out. On Monday, Uber partnered with Autobrains and Nvidia Drive to launch a robotaxi testing programme in Munich — OEM-agnostic, Level 4, pending regulatory approval. Europe becomes the third geography after Las Vegas and Dubai, where the AV marketplace thesis is being built. On Tuesday, Uber restructured its HR division, cutting nearly a quarter of it under new president Jill Hazelbaker, getting leaner as it expands. UBER drifted lower in the broader selloff. The thesis is building, not breaking.

No trades this week. We came in defensively positioned. We stayed defensively positioned.

Looking Ahead

Everyone is talking about the June 16-17 FOMC as the key macro event. We’ve been talking about it for weeks. But the event that sets the table for that meeting arrives first.

Wednesday, June 10: May CPI at 8:30 a.m. ET.

This is the last major inflation reading before Warsh chairs his first meeting. April printed at 3.8% — above estimates, driven by energy. With rate hike odds already above 50% and today’s jobs report removing any excuse for inaction, Wednesday’s number could move markets before the FOMC even convenes. It is the most important data point of the next two weeks.

One broader question to close on: when a strong jobs report — historically a sign of economic health — is the thing that triggers the Nasdaq’s worst day in over a year, what does that tell you about the fragility of the current setup?

That’s the conversation to have next week.

Have a good weekend. Stay hydrated. 💧

Follow our portfolio live on eToro: here

*The views, thoughts, and opinions expressed in this newsletter are solely my own and do not reflect the official policy or position of VTEX or its affiliates. All analysis provided here is based on public information and is for educational and entertainment purposes only.

At what point does a great story become too good a story? That is the question we have been sitting with this week. The S&P and Nasdaq hit new all-time highs, and we have started to prepare for what might be the top. Let's dive into why.

and both surged over 30% in a single session — Dell reported revenue of $43.8B and an AI server backlog of $51.3B. Full-year AI server revenue target raised to $60B. This is not a small beat. This is a company being reborn by AI hardware demand. Snowflake reported that Q1 product revenue grew 34%. They announced a $6B multi-year AWS infrastructure commitment. The stock had been down 19% in 2026 before the print. In one day, those losses vaporised and then some.

Anthropic closed a $65B Series H at a $965B valuation: that is nearly $1 trillion. For a private company. The company’s annualised revenue has shot up from 9 billion to $47B in just 5 months. That is a 500% increase. The trend will likely flatten out, but this is the world´s fastest-growing company, being valued at 20x revenue. Keep that number in mind

Speaking of valuations and revenue multiples, let’s quickly discuss the SpaceX IPO. The numbers are now public, and the company reported $18.7B in revenue last year at a net loss of $4.9B. At the intended valuation, that is roughly 94x revenue on a loss-making business that only grew 15% YoY in Q1. The argument for SpaceX is not earnings. It is narrative — Starlink, Starship, the manned Mars mission. Real achievements. But a company valued at $1.75T is pricing in a future that has to go perfectly for many, many years in a row.

We do not hold SpaceX or Tesla. Both are on our explicit avoid list. But here is our end-of-year prediction: the SpaceX IPO will pop on day one, generate enormous headlines, and make a lot of retail investors feel very smart for about six to eight weeks. Then the valuation will start to crater. There is no earnings anchor. There is no dividend. There is only narrative, and narrative is a depreciating asset once the IPO confetti settles.

When that happens, we predict Elon will move fast. He will argue that the merger of Tesla and SpaceX is a strategic necessity, specifically, that unifying both companies is required to realise the Terafab project. It will be framed as industrial destiny. What it will actually be is a valuation rescue operation. We are not recommending either company. We are just noting that when this is the most exciting story of the cycle, the cycle is probably late.

Back on April 4th, we called the bottom in our newsletter — we said the Iran war panic and the Warsh Fed fear had created a buying window, and we went all-in. The April recovery proved that call right. Now we are saying the opposite. The risk-reward has shifted fundamentally.

Let us be fully transparent about what that means in practice. This week, we deployed approximately 19% of the portfolio into IB01.L — the iShares 0-1 Year Treasury Bond ETF. This is the highest cash and short-duration bond buffer we have held in over 18 months. If markets continue up from here, that 19% earns roughly 4% annualised while the rest of the portfolio earns whatever tech earns. We are accepting that drag deliberately. Kevin Warsh's first FOMC meeting is June 16-17. He has pledged active quantitative tightening. The April CPI was 3.8% — a three-year high driven by Iran war energy disruption. The Strait of Hormuz remains closed. QT without rate cuts, at 3.8% inflation, with equity multiples at all-time highs, and an IPO wave about to hit — that is not a recipe for continued market strength. It is a recipe for a second-half correction. We want dry powder for that correction..

Let's also be clear about what this is not: we are not moving 60% to cash and watching from the sidelines. We still have a large, active exposure to the market, and we are staying bought in. GOOG, AMZN, UBER, RDDT — the core positions are untouched, and we have no intention of changing that. This is a dial, not a switch. If the rally keeps extending into the second half, we may well consider pushing the cash buffer beyond 20% — raising it incrementally as the risk-reward continues to shift. We are not calling the top with certainty. We are managing for the scenario where this turns out to be the top.

Our Portfolio

Besides the cash buffer, this week we added more UBER shares. The Delivery Hero bid — €11B, representing Uber’s continued push to dominate European food delivery alongside its autonomous vehicle strategy. This is a global mobility and logistics marketplace. The market keeps mispricing UBER, so we added more on the intra-week dip.

had a strong recovery this week, swinging up over 18%. The AI training data economics story reinforced why we hold it: Reddit’s $60M API deal was cited this week as a benchmark in a rapidly emerging “AI data as a sellers’ market” narrative. Combined with 60% ad revenue growth in Q1, we continue to hold with high conviction.

Our YTD return stands at approximately +8%, while the S&P 500 sits at around +10.5%. That gap is largely explained by our real economy bets — ACN, NVO, BX, IHI — which are a deliberate hedge against the inflation and rate environment we have been calling all year. They are dragging near-term performance and providing structural insurance at the same time. We are not complaining. +8% in five months is excellent. We are not being greedy. And that is precisely why we are comfortable moving even more conservative now — this is not a panic, it is a plan.

A Look Ahead

June 16-17: Warsh’s first FOMC decision. This is the most important date on the macro calendar. The market is not fully pricing in active QT. If Warsh signals faster balance sheet reduction than expected, the reaction in tech multiples will be fast and meaningful. We will have cash ready.

The Iran situation. Oil fell 3% mid-week on reports of a draft Strait reopening framework. Then, US forces launched new strikes. The ceasefire is not done. If it closes — and we think it eventually will — we will be watching to see if the drop in oil prices leads to lower inflation or not.

Have a good weekend. Stay hydrated. 💧

Follow our portfolio live on eToro: here

*The views, thoughts, and opinions expressed in this newsletter are solely my own and do not reflect the official policy or position of VTEX or its affiliates. All analysis provided here is based on public information and is for educational and entertainment purposes only.

Here is the question that is becoming the most important one in tech right now: What does AI actually cost?

The first wave of enterprise adoption was all about experimentation. Pilots, proofs of concept, and innovation budgets are approved with a loose definition of ROI. That phase is ending. Companies are now sitting on AI spend that has to justify itself, and the conversation is shifting to something much more uncomfortable: cost per token, at scale, across the whole organisation. That shift changes who wins the AI race. And it is exactly why Google’s annual developer conference this week mattered more than the headlines suggested.

Google I/O took place on Tuesday, and it was one of the most significant keynotes in years. Sundar Pichai announced that Gemini now has one billion monthly active users, effectively overtaking OpenAI in monthly active users. Token throughput has hit 3.2 quadrillion per month, up 7x year-on-year. A new personal AI agent called Gemini Spark runs 24/7 on dedicated cloud servers. But the announcement that caught our attention most was quieter than all of that: Gemini 3.5 Flash, the new workhorse model, runs at 4x the speed of its predecessor at less than half the cost of frontier rivals. That is the real story. Google controls the full stack from custom chips to cloud to the model itself, which means they can make AI cheaper in ways that OpenAI, Anthropic, or even Microsoft simply cannot match. When CFOs start asking how much this actually costs to run at scale, Google is the only one in the room with a truly compelling answer.

NVIDIA () also reported its Q1 2026 earnings this week, and the numbers were extraordinary — revenue up 85% year-on-year, another record quarter, with data center demand showing no signs of slowing. And yet the stock fell after-hours. Nvidia is now priced at a forward P/E of less than 25, even below the S&P average. This is because the market is looking ahead to 2027 and 2028, wondering what will happen when every hyperscaler buying Nvidia chips today is quietly building custom silicon to replace them. The cost curve argument applies here too — the direction of travel in AI infrastructure is toward cheaper, purpose-built chips, and Nvidia’s long-term ceiling is becoming visible even as the near-term numbers stay extraordinary.

Then there is SpaceX, which filed this week for what could be the largest IPO in history, at a rumoured $1.75T valuation on roughly $15B in revenue. Our view has not changed: stay away. There is no fundamental anchor for that number. Flight 12, the first v3 Starship, launched last night, and even though it was spectacular as always, the flight was far from perfect. Watch the spectacle. Do not buy the narrative. We have been following SpaceX closely for nearly 15 years now — I still remember vividly the first Falcon 9 landing back in 2015 — and if there is one thing we have learned, it is that space is slow. Data centers in space are a far-fetched dream that will take decades to become a reality.

On the macro side, the EU cut its 2026 growth forecast down to just 0.9%, with the Iran war’s energy shock still working its way through European corporate margins. The more interesting data point was the UK, where CPI actually fell to 2.8% in April, with government energy subsidies absorbing the Hormuz shock better than expected. Japan came in at 1.4% for similar reasons. For our VXUS and VGK positions, the EU headline is uncomfortable, but lower valuations and a deflationary surprise out of the UK keep us holding. The unlock for European equities is still the same thing it has been for months: the Strait of Hormuz reopening. Every week it stays closed is another week the ceiling on European stocks stays artificially low.

Portfolio Update

This week: +0.57% | YTD: ~+7%

In other news related to our portfolio, Blackstone had a great week after the $5B Google TPU joint venture was announced on Monday. This is exactly the kind of move that explains why we have not sold — Blackstone is not a distressed private credit firm, it is a capital allocator that just bet big on AI infrastructure alongside Google. Jon Gray’s call for their best-ever year of IPOs in H2 still stands. We are holding.

As we flagged last week, we have now opened a new position in Berkshire. Greg Abel’s first quarter as Berkshire CEO was genuinely interesting. He exited 16 positions — including Visa, Amazon, and UnitedHealth — and tripled Berkshire’s Alphabet stake to nearly 58 million shares. We took that as confirmation rather than coincidence, and opened a small position in Berkshire Class B this week. The $300B+ cash pile that Berkshire holds is a genuine advantage in an environment that is likely to get rough in the second half of the year for stocks.

Looking Ahead

The thing sitting at the top of our mind going into next week is not a stock; it is the calendar. We are 25 days from Kevin Warsh’s first FOMC meeting as Fed Chair, and he is not Powell. He has pledged quantitative tightening, he has a 3.8% CPI print giving him cover to act, and a jobs market with no cracks giving him no reason to rush a cut. The Iran situation offered a small sign of life this week — Trump called off a new attack at the request of Gulf states and mentioned serious negotiations underway — but we have heard that before.

What are you watching most closely going into June? Drop a comment below.

Have a good weekend. Stay hydrated. 💧

Follow our portfolio live on eToro: here

*The views, thoughts, and opinions expressed in this newsletter are solely my own and do not reflect the official policy or position of VTEX or its affiliates. All analysis provided here is based on public information and is for educational and entertainment purposes only.

April’s US inflation print landed like a verdict. The reading came in at 3.8%, a three-year high, with energy accounting for more than 40% of the monthly increase. Gasoline was up 5.4% in April alone. Americans are now paying an estimated $1 billion per day in extra fuel costs compared to before the Iran conflict began. Core CPI hit 2.8%, well above the Fed’s 2% target. This is the Iran war’s economic impact arriving, and it’s not small.

The timing of all of this matters. First, because there are still no signs of the Strait of Hormuz reopening, and second, because Kevin Warsh was confirmed as Fed Governor this week, he is on track to become the next Chair, replacing Jerome Powell. Warsh has pledged active quantitative tightening, meaning the Fed is about to switch from passive to aggressive on balance sheet reduction. For QT not to hurt equities, it would need to be accompanied by interest rate cuts. But with the labour market showing no major cracks and inflation moving in the wrong direction, rate cuts appear to be completely off the table.

We’ve been talking for weeks about the tale of the two economies, and this week was no different. While the macro narrative was filling headlines about inflation, the AI infrastructure machine kept moving. Anthropic is reportedly weighing a fundraise at a $1 trillion valuation, potentially surpassing OpenAI, whose ARR Anthropic has already overtaken. Alphabet broke Berkshire Hathaway’s yen bond record by issuing ~$3.6B in bonds specifically to fund AI infrastructure. New research confirmed that agentic AI workloads consume 1,000 times more compute than a standard chatbot query, and it is becoming increasingly common to see reports of employees engaging in what is called “token maxing”: artificially inflating their AI usage to signal productivity at work, which, if anything, only adds fuel to the underlying demand curve.

None of this surprises us. The AI build-out remains structural and mid-cycle. But we’d be doing you a disservice if we didn’t flag something: the market has largely priced this in. With a wave of AI-linked IPOs on the horizon, we need to at least plan for the possibility that the narrative will peak right as the IPO window opens, not because the underlying technology fails, but because that’s historically what markets do. Institutional investors first need to pass these valuations on to retail investors, and only then will the bubble be allowed to correct. We’ve seen this script before.

To finish out the week, Trump’s Beijing summit with Xi concluded without major public announcements, sending Treasury yields above 4.5% and Nasdaq futures down over 1%. Xi’s warning that Taiwan is the core of the entire US-China relationship rattled markets. But here’s what we think the market is missing: the CEOs of Nvidia, Apple, and Elon Musk were all in that room. It is very unlikely they left without securing meaningful benefits for their companies. Keep an eye out next week for announcements related to these companies gaining better access to the Chinese market.

💼 Portfolio Update

Weekly performance: -0.87%. YTD: +6.6%.

Nubank ( ): -14.8%. NU reported Q1 earnings on Wednesday that were, objectively, excellent. Revenue surpassed $5 billion for the first time. Net income up 41% year-over-year. ROE of 29%. Mexico hit break-even this quarter. US banking charter conditionally approved. And the stock dropped nearly 15% across the week with no identifiable fundamental trigger. This one goes in the “price action we respectfully disagree with” file.

Trade this week: We added to Accenture ( ) at around $159/share, continuing to build our enterprise AI consulting position on weakness. ACN’s Q1 confirmed a meaningful surge in AI contract bookings, and earnings are still forecast to grow. The stock now trades at a forward P/E below 12, historically cheap for a business at the intersection of enterprise transformation and AI adoption. Companies will not implement AI alone. This is precisely why we like ACN at these levels.

🔭 Looking Ahead

The inflation trajectory is what sits at the top of our watchlist, and there is really only one thing that can meaningfully move it in the near term: the reopening of the Strait of Hormuz. Energy is driving over 40% of the current inflation reading. Gas prices are expected to exceed $5/gallon this summer. A reopening would be the single biggest deflationary catalyst available right now, and the market would respond immediately.

In the equities space, Berkshire Hathaway is currently the only company we are actively evaluating. The stock has been essentially flat for nearly two years, but more importantly, it offers valuable diversification. Aside from a shared position in Google, there is virtually no overlap between their holdings and our portfolio. Furthermore, in a macroeconomic environment defined by quantitative tightening (QT) and high inflation, Berkshire's massive cash reserves are a distinct advantage rather than a liability. This defensive, cash-heavy posture perfectly aligns with the strategy we have advocated throughout this newsletter, making Berkshire an increasingly compelling prospect.

Do you see the 3.8% inflation print as a temporary Iran-war artifact, or the beginning of something stickier? Let us know in the comments.

Follow our portfolio live on eToro: here

Have a good weekend. Stay hydrated. 💧

*The views, thoughts, and opinions expressed in this newsletter are solely my own and do not reflect the official policy or position of VTEX or its affiliates. All analysis provided here is based on public information and is for educational and entertainment purposes only.

The S&P 500 and Nasdaq closed at record highs this week. Ships exchanged fire in the Strait of Hormuz. A federal court struck down the President’s tariff policy. And the jobs report just came in with 115,000 payrolls added in April, beating the 55,000 forecast. Our portfolio ended the week up +1.9%, bringing our year-to-date return to +7.5%. Having said that, the caution is real: a war in the Gulf, record highs that leave little room for error, and a Fed transition nobody can fully read. But this week, the data made the thumbs-up hard to argue with. Here’s why.

The AI capex wave keeps growing

We’ve been watching the AI infrastructure buildout for over two years now, and the numbers keep getting bigger in a way that feels almost unreal. New data this week showed that AI infrastructure investment was responsible for two-thirds of US GDP growth in Q1 2026, surpassing the record set by the internet boom in 1999. We just broke the dotcom era’s record for the largest tech contribution to GDP growth in history. The difference is that this time, there is no sign that the hyperscalers are slowing down. According to new analysis, the combined free cash flow of the major hyperscalers is expected to collapse to just $4 billion in Q3, their lowest level since 2014. In other words, despite generating vastly more revenue than ever before, they are ploughing virtually every dollar back into AI infrastructure. AMD’s 38% revenue surge this quarter and Nvidia’s $500 million infrastructure deal with Corning are just further signals of a capital deployment cycle that is still accelerating, not peaking. The market is hitting records because the future expected capex and revenue of these companies keep going up at an astonishing pace.

A federal court ruled the President’s 10% tariff illegal

This week, a panel of federal judges found that the administration misused an obscure provision of the International Trade Act of 1974 to impose a 10% tariff on all imports. Importantly, this is not the first time the tariffs have been struck down. We covered an earlier Supreme Court ruling back in February, and at the time, we asked the question that has now been answered: What happens to the tariffs that have already been paid? They are being refunded. The administration launched the CAPE portal in April to process $166 billion in tariff repayments, with over 56,000 US importers already registered. The scale of what’s coming back is significant: Walmart is due $10.2 billion, Target $2.2 billion, Nike $1 billion. This is not a hypothetical tailwind for corporate margins anymore; it is already in motion. A second court ruling now makes any return of broad tariffs even harder to sustain legally. With gas prices remaining elevated due to the ongoing Iran conflict, a deflationary legal shock to trade policy is exactly the kind of offset the Fed needs to hold its current course without being forced into action.

The April jobs report just dropped

April came in at 115,000 nonfarm payrolls, comfortably ahead of the 55,000 consensus forecast, with the unemployment rate holding steady at 4.3%. This is about as close to a Goldilocks print as the Fed could hope for: enough jobs to show the economy is not breaking, with wages slowing enough to keep inflation in check. We remain genuinely torn on what AI is actually doing to employment; the displacement story has not materialised anywhere near as badly as predicted, but the continued decline in information sector jobs (-13,000 in April) is a quiet trend we are watching closely. But the direction of travel is encouraging: a resilient labour market gives the new Fed leadership more room to be patient, which is exactly what we want to see heading into the second half of the year.

Portfolio Update

This was a full hold week, no new buys, no sells, and the portfolio rose +1.9%, bringing our year-to-date return to +7.5%.

What carried us: Our tech satellite positions delivered the week’s alpha. led the portfolio with +8.89%, catching the tailwind from the broader AI chip narrative and the Corning deal. gained +4.7% as the Nasdaq surged to record highs. Alphabet added +3.39%, reflecting its all-time high this week. Our European defense position, Airbus (AIR.PA), quietly gained +3.2%.

What dragged: Reddit ( ) pulled back -4.19% as risk-off sentiment on Friday reset some of the week’s gains in more volatile tech stocks. Waste Management (WM) fell -4% despite no negative company news. Visa (V) slipped -3.2% following news that the UK’s Financial Conduct Authority opened an antitrust probe into payment processors, a risk worth monitoring but not a thesis-breaker for the position. Full story here.

Looking Ahead

Being at +7.5% YTD with seven months still to run is a good position to be in. This is the moment when a disciplined investor should start getting a little more conservative, not more aggressive.

Our single biggest priority over the coming weeks is rebuilding our strategic cash buffer back to the 15–20% target. After going all-in during the February/March correction and riding the April recovery, we are currently sitting at minimal cash levels. The plan is to rebuild that buffer systematically into short-term bond ETFs — specifically $IB01.L, which currently yields around 3–4% while we wait for the right entry points.

Follow our portfolio live on eToro: here

Have a good weekend. Stay hydrated. 💧

*The views, thoughts, and opinions expressed in this newsletter are solely my own and do not reflect the official policy or position of VTEX or its affiliates. All analysis provided here is based on public information and is for educational and entertainment purposes only.

What a week. What a month. April has officially gone down as a historic month for our portfolio. We landed a return of +11% for the month, clawing back from being down earlier in the year to sit securely at +5% YTD.

The overarching story right now is a massive divergence in the global economy. The “old economy” continues to wrestle with sticky inflation and a massive energy shock from the Iran war. Meanwhile, the tech sector is living in an entirely different reality, powered by an unstoppable wave of AI capital expenditure and unheard-of growth from the hyperscalers. This week was basically the Super Bowl for earnings as all major tech companies reported results, and the AI story has never been stronger.

Let’s dive into how our favorite stocks performed:

• Amazon ( ): Our top pick for the year is on fire. The market finally realized its AI infrastructure spend is yielding massive real-world revenue. AWS growth accelerated to 28%, and North American business margins have started to expand just as we predicted on January 31st. The stock surged 30% in April alone. We talked about how Amazon at $200 was a bargain, and at this point, it is unlikely we will ever see the stock back at those levels.

• Alphabet ( ): Google delivered a monster earnings beat driven by Google Cloud, which grew a staggering 63%. During the earnings call, they also discussed how AI models are driving increased engagement and ad spend in the core business. Even better were the CFO’s comments that they plan on growing CapEx even more in 2027, validating the thesis that the AI buildout is far from over. The Big Four confirmed they are on track to spend a combined $650 billion on AI infrastructure in 2026, and Google is proving that these investments generate immense returns. Google climbed 11% this week alone and is now well-positioned to become the largest company in the world at some point this year, a prediction we made back on November 22nd.

• Microsoft ( ): Microsoft delivered a solid earnings beat driven by Azure’s 38% growth. More importantly, they made headlines by clarifying their agreement with OpenAI, effectively securing ownership of the tech until 2032.

• Reddit ( ): To close out the week, we received amazing results from Reddit on Thursday afternoon. Their ad business is still growing at over 60% per year, and the stock climbed 15% on Friday alone.

• Meta ( ): The only lowlight in earnings week was Meta, who announced earnings and saw the stock drop. Thankfully, we were completely right to avoid it. We have been staying away from Meta due to mounting legal liabilities. As we discussed previously, social media is facing its “Big Tobacco” moment. The market’s fear of further legal problems will likely continue to weigh on Meta for several years.

Portfolio Update

This is the type of week that makes us proud. While our core, low-cost index funds ( , ) climbed less than 1% this week, our single-stock bets carried the day with total portfolio performance up nearly 3% for the week. Our structural long bias towards tech (>50%) and our outsized bets on Google and Amazon are exactly why we outperformed.

But outside of tech, several of our holdings also reported this week. Here is how the rest of the Bliss portfolio held up:

• Visa ( ): Earnings generated a solid pop (+8%) as revenue climbed, driven by increased consumer spending. This serves as a forward indicator that consumers are still spending despite inflation fears.

• Waste Management ( ): Solid earnings results showed that the business is strong and growing at a steady pace. This is precisely why we like the trash business: it’s predictable, necessary, and always delivers.

• Airbus: Not all news was great in the old economy. Airbus earnings missed the mark, and the stock dropped, highlighting the ongoing supply chain and production struggles in aerospace.

Looking Ahead

The AI party is booming, and we are riding the wave.

All we need to finish landing a perfect earnings season is a good announcement from Uber ($UBER) on Tuesday. We have been accumulating for months now, and we are sitting tight for the robotaxi story to play out.

Apart from that, if the market continues to rise over the next few weeks, we will resume building a strategic cash buffer to prepare for future dips.

Follow our portfolio live on eToro: here

Have a good weekend. Stay hydrated. 💧

*The views, thoughts, and opinions expressed in this newsletter are solely my own and do not reflect the official policy or position of VTEX or its affiliates. All analysis provided here is based on public information and is for educational and entertainment purposes only.

For the second week in a row, the market basically moved sideways, and we closed the week down 0.4%. The broader panic around the Middle East has largely settled. While we wait for the details of a long-term resolution to the conflict, the market’s attention has clearly shifted back to familiar territory: inflation and the ever-looming impact of AI on the labor market. It feels like déjà vu from January, where every week we were discussing macroeconomics, central bank drama, AI circular deals, and the SaaSpocalypse. Honestly, it’s a relief to be back.

Here are the main news stories of the week:

• Fed Chair Senate Hearing: Kevin Warsh, the nominee to succeed Jerome Powell, faced his first Senate confirmation hearing. Pushing back against White House pressure, he explicitly pledged independence, assuring lawmakers he won’t be anyone’s “human sock puppet.” More importantly, Warsh was crystal clear about his intention to ease the Fed balance sheet and actively remove money from the economy. Looking forward, we are hyper-focused on his target pace for this quantitative tightening, because a rapid liquidity drain will act as a very strong force against equities in the second half of the year. Hopefully, any reduction in the Fed balance sheet will be accompanied by a reduction in the interest rate to soften the blow to the economy.

• Blackstone’s Q1 Beat: We are currently sitting on a roughly 16% unrealized loss on Blackstone thanks to the liquidity panic ripping through the private credit markets. But this week, they announced an earnings beat, with profit surging 25%, proving exactly why we aren’t panic-selling. While the headlines focused on President Jon Gray forecasting the firm’s “best-ever year” for IPOs, we are looking under the hood. Their private credit fund’s 0.4% default rate is holding steady. More importantly, with $190 billion in cash, Blackstone isn’t a victim of this credit storm. They are the predator waiting for others to default so they can swoop in as the lender of last resort. Speaking of Gray’s prediction for a record IPO year…

• SpaceX buys Cursor: Elon Musk’s SpaceX announced a staggering partnership with AI coding startup Cursor, complete with a provision to acquire them for $60 billion. This is a massive positioning move ahead of SpaceX’s rumored $1.75 Trillion IPO. Cursor, combined with xAI has the potential to field a true rival against Anthropic in the Enterprise coding market. Having said that, our view on the IPO remains the same: stay away.

• Apple’s New CEO: Tim Cook’s long-expected departure was announced this week. Cook will step down as CEO on September 1, 2026, passing the baton to John Ternus. Cook has been a steady, diligent operator who built Apple into the world’s most effective supply chain powerhouse. We fear Apple is now a company guided more by supply chain instead of product design, and we hope Ternus brings back a spark of much-needed product innovation. Cook’s legacy should be defined by smash hits like the Apple Watch and AirPods, but we feel he will be remembered most for the Apple Vision Pro disaster and for his disciplined decision to break away from the rest of big tech and not enter the trillion-dollar AI CapEx arms race.

AI CapEx is still accelerating, and it has gotten us thinking about who will eventually pay the price. Spoiler alert: it’s you and me.

The End of your cheap AI Subscription

There is a reason we are seeing this frantic rush of circular deals and AI IPOs. The time has come for these companies to prove that they can turn a profit, which means the era of the cheap subscription and VC-subsidized AI is about to come to an end.

Earlier this month, Anthropic restricted the viral OpenClaw agent tool, forcing users to pay up or get cut off. It was a wake-up call for the industry. Between 2024 and 2029, research firm Gartner estimates that capital investment in AI data centers will reach a jaw-dropping $6.3 trillion.

To avoid an unmitigated disaster for investors, these AI labs need to generate a return on that capital. To hit even a baseline 7% return, the industry needs to earn close to $2 trillion a year by 2029. To do that, AI companies will need to not only grow in the enterprise market but also end their generous B2C flat-rate tiers, switching to metered token billing, and squeezing developers to actually start paying for the immense compute they consume.

The takeaway? We are currently living in the golden age of cheap, heavily subsidized superintelligence. Take advantage of it now to automate your workflows, build your side-hustles, and upskill yourself, because the bill is coming due. Bliss has spent 14 million tokens over the last 30 days optimizing his open-source app, and we are confident that the cost to Anthropic was a lot higher than the price of our monthly subscription.

Portfolio Update

While our aggressive “Buy the Dip” strategy during the Q1 panic engineered an 11% surge in April, this week brought a slight reality check. We are currently sitting at roughly +3% YTD.

During the week, our global ETF satellites slightly underperformed. Our core Finance sector holdings were the worst-performing sector in the portfolio, down about 4% on the week. The highlight again was Amazon. The stock has been on a massive run-up, closing up another 5% this week; the stock is now up 30% in April.

Looking Ahead

As we look toward next week, earnings come into full view with our two largest holdings, Google and Amazon, reporting results. We are watching these closely for very different reasons:

• Amazon is currently our Top Pick, and we want to see if their AI infrastructure spend continues to yield massive AWS revenue.

• For Google, we are looking for forward guidance on ad revenue and for any updates on how they plan to navigate the massive legal precedents forming around social media and product design liabilities. We discussed this a couple of weeks back in this newsletter.

Expect the labor market and general macroeconomic data to regain total dominance over market sentiment. Right now, we are sitting tight and not moving, but if we see the market rise back over the next few weeks, we will start building a cash position again.

If you enjoyed this week’s breakdown, do us a huge favor and share it with a friend who is still relying on a $20 AI subscription.

Have a good weekend. Stay hydrated. 💧

Follow our portfolio live on eToro: here

Want to track your own finances as we do? Check out our open-source Bliss app and drop us a star on GitHub if you like it: https://github.com/danielvsantos/bliss

*The views, thoughts, and opinions expressed in this newsletter are solely my own and do not reflect the official policy or position of VTEX or its affiliates. All analysis provided here is based on public information and is for educational and entertainment purposes only.

VTEX Day is always a blast because we get to reconnect with the ecosystem and see the live reactions to our latest tech. This year felt truly special. Not only did VTEX unveil a massive wave of new AI-powered commerce products during the Vision keynote, but we also saw the official launch of the #BrazilianEngineering manifesto. It’s a push to show the world that Brazil isn’t just about commodities; it’s a global powerhouse for software, technology, and applied intelligence. From Embraer jets to VTEX’s commerce platform, Brazilian engineering is one of the most disruptive, creative, and efficient in the world.

But while we were running around the event, the markets decided to pull off an absolute miracle. And before we get into why, let’s look at the numbers. We made another 5% this week alone. We are on an absolute tear. Just like that, the portfolio is back up 3.6% YTD, and an insane 11% in April alone. Our aggressive “Buy the Dip” approach from February and March is officially paying off.

Here is what drove the craziness this week:

• The Iran Rally: Tensions with Iran continue to cool down. It is getting increasingly rare to hear talk about attacks in other Gulf states. While shipping lanes in the Strait of Hormuz haven’t normalized just yet, life in the region is starting to pick back up. The markets have apparently assumed that we will not go back to full hostilities and U.S. bombing campaigns. The market loves certainty, and equities absolutely took off.

• Earnings Season Got Off to a Great Start: U.S. banks started to report Q1 earnings this week, and most large banks soundly beat expectations, driven by record trading revenue, strong capital markets activity, and a resilient consumer.

• Amazon is hitting new records: A few months ago, Wall Street was hyperventilating over how much money tech giants were burning on AI. That panic is over, and Amazon is the poster child, peaking into record territory this week. Investors are realizing Amazon’s heavy AI infrastructure investments are already translating into massive, sustainable revenue growth for AWS, and that the boom in Anthropic will be massively beneficial for Amazon.

• The Software Rebound: Tech is bouncing back, but stock picking actually matters again. Microsoft completely broke out this week as a top performer after a brutal couple of months. If you want further proof that software is back, look no further than the IGV ETF (which tracks the broader software sector)—it popped an incredible 14% this week alone.

Portfolio Update

• Sold: We sold our position in Digital Ocean. We locked in a massive 37% gain over just two months. We love letting our winners run, but when a short-term satellite thesis plays out perfectly, you take the cash and walk away.

• Upped: We bought more Uber. It didn’t pop as aggressively as the rest of tech this week, but we are playing the long game here. They keep quietly making massive inroads in their push towards Autonomous Vehicles (AVs), and we are happy to keep accumulating.

Looking Ahead

Earnings season will kick into high gear over the next couple of weeks, and we are watching closely to see if the AI revenue numbers hold up.

What are you buying right now? Do you think this rally holds, or is it a mirage? Drop a comment or reply to this email to join the discussion.

Follow our portfolio live on eToro: here

Have a good weekend. Stay hydrated. 💧

*The views, thoughts, and opinions expressed in this newsletter are solely my own and do not reflect the official policy or position of VTEX or its affiliates. All analysis provided here is based on public information and is for educational and entertainment purposes only.

This is easily the biggest week in the short history of the Bliss ecosystem.

We aren’t just celebrating a massive 4% portfolio rip or the fact that the markets finally stopped holding their breath over the Iran ceasefire. The real news is that I’m finally revealing a “secret” project I’ve been obsessively building behind the scenes for over a year. Before the rest of the world sees it on LinkedIn next week, I wanted to give you all a sneak peek.

But first, let’s look at what is happening in the markets.

The Market Climbs Back Fast

We made over 4% this week alone. After a difficult February and March, we are now down just 2% for 2026. Our patience is paying off, proving exactly why trying to time the bottom or actively managing a portfolio during crises is a loser’s game. Most of the money in this market is made on a handful of high-velocity days; if you get caught on the sidelines, you miss the rip.

As you all know, the reason the market is climbing back is the optimism around a ceasefire in Iran. Back on March 21st, we wrote:

“There is really only one way out of this mess and luckily, that relies on Trump protecting his own ego. We just have to hope he realizes this isn’t going anywhere, announces that the war has been won, and goes home.”

Thankfully for all parties involved, it looks like this is what will happen. The ceasefire news acted as the exact relief valve the market had been waiting for. We certainly expect more threats and posturing from the administration during negotiations, but the market’s reaction confirms that most investors expect the worst of the war to be over.

Amazon’s Visionary Future

Andy Jassy’s shareholder letter dropped Thursday, it’s definitely worth the read and confirms what we’ve been saying for months: Amazon isn’t just an e-commerce giant; it’s an AI infrastructure beast. In response, is up 14% this week. The market is finally waking up to the fact that their investment in Anthropic, which just passed OpenAI with a $30 Billion run rate, is the trade of the decade.

Introducing Bliss: The open-source, self-hosted personal finance manager

You all know that finance interests me. What you might not know is that for the past 15 years, I have been tracking my finances down to a level of detail that many people around me have called “a cry for help”.

I’ve categorized over 30,000 transactions with every single expense since 2010 across multiple currencies, building complex spreadsheets that allowed me to manage my finances as if I were a company, with complete P&L statements and multi-account portfolio tracking

About a year ago, with the advances in AI coding agents, I started building a system to replace my Excel sheets. By the time I realized it, I was thinking of developing the solution as a product and launching a business around it.

About a month ago, it occurred to me that the best use of this software would be to make it available to everyone as a free, open-source, self-hosted platform. I have truly poured everything I know about software into this project. After 15 years of architecting web systems and roughly 800 hours of development over the last year, this is my representation of what modern financial software should look like. Bliss comes with;

• Multi-Currency Financial Summary: A normalized view of your net worth across USD, EUR, BRL, and more.

• Real-Time Portfolio Tracking: Live integration with global market data to track equities and satellite positions.

• AI-Powered Classification: A 4-tier waterfall engine that categorizes transactions with enterprise-level accuracy.

• AI Financial Insights: Contextual reasoning over your data to identify trends spreadsheets miss.

• Comprehensive Expense Dashboards: A clean, high-velocity UI designed to navigate 30,000+ transactions.
  

Bliss is built on a modern technical foundation that ensures your data remains your own. It’s multi-tenant ready, composable by design, encrypts your data, and has an enterprise-grade CI pipeline with over 1000 automated tests. All of this comes in a well-documented, self-hosted package that anyone can run on their own hardware.

Building this also taught me how to weaponize the modern AI toolchain, from Cursor to Antigravity and, more recently, Claude Code. I’ve developed a “Spec-Driven” methodology that I’ve since used to rapidly develop tools for a host of solutions, both personally and in my role at VTEX. Ill be sharing more details on this methodology on linkedin in the coming weeks.

I’m officially taking the repository public on LinkedIn this Monday. But today, I’m inviting you to get the exclusive first look:

• See it in action: Access the live demo with mock data at app.blissfinance.co Navigate around and let me know what you think.

• Explore: Landing page and API references at www.blissfinance.co

• Drop a Star on GitHub: Help us build momentum before the world sees it on Monday. https://github.com/danielvsantos/bliss

If you’re on LinkedIn this Monday and see the announcement, a comment or a share would mean the world.

With the launch of this app, the Bliss financial ecosystem is now complete: The newsletter, the public portfolio, and now the tool. I hope you enjoy it.

Portfolio Update

Back to our eToro Portfolio. This week, we only made one move. We officially sold our Nebius ( ) stock. After a 30% climb this week, we walked away with a 60% total profit. We love the company, but when a satellite position hits that kind of velocity, we harvest the gains.

Follow our portfolio live on eToro: here

Have a good weekend. Stay hydrated. 💧

*The views, thoughts, and opinions expressed in this newsletter are solely my own and do not reflect the official policy or position of VTEX or its affiliates. All analysis provided here is based on public information and is for educational and entertainment purposes only.

Finally, a good week for our portfolio. We made 4% this week, partially clawing back some of our Q1 losses. We are now sitting at -6% YTD. We are still trailing our benchmark, the S&P 500, which is currently at -4% YTD. But the feeling is that the tide is turning, and now the question is:

Has the market hit bottom for the year?

We believe so. Earlier this week, as expectations grew that the conflict in Iran might be coming to an end, we got a glimpse of what happens to a coiled spring market: the Nasdaq exploded higher by 3.83%, and the S&P 500 rose 2.91% in Wall Street’s best session since February.

But it was too good to be true, and on Wednesday, Trump gave a primetime speech stating the U.S. would “hit [Iran] extremely hard over the next two to three weeks.” At this point, what we described as “Scenario A” in last week’s newsletter is now looking more and more likely: A short-term escalation followed by an abrupt exit with no boots on the ground.

On the macroeconomic front, the data is starting to show the true impact of the war. Eurozone inflation jumped to 2.8%, fueled by the energy crisis, putting pressure on the ECB to consider rate hikes. Furthermore, the IMF revised global growth down to 3.1% for 2026. Another important factor for our portfolio, and for investors at large, is that this geopolitical escalation is proving once again that the U.S. Dollar is still the ultimate haven. This week again, we saw gold, silver, bitcoin, and other major currencies all struggling to hold their value against the dollar.

On the bright side? The March jobs report came out on Friday, and there were no signs of mass AI layoffs. The report revealed a surprising 178,000 workers were hired in March, and the unemployment rate dropped to 4.3%. The data is not amazing, but at this point, as long as we avoid the doomsday scenario of AI layoffs, we see this as good news.

A Historic AI Leak

Speaking of AI, for the third time in the last month, Anthropic makes the newsletter. But this time, it’s due to one of the largest blunders in recent tech history. During an update, they accidentally published a map file that allowed users to download the source code for Claude Code.

In a matter of days, the code has spread widely across the internet, with dozens of free copies popping up. Anthropic’s legal team is currently scrambling, firing off aggressive DMCA takedown notices to GitHub and other hosts in a frantic attempt to scrub the leaks from the web. But as anyone who has used the internet knows, you can't put the toothpaste back in the tube. The experience offered with Claude Code is what allowed Anthropic to leapfrog the competition over the past three months, and now it is out in the open. I’m sure the teams at Google and OpenAI couldn’t be happier right now, and I definitely expect these features to make their way into their products in the next couple of weeks.

Meanwhile, Microsoft isn’t wasting time trying to catch up. They are jumping back into the frontier model race, as they announced three new proprietary “MAI” foundational models via their Foundry.

Space is Back!

Space exploration is back in the news in a big way. NASA successfully launched the historic Artemis II mission, taking humans for a round-trip around the moon for the first time in over 50 years. Crucially, this will likely be NASA’s last moon mission before kicking off a new era in the Artemis program led by the booming American private space industry.

Speaking of the private space industry, SpaceX officially filed for what is expected to be the largest IPO in history. We have been die-hard fans of SpaceX for many years now, religiously following their developments. I am a true believer that Starship and the Raptor engine are the pinnacle of human engineering—but there is no financial justification for the valuations currently being floated. $1.5 trillion for a company with $15 billion in revenue? At this point, all we can say is that we recommend staying far away from the IPO.

Instead, we prefer Amazon (). Amazon’s LEO satellite network (Project Kuiper) is picking up heavy momentum with another Atlas V launch sending an additional 30 satellites just today, alongside a new partnership with Delta Air Lines to provide in-flight Wi-Fi, directly challenging Starlink’s dominance.

Portfolio Update

This week, we held all positions steady, riding the rebound. We want to give a massive shout-out to the 5 assets in our portfolio that have maintained a positive return year-to-date: Individual stocks DigitalOcean () , Nebius () , Waste Management (), and ETFs and .

Here are some other news stories from the week that strengthen the core arguments we shared in our 2026 playbook:

• Uber (): Making massive strides in the Autonomous Vehicle space. WeRide and Uber officially launched their driverless robotaxi service in Dubai, signaling global momentum.

• Nubank (): Getting ready to make waves with their U.S. entrance. Following their heavy advertising investments with Mercedes in F1, they are stepping into the physical limelight this Saturday with the opening of the new NU Stadium for Inter Miami.

• Novo Nordisk () : In the highly competitive pharma space, they announced a novel subscription model for Wegovy, aiming to lock in patients and aggressively defend market share against Eli Lilly.

Looking Ahead

We will continue to ride the “buy the dip” momentum while monitoring the macro landscape. (And hopefully, Bliss the Capybara can keep doing his happy dance).

What are your thoughts on the SpaceX IPO valuation? Are you buying in or staying away? Let us know in the comments.

Follow our portfolio live on eToro: here

Have a good weekend. Stay hydrated. 💧

*The views, thoughts, and opinions expressed in this newsletter are solely my own and do not reflect the official policy or position of VTEX or its affiliates. All analysis provided here is based on public information and is for educational and entertainment purposes only.

It’s been a dizzying week in the markets. This has been the worst week for equities since the start of the war. Heavy selling pressure has dragged both the S&P 500 and the Nasdaq more than 10% below their all-time highs. We are now in official correction territory, sending both indexes all the way back to where we were in August. Bliss the Capybara is officially hiding under his desk.

The most dominant story is obviously the war and the oil crisis. But before we dive into the geopolitical chaos and the latest whiplash from the White House, there are two massive tech stories impacting markets that we need to discuss.

1. Social Media’s “Big Tobacco” Moment

On Wednesday, a massive legal bombshell hit social media companies. A Los Angeles jury found Meta and Alphabet liable for a 20-year-old woman’s severe mental distress caused by social media addiction, awarding her $3 million in damages. The monetary amount in this single case isn’t what matters; it’s the legal precedent.

For decades, social media companies have hidden behind Section 230 of the Communications Decency Act, claiming they aren’t responsible for user-generated content. However, the plaintiffs in this case successfully bypassed this shield by arguing something entirely different: it’s the design of the product itself. They argued that the infinite scroll, autoplay, and algorithmic feeds act as a dangerous, addictive machine.

During the trial, internal memos surfaced proving that Meta knew these features acted like poker machines on adolescent brains, yet they continuously prioritized engagement over safety. With more than 2,000 similar lawsuits waiting in the wings across the country, the floodgates for legal liabilities are wide open. Stocks across the social media sector dropped heavily, with Meta alone shedding over 8% on Thursday.

What happens next? We believe the most likely outcome is a wave of massive settlements, followed by social media companies being forced to limit the use of their products for anyone under the age of 16. The era of unchecked growth at the expense of youth mental health is facing its first real threat. Many countries have already banned social media for early teens, and we think Meta and Alphabet will have to get behind age-gating efforts to avoid even larger repercussions. Expect a future where social media is as unreachable for children as cigarettes.

2. Anthropic’s Continued Disruption & The “Mythos” Leak

Tuesday brought another reality check for enterprise software. Anthropic launched a “computer use” functionality for Claude, allowing the AI to take direct control of a user’s desktop to autonomously complete workflows. Combined with a feature called Dispatch that allows you to talk to Claude from your phone, these features basically emulate a lot of the functionality that made early AI agents like OpenClaw go viral in the first place. These launches triggered yet another sell-off across the enterprise software sector.

We covered this impact a few weeks back in Panic in the Cloud: Is software dead? Nearly two months have passed, and there is still no evidence that these features are negatively impacting the revenue and earnings of major software players. SaaS companies with a deep data moat still seem safe and genuinely undervalued.

On Thursday, though, we got a whole other story. A leaked blog post revealed Anthropic’s next-generation model, “Mythos,” is nearly ready for launch and vastly outperforms its predecessor, Opus 4.6. The leak explicitly detailed how Mythos possesses such advanced autonomous capabilities that it poses a massive, systemic cybersecurity risk. The market reaction was swift and brutal. Cybersecurity stocks tumbled as investors questioned whether we are about to enter an era where every user with a Claude subscription is capable of a significant cyberattack.

3. The Geopolitical Standoff: Three Scenarios for the War

Now, back to the macro picture. Last weekend, markets were bracing for a massive energy shock when Trump issued a strict 48-hour ultimatum to obliterate Iran’s energy infrastructure if it didn’t allow safe passage in the Strait of Hormuz. Then, abruptly on Monday, he postponed the ultimatum for 5 days, citing “productive conversations” and a “significant prize” regarding oil. Iran denied any ongoing conversations. On Thursday, Trump announced again that he is postponing any attacks on energy sites for another 10 days.

Our Take: The market’s trust in this narrative and in anything Trump says at this point is essentially zero. Iranian officials aggressively denied that any negotiations were taking place. As we discussed in last week’s issue, State of the Market: A Spiraling War, one thing is clear: the US wants out of this conflict really badly, but Trump’s ego makes it hard to just walk away.

If you listen to Admiral Cooper´s latest updates on X, he is clearly hinting that the US is close to its military objective, mentioning that 92% of the Iranian navy is gone, as is 2/3 of their missile production capabilities. We believe there are three possible outcomes from here, with very different impacts for the world. Let’s go over them from most optimistic to pessimistic:

• Scenario 1 (The Swift Exit): The US and Israel launch a massive wave of attacks as one last salvo. Trump just stated at an event in Florida that there are exactly “3,554 targets left” to hit in Iran and that the US will get it done “pretty quickly.” In this scenario, the military ramps up to wipe out those remaining targets, claims victory in the next 7-10 days, and immediately exits to a defensive posture in the Gulf. This scenario yields the least amount of casualties and the most positive impact for financial markets.

• Scenario 2 (The Slow Fade): “Talks” start to happen and drag on for weeks or months while the attacks slowly lose steam. The cost of the war is such that the US gradually decreases its effort until the conflict just dies out. In this scenario, markets remain under pressure for several more weeks but eventually start to recover as the strait gradually opens, thanks to the involvement of other nations. The humanitarian toll is massive as the war drags on in Iran, and other countries largely dependent on oil coming from the strait start entering a state of emergency (e.g., the Philippines). Today, this seems like the most likely outcome.

• Scenario 3 (The Worst-Case): Trump authorizes the use of ground troops to take possession of Kharg Island. If that happens, buckle in.

What scenario do you think is most likely? Do you believe in some other outcome? Let us know in the comments.

Portfolio Update

Here is what we are doing with our own money, following our Core/Satellite strategy while we wait for the volatility to pass.

• Bought the Core: We executed our monthly Dollar Cost Averaging (DCA) strategy this week. Despite the indices falling into correction territory, we stick to the plan. We deposited fresh cash this week and allocated it directly into our main Core ETFs ( , , ) to maintain our broad market exposure.

• Bought More Uber ( ) & Reddit ( ): On the satellite side, we used the tech pullbacks to add to our positions in Uber and Reddit. Both companies continue to show strong structural momentum, and we are happy to accumulate shares at a discount.

Looking Ahead

We are strictly in the mode of waiting and executing our monthly buys. The market will swing wildly on every White House press briefing, but our job is to ignore the noise and stick to the schedule.

Follow our portfolio live on eToro: here

Have a good weekend. Stay hydrated. 💧

*The views, thoughts, and opinions expressed in this newsletter are solely my own and do not reflect the official policy or position of VTEX or its affiliates. All analysis provided here is based on public information and is for educational and entertainment purposes only.

A lot has happened since we last spoke. The year has officially turned sour, and our portfolio is down about 7% YTD. However, that is in line with the Nasdaq and about 1% below the S&P 500's drop. Let’s dive right into the main news stories of the week and our view of how they are impacting financial markets.

The News: War, Inflation, and a $1 Trillion Silver Lining

• The War: Two more weeks have passed, and the war doesn’t seem any closer to finishing. The Iranian regime is still defiant, there are no signs of a popular uprising, and it is becoming obvious that it is impossible to keep this going much further. The economic and political cost of a protracted effort—including damaged F-35s and aircraft carriers needing repairs—is simply not something the U.S. can bear right now. Trump dramatically miscalculated here; from the public’s reaction to the regime’s resilience, this will go down as a massive blunder. Markets initially shrugged off the economic impact until it became clear that the Strait of Hormuz (which handles roughly 20% of global daily oil consumption, or ~21 million barrels) would remain closed for the foreseeable future. The situation became much more dramatic on Wednesday when Israel and Iran started trading attacks on critical oil and gas infrastructure, ratcheting up pressure on Gulf states. There is really only one way out of this mess, and luckily, that way out relies on Trump protecting his own ego. We just have to hope he realizes this isn’t going anywhere, announces that the war has been won, and goes home. His administration had initially promised a 4-to-5-week total effort, which is about 10-15 days away, so let’s hope they stick to that timeline, simply declare victory, and exit.

• Labor and Inflation: Plenty of economic updates also came out. On the labor front, we got mixed signals. The February jobs report was bad, but jobless claims remain low, meaning the market is essentially in a strict “no hire, no fire” mode. For all the AI apocalypse prophets out there, mass layoffs should have made a true dent in the labor market by now, but that simply is not happening. As we’ve been saying for the past month, every week the labor market doesn’t crater is a massive win. Inflation, however, is a whole other story. The recent CPI reading is concerning, and it is directly tied to the war effort and the resulting energy shock. Wholesale inflation climbed 0.7% month-over-month and is now solidly over 3%. If oil prices stay elevated for another 4-6 weeks, the damage will be nearly impossible to ignore, and inflation will climb even higher. On the other hand, some argue that inflation can cool down quickly. Yes, that assumption relies on the war ending (which would cause oil prices to drop), but also because we are exactly one month away from the anniversary of the tariff impacts. Once those tariffs hit the one-year mark, they drop out of the Year-over-Year inflation calculation, creating a possible sudden drop in headline inflation.

• Central Banks Hold the Line: Unsurprisingly, this macroeconomic data has forced central banks to pivot. The Fed, ECB, and BoE are officially moving away from the narrative of impending interest rate cuts, with the Fed holding steady this week. The Australian central bank (RBA) even increased rates. The market is now pricing in just one rate cut from the Fed through the end of the year. Central banks are going to be cautious. In light of the geopolitical drama, they will opt to simply “wait and see,” leaving rates higher for longer, which will prove to be a drag on markets, specifically the financial sector.

• The Private Credit Panic: Another news story gaining traction is the impending crisis in private credit markets. Let us explain what’s going on. Following the 2008 financial crisis, traditional banks retreated from corporate lending due to heavy regulations. Alternative asset managers eagerly stepped into the void, fueled by a historic zero-interest-rate environment. What started small has ballooned into a massive $1.7 trillion market. But the era of free money is over. The alternative asset space has been rocked by a severe drawdown over the last few weeks, with heavyweights like Blackstone ($BX) seeing 30% drops. This was triggered by the downturn in the software sector, combined with rising default rates among larger companies. This downturn even led BlackRock and Blue Owl to put limits on how much money investors can redeem from their funds, which only further fueled panic. Over the past two weeks, we have studied this topic in depth. While there is definitely an elevation in market risk, we don’t think this is 2008 all over again. We are especially confident in the prospects of our core holding, Blackstone. More on that below.

• The AI Boom is Alive and Well: We do have one incredibly positive note. This week was the Nvidia GTC conference. Jensen Huang took the stage and officially committed to a massive $1 trillion backlog in orders for their Blackwell and next-generation Vera Rubin chips through 2027. For those who watched the event, it really was a feel-good story; the real-world impact of AI is starting to become tangible. The examples shown of how AI is advancing education, medicine, and general robotics were truly inspiring. It was also amazing to see the rate of improvement in the power efficiency of these AI systems (with newer chips operating at a fraction of the energy cost of their predecessors). Everyone talks about power being the bottleneck for the AI revolution, but they forget to account for how fast efficiency is improving. The build-out of the world’s AI infrastructure is truly just getting started. We are staunch believers in this and have invested in the companies that are set to benefit most. Yes, the world is a messy place, and war, inflation, and corrupt politicians might dominate the news, but in the background, the greatest technological revolution the world has ever seen is advancing rapidly.

The Bottom Line: Right now, we do not see much of a place to hide from this short-term volatility besides cash and bonds. Even gold dropped sharply to close out the week and is down more than 20% from its all-time highs. Lots of people have run to oil companies, but we believe that is too volatile, trading entirely on whatever happens in the Middle East on a day-by-day basis. For the past three weeks, our positions have remained the same: Don’t be like our friend Bliss, who is trying to hide behind gold & oil. At this point, it is best to just Do nothing.

Portfolio Updates

Despite the drawdown and all the geopolitical drama, we had some good news over the past two weeks. We are continuing to see news that validates our 2026 playbook:

• Uber ( ) is the Future of Robotaxis: Uber delivered a massive catalyst, perfectly aligning with our AV predictions. They just announced a multiyear partnership with Amazon’s Zoox to deploy fully autonomous robotaxis on the Uber app in Las Vegas this summer, and in Los Angeles by mid-2027, alongside a massive $1.25B investment in Rivian. Uber just keeps piling up partnerships with everyone possible. The AV market will be very crowded in the future, but Uber will remain the marketplace and extract most of the value.

• Nebius ( ) and the Growth of NeoClouds: Our NeoCloud bet is paying off beautifully. Nvidia announced a $2 Billion investment in Nebius to build 5 gigawatts of hyperscale AI cloud infrastructure by 2030. Nebius is positioning itself as the ultimate cloud platform for the agentic AI era. Besides this, they received a massive $27 billion contract from Meta. We love the NeoCloud stories, but we like Nebius’s position as the leading player from the EU even more.

• Accenture ( ) Earnings: Their consulting business is alive and well, and they are starting to see a tangible AI boom in their enterprise contracts. We had predicted that companies would run to Accenture to “AI enable” their businesses, and this week’s earnings showed exactly that. Accenture increased its revenue outlook for the year while also improving margins.

• Holding Firm on Blackstone ( ): Blackstone took a massive hit recently amid the private credit panic but has since regained part of its value, bouncing up 10% this week. Blackstone is a $1.27 trillion fortress. Private credit is an important part of their business, but based on our calculations, it is far from an existential threat. Our thesis for buying Blackstone was based on their real estate exposure (data centers) and their private equity gains ahead of an upcoming IPO boom. We are holding.

Note: If this whole story on the private credit market interests you, please reach out to me, and we can discuss it further. It is a topic that I’ve studied in depth, but it’s too boring to put in the newsletter.

Looking Ahead

The market is testing everyone right now. But we built this portfolio for professionals seeking long-term capital appreciation. This quarter has definitely not been our best performance-wise, but looking on the bright side, we have made amazing purchases. We are buying high-growth, generational companies for the same prices they traded at nearly 6 months ago (like Nvidia).

Follow our portfolio live on eToro: here

Have a good weekend. Stay hydrated. 💧

*The views, thoughts, and opinions expressed in this newsletter are solely my own and do not reflect the official policy or position of VTEX or its affiliates. All analysis provided here is based on public information and is for educational and entertainment purposes only.

It’s been a heavy, dark week. With the outbreak of conflict in Iran and the massive disruptions across the Gulf countries, we frankly wanted to avoid talking about markets altogether. It feels trivial to look at P/E ratios when there is a real human toll unfolding. To anyone stuck in transit, waiting out airspace closures, or trying to get home to their families, we are sending you our best wishes. Let’s keep this one short and dive right in to the most important stories of the week.

• Impact of the war: Markets opened sharply lower on Monday following the start of the war. As the reality set in that this conflict will likely drag on for weeks, oil surged to close the week at over $90 a barrel. The market followed its usual wartime playbook: oil climbs, the dollar strengthens, defense stocks rally, and everything else bleeds. This time was no different. The most visible casualty was the aviation sector. With Middle Eastern airspace shut down, airlines like United, Delta and IAG took an absolute beating. If we have one takeaway here, it’s this: whether it takes one, two, four, or eight weeks, this war will end. When it does, we expect oil prices to fall sharply and fast. That’s exactly why we are staying far away from oil stocks.

• Economic indicators: Friday morning delivered a brutal blow. The US economy lost 92,000 jobs, and unemployment ticked up to 4.4%. Normally, this screams “rate cut!” But with oil spiking due to the war, the Fed is stuck in a nasty corner. We will now have to wait and see the true inflationary impact of the war before discussing any further rate cuts.

• Dario Amodei’s Masterclass: While the world watched the Middle East, the US tech sector experienced a massive earthquake. It all started when the Pentagon demanded Anthropic remove its safety guardrails and allow the government to use Claude for mass domestic surveillance and the development of autonomous weapons. CEO Dario Amodei drew a hard line, said no, and in a single move might have become the most influential person in the AI world.

  • The Retaliation: The government then designated Anthropic a “supply-chain risk”, a label usually reserved for foreign adversaries like Huawei. Amodei went public with his defiance, alleging this was clearly retaliation from the government and that Anthopic would challenge this in court. In a CBS interview worth watching, Amodei laid out his rational arguments and came out looking like the good guy (just take a look at the comments on the video).

  • OpenAI’s Mistake: Following this announcement, Sam Altman swooped in with one of the dumbest moves we’ve seen, announcing OpenAI had reached an agreement to supply the Pentagon. It backfired spectacularly. Nearly 900 employees across Google and OpenAI signed an open letter protesting the deal, forcing Altman into a messy PR damage-control mode amid a trending “Delete ChatGPT” campaign.

  • The Unintended Boom: By refusing to bend on its ethical red lines, Anthropic and Amodei executed the greatest marketing move in AI. Consumers and enterprises aggressively rallied behind the “trustworthy” alternative, and Claude surged to #1 on the App Store. Adding to the good news, on Thursday, news broke that Anthropic’s ARR has skyrocketed from $10 billion in December to a staggering $19 billion in just two months. OpenAI, on the other hand, seems to be hovering around $25 billion, growing 17% since December. This is amazing news for Anthropic, but also for our top holding, . Amazon owns a massive equity stake in the company, and AWS stands to be the primary beneficiary of Anthropic’s rapid growth over the last few months.

Portfolio Update

We deposited fresh cash this week. Instead of panic-selling into a war, we spent the money doubling down on our previous predictions.

• Bought the Dip in Non-US ETFs: and dropped sharply amid the global sell-off and strengthening of the dollar. We took advantage of the extreme volatility to add to our positions.

• We allocated more capital to Airbus, Nvidia ( ), and Blackstone ( ). Blackstone has continued to fall sharply amid fears of a bubble in the private credit market. More on that in next week’s newsletter.

A look ahead

Right now, our only strategy is to “wait and see.” We aren’t making any sudden moves, and our primary hope is simply that this war ends swiftly with minimal human suffering.

What are your thoughts on Anthropic’s stand against the Pentagon? Hit reply and let us know.

Follow our portfolio live on eToro: here

Have a good weekend. Stay hydrated. 💧

*The views, thoughts, and opinions expressed in this newsletter are solely my own and do not reflect the official policy or position of VTEX or its affiliates. All analysis provided here is based on public information and is for educational and entertainment purposes only.

We are closing out the month of February, and this week was no different than the rest of the month. Massive market volatility driven by Trump tantrums, AI fears, and tech earnings.

Let’s get into this week’s news:

• Tariffs & AI Doomsday: Both our portfolio and the market spent the entire week recovering from Monday’s drop. The drop was driven by two factors: President Trump announcing his 10% retaliatory tariff after the Supreme Court ruling, and a viral blog post from Citrini Research outlining a 2028 “Global Intelligence Crisis” (an AI doomsday scenario predicting massive white-collar unemployment).

• NVIDIA’s perfection wasn’t enough: Tuesday and Wednesday, markets climbed in anticipation of NVIDIA earnings. Perfect results were delivered; they beat expectations and raised guidance, but the stock still dropped 5%. Why? Because the market is not convinced that the AI buildout will continue at this pace. When investors are uncertain, they look for stable, boring businesses, and this week we saw another glaring example of this irrationality: Walmart and Nvidia are currently trading at the exact same trailing P/E ratio. Take a minute to really think about that.

  • Walmart ( ), a business that is predicted to grow 5% YoY with a 3% net margin, is trading at a 45 P/E and a 42.7 forward P/E.

  • Nvidia ( ), the best business in the world, which has a 55% net margin and is projected to grow 77% YoY, is trading at the same 45 P/E but a wildly cheaper 22.7 forward P/E. Investors are so terrified of the unknown that they are willing to pay AI-level premiums for the absolute certainty of toilet paper and bananas. bliss is having a hard time wrapping his head around this one.

    

    Remember: The market can stay irrational longer than you can stay solvent. In these times of uncertainty, only invest in individual stocks money that you won’t need in the short term.

• The $840 Billion OpenAI round: Just as the Friday closing bell rang, news broke that OpenAI secured an $730 billion pre-money valuation in a monstrous $110 billion funding round. Amazon is throwing in $50 billion, SoftBank added $30 billion, and Nvidia pitched in $30 billion. We are still digesting what this means for Amazon. We previously mentioned we didn’t want to see Amazon get involved, but in a separate press release, it was announced that they are getting massive spend commitments on AWS, and even a commitment to train models on Trainium 2 chips. If that is the case, AWS growth should accelerate even further, which will be a strong tailwind for Amazon( ) in the coming quarters.

• The AI Labor Paradox & Inflation: Also on Friday, we got a mix of economic indicators. On one side, jobless claims came in under expectations, showing that AI layoffs are not exploding, but at the same time, we saw Block announce a massive AI-driven layoff, cutting 40% of their workforce. In the inflation world, PPI came in above expectations, showing that wholesale inflation is still sticky around the 3% mark. In general, these indicators reduce the likelihood of further Fed rate cuts before June.

• Netflix Walks Away: Netflix smartly withdrew its bid for Warner Bros. Discovery. We wrote back in November that this was a great acquisition for Netflix, but everything has a price. To a certain extent, it’s highly refreshing to see that Netflix ( ) executives are not willing to buy assets at any price, and that is exactly why the stock climbed on the news.

Portfolio Update

We spent all week trying to recuperate from Monday’s losses. We were nearly there, but a Friday drop dragged us back down, closing the week at -1.1%. The main culprits? Nubank and Novo Nordisk, which were down 15% and 20%, respectively.

• Nubank ( ) : Nu absolutely crushed earnings. Revenue grew 45% YoY, and revenue per active customer grew 27% year-over-year to $15. Delinquencies dropped to 6.6%. The kicker? They received conditional approval for a U.S. banking charter for 2026. Despite this, the stock dropped 15% this week. This drop does not worry us at all. We are very long on Nu, and this quarter just reinforces our thesis.

• Novo Nordisk ( ) : We took a massive beating here, down 15% for the week. Shares plummeted to a five-year low after its new obesity treatment, CagriSema, showed less weight loss in trials than Eli Lilly’s Zepbound. Unlike Nubank, this drop worries us. We are starting to lose faith in Novo, but we aren’t capitulating yet because at this point the company is trading at a forward P/E of 12. We are holding the line while we reassess the weight-loss drug war.

A Look Ahead

This morning, breaking news came in that the US has launched a massive military operation in Iran. Crypto prices are already dropping heavily across the board in response to the geopolitical shock, indicating that we could be in for another terrible Monday.

What do you think about the Walmart vs. Nvidia valuation, or the new $840B OpenAI price tag? Let us know in the comments.

Follow our portfolio live on eToro: here

Have a good weekend. Stay hydrated. 💧

*The views, thoughts, and opinions expressed in this newsletter are solely my own and do not reflect the official policy or position of VTEX or its affiliates. All analysis provided here is based on public information and is for educational and entertainment purposes only.