Here’s a preview of what we’ll cover this week: 

Macro: Don’t Worry About The US Economy; Is AI replacing employees?

13F Highlights: The Rotation Happened Inside Semis; Hedge funds Bought Quality; The Money Went To Microsoft; Hedge Funds Know AI Needs Power

Markets: Collegium: Management Agreed With Us In Cash; Western Digital: Mean-Reverting Quality

Lumida Curations: Aswath Damodaran on Meta’s Trillion-Dollar AI Capex Bet; Travis Kalanick on Building the Operating System for Atoms

I remember the Dotcom cycle vividly. 

The Netscape IPO was the wakeup call. 

I remember Gateway and Dell becoming household brands. 

Books-A-Million went up a gazillion percent on Thanksgiving week. 

3Com spin-off and Goldman’s IPO were the top signal. 

I remember partners at McKinsey day-trading stocks. 

I remember the book ‘The New New Thing’ by Michael Lewis. 

I remember everyone wanted to be a VC. 

I remember going to a lavish Global Crossing event and wondering why they were letting a college kid like me in. 

(Answer: ‘It’s all about eyeballs’). 

I remember Frank Quattrone at Credit Suisse was a legend. 

So were Mary Meeker and Henry Blodget. 

I remember the investment banker at JP Morgan not knowing what Geocities did. 

I remember VerticalNet saying their insane valuation meant that the market was telling them to do M&A and thinking how strange that sounded. 

I remember DoubleClick and the birth of Silicon Alley in NYC and Flatiron Partners. 

I remember Wall Street investment bankers quitting to join startups. 

I remember Krispy Kreme going nuts in its IPO.  

I remember company issued press releases announcing a website causing stocks to surge. 

Where are we now?

Markets remember history. 

There is far more discipline in public markets today. 

There is also far more earnings growth, margin, and productivity growth. 

This is not a bubble. 

Too many of the leading companies have reasonable valuations matched with strong earnings growth.

We do not have a euphoria or ‘permanently higher plateau’. 

We have skepticism and a wall of worry. 

Check out the below snapshot in the newly launched ‘Feed’ Section of the Lumida Invest app (which triggered this post). 

The feed is an infinite scroll that focuses solely on curated market insights.

If you are a markets junkie, you should NOT visit the feed.

We are embedding curations of thoughtful podcasts, too, so we can get more of our time back.

Download the app at www.lumidainvest.com.

On a side note, I’ll be in San Francisco next week, and would love to connect with the Lumida Tribe there. Drop me a DM on Telegram, or reply to this newsletter if you’re available.

13F filings came out this week.

These filings update automatically in the Lumida Invest app.

We have also created a 13F grid that highlights the top positions, additions, and reductions of over 70 hedge fund managers in Q2 2026. We also note the key insights around popular stocks, and whether they were added, sold, or retained as top positions.

You can download the grid here.

These 13F filings tell you how managers were positioned around June 30.

That is right before at midyear mark, after which we saw the mighty rotation unfold.

Here are a few themes that were clear across hedge fund positioning.

The filings tell an interesting story around semis.

Hedge funds weren’t selling everything inside semiconductors. 

The managers rotated from the expensive supply chain names to the cheap cloud compute stocks.

Druck also sold memory stocks before July's semiconductor crash. He exited Micron, Broadcom, Bloom Energy, Sandisk and Lumentum.

Atreides (Gavin Baker) closed Lumentum and GlobalFoundries entirely and cut Astera Labs by nearly two-thirds.

Brad Gerstner’s Altimeter also closed Broadcom entirely.

We saw the same at Renaissance, which dumped Micron - again, very well timed.

Now, look at what those same managers did not sell.

Nvidia stayed. TSMC stayed. And, in several books they added.

Nvidia is now the most popular holding amongst the 13Fs we track.

Amazon deserves a mention as well. Druckenmiller made it 3% of their portfolio in this quarter. Jane Street, Coatue, and Millennium also added significant sizes.

Funny thing, while these 13Fs are adding Amazon, Jeff Bezos is selling the stock, and he sold $5B of it.

We think the stock has likely topped here, but it can run on fundamentals. The revenue growth is backed by consistent FCF generation.

We noted before these 13F drops that we thought Nvidia and Microsoft might be this years Google. The HF managers seem to agree, and both have recovered.

Nvidia still has good value and we can see the name getting to $275. Microsoft likely earns a ‘market return’ after its 40% rally. We trimmed recently to fund positions such as AppLovin.

Brad Gerstner’s Altimeter held Nvidia at 20% of the entire portfolio, its single largest position.

Renaissance also added aggressively to Nvidia, taking the position to $1.4B, its largest position now.

Brad Gerstner also added to TSMC, now a 7% position in his portfolio.

We like TSMC. Taiwan Semiconductor is like Scottie Pippen to Michael Jordan, aren’t they? Reliable player, never really gets the credit it deserves. (Jordan is obviously Nvidia.)

So, managers were not selling AI. They were selling names that had run above their fundamentals, and were due for a correction.

We think Altimeter got his NVDA and TSM calls right.

Nvidia is trading at its cheapest forward earnings multiple in five years, while FCF yield is at the highest of its 5-year range.

Revenue is estimated to grow 82% and EPS 88% in the current fiscal year.

Return on equity is 114%.

PEG is 0.51x against sector PEG of 1.5x. 

You are being asked to pay an index-level P/E multiple for a business growing earnings at 88% with triple-digit returns on capital. 

That is not a bubble multiple. That is a business whose fundamentals have outrun its share price.

TSMC is the same argument with a wider moat and a lower profile.

There is exactly one firm on earth that can fabricate the leading-edge silicon the entire AI buildout depends on. And, that’s Taiwan Semiconductor.

Revenue is estimated to grow 40% with EPS increasing 60%.

They have 40% ROE, with a below-index earnings multiple of 19.2x.

The forward P/E sits right in the middle of its five-year band, about 20% off its peak.

The investment thesis is clear here. 

Another interesting selection for HFs in semis was Oracle. 

Multiple hedge funds initiated or added to their positions here.

Point72 tripled its existing position, taking it to about 1% of their portfolio. Millennium management also doubled their last quarter’s position.

Ken Griffin, AQR, and Jane Street also bought the stock. 

Funny thing, Lumida also bought it, but we bought it cheaper than all of them, and right around the bottom. 

We don’t nail it all the time like that (in fact, we bought Microsoft too early in December of last year when it tagged the 200 DMA, then we cut it.).

The yellow indicates 30th June - the cut-off of 13F filings. The white arrow is our buy. 

This suggests Oracle still has more room to go.

The fundamental case for Oracle relies on OpenAI’s financing and the reality of the $600B+ backlog.

As soon as Nvidia’s $250B financing news went live, the backlog started to look real. I did an FSD stream, titled Nvidia is the lender of last resort, discussing what Nvidia’s financing to OpenAI means for markets.

Oracle’s revenue is estimated to grow 33%, with return on equity of 53%. 

The stock trades at Forward P/E of 18.7x with a PEG of 0.64.

A 33%-growth cloud franchise at 18.7x forward is not a demanding ask.

Warren Buffett’s Berkshire added 40% to their Delta Air Lines position.

Druckenmiller’s Duquesne family office tripled its United Airlines stake, making it 2.5% of their portfolio weight.

Quality Stocks

Bill Ackman's Pershing Square opened a brand-new position in S&P Global (SPGI) after the stock had fallen about 28% in 2026 alone.

A common buy in quality was Morningstar (MORN). We have been bullish on the stock since March 2026, when we first wrote about it. 

Cliff Asness's AQR Capital more than doubled their MORN position. Paul Tudor also increased their position by about 90%.

Jane Street, Ken Griffin's Citadel also added meaningfully.

The valuation is the punchline for Morningstar.

It trades at a forward P/E of 16.1x, lowest of its 5Y history.

The FCF yield is around 6.4%, highest of its 5Y range.

Now the business quality underneath it.

Earnings are growing double digits, and are projected to grow 25% in 2026. Return on equity is 32%.

And the number that should stop you: a 12.6% buyback yield.

Morningstar is retiring more than an eighth of its shares a year while trading at the bottom decile of its historical multiple.

This is a subscription data and ratings business. Recurring revenue, embedded in client workflows, effectively impossible to rip out once installed.

It is almost the same toll-booth economics Ackman just paid up for at S&P Global — available at a materially lower multiple.

We own Morningstar.

Microsoft saw massive buying across hedge funds we track. 

Bill Ackman’s Pershing Square added, taking it to 12% of the portfolio, a top 3 position. 

Altimeter, Coatue and Whalerock added. 

So did Tudor Investments, Jane Street, and Citadel.

We have been writing frequently about Microsoft, and it has paid back well. 

It is up about 41% since the quarter started.

We have started to trim Microsoft now.

Despite the move, Microsoft's forward P/E of 25.1x is still at lower levels of its 5Y history. That tells you how far the stock had dropped.

The fundamentals of Microsoft don’t need much mention.

Return on equity is 34%. Revenue is estimated to grow 18% and EPS 14%.

You are getting the most durable enterprise franchise in software at a multiple the market has rarely offered. Read our analysis of their latest earnings report.

Microsoft's free cash flow conversion is at the low end of its own history because it is spending enormously to build AI capacity.

So markets are questioning: are you willing to fund the buildout before you can see the return?

With Microsoft we are, and the reason is the revenue attach. 

Microsoft is not spending into a hope. It is spending into an installed base of enterprise customers it already bills every month, with Copilot and Azure AI services sold into seats it already owns.

That is a fundamentally different risk profile from spending to build an audience you still have to monetize.

Which brings us to Meta.

Meta appears to be exiled and sleeping on the couch these days.

The advertising engine is excellent: high margin, structurally growing, and still taking share.

The problem is capital intensity.

Meta's capex has ramped materially, to over $100B expected in 2026.

That elevated spend means Meta is burning cash today for a possible return in the future.

Markets do not like sitting through those periods of uncertainty, and that is why the stock has continued to underperform despite a healthy core.

Hedge funds took their chips off Meta. 

Altimeter cut roughly a third of its position. Coatue trimmed as well.

If Meta signs a cloud deal with a frontier lab, the stock will have a pop. But, the main issue for Meta is that other assets are also attractively priced and they don’t have a ‘show me’ requirement.

Having talked about Meta and Microsoft, it’s only fair to talk about Google now. 

Google saw broad institutional buying. It is now Berkshire's second-largest position, totaling about 15% across both share classes.

Renaissance Technologies also added aggressively, tripling their position from last quarter.

Paul Tudor Jones also added here.

What do we like here?

Vistra (VST) - we talked about it in our last newsletter. 

Vistra is Peter Theil’s second largest position. Druckenmiller owns Vistra too.

Earnings are expected to grow double-digits in 2026 and beyond. 

Forward P/E sits at 14.5x, having compressed from its 3Y high of 33.6x on Sep-25.

A PEG of 0.39x on a power generator is an unusual thing to see.

Vistra is also returning capital through buybacks on top of the dividend, which most regulated utilities cannot do.

The stock is poised to benefit from the buildout.

As we noted recently, quite a few of the IPPs are cheap: Talen Energy, NRG, and Constellation.

It has nothing to do with ‘sports as an asset class’ as The Economist suggests.

It’s a powerful tax shield.

My guess is he is preparing to offset a boatload of carried interest income.

If you own a sports team, done correctly, you can get a deduction against income.

The goal in acquiring a sports team is to set up a ‘non-passive’ deduction.

(A passive deduction can only offset passive income.)

Active deduction follows because of ‘material participation’ under Rule 469.

It gets better.

The amount you can deduct yearly is based on a 15-year depreciation schedule.

That’s faster than the 27.5 year deduction found in residential real estate investments.

So, why now?

Kushner is about to see his W2 carried interest explode from his SpaceX, Stripe, and OpenAI investments.

That’s it.

Is it a free lunch?

No. There is ‘depreciation recapture’ at sale. Still, you are compounding on the deferred tax bill.

I wouldn’t be surprised if he is waiving his salary - typical for sports team owners.

What happens when you combine Investing, Tax, and Estate coherently?

That’s called Wealth Architecture.

Warren Buffett is the master of linking investing and tax.

One of his favorites is depreciating goodwill expense.

That’s also over a 15 year schedule.

When you acquire a company that has a quality brand, that will show up as ‘goodwill’.

See’s Candies, Dairy Queen, NetJets etc - they all have goodwill amortization.

This tactic is actually mentioned in one of his biographies.

Mark Cuban handled the Mavericks very well. He also grew the equity value at the same time. Net net, he transformed high income tax into lower taxed capital gains.

That’s a trifecta.

Warren Buffett was arguably the first investor to take advantage of "float".

Insurance premiums paid generate cash inflows that are not taxable.

And, those premiums can be used as a source of funding to buy productive assets.

(Rental income, by contrast, creates cash inflow but is fully taxable as income.)

What Kushner is doing is something best described as 'Tax Float.'

When you acquire a sports franchise, a big chunk of the purchase price can be amortized over 15 years.

Here's the key idea:

You own an asset that is appreciating economically while it is depreciating for tax purposes.

Those deductions reduce taxable income today.

If you eventually sell, some of those deductions may effectively come back through depreciation recapture.

But that could be 10, 20, or 30 years later or never (if you pass to your kids).

In the meantime, you have the use of money that otherwise would have gone to the government.

That's Tax Float.

And like insurance float, its value comes from duration.

The longer you can defer the liability - and the higher the return you earn on the capital in the meantime - the more valuable the float becomes.

The US economy continues to be on a solid footing. 

Let’s start with the labor market. 

Initial jobless claims edged up to 209,000 in the week ended August 7, but stayed close to their historical lows.

The four-week claims average is now at its lowest level since October 2022.

Continuing claims also came in lower. 

The labor market is supported by a growing economy.

The Weekly Economic Index rose 2.7% YoY last week, consistent with roughly 3% real GDP growth.

The Atlanta Fed's GDPNow model currently projects 5.8% real GDP growth in Q3.

Pause, and read that again.

The United States is growing like China. China is growing like Europe. Europe is growing like… well, Europe isn’t growing much.

The composition matters more here than the headline number.

Fixed capital spending is doing the heavy lifting, running at 9.7% YoY growth. That’s datacenter investment.

Corporations are investing in new machinery and equipment at the highest pace over the last year. 

Corporate spending is a signal on the broader economy. 

Companies only increase their investments, when they have high clarity on strong future demand. 

Corporates have strong earnings with all time high margins, and they continue to see a future pathway for higher profitability.

When margins are strong and companies are investing in capacity, payrolls are the last thing management touches.

We have seen it across history (see chart below), and this is the same tape.

This week, I came across an article on where the new demand for labor is coming from, and it’s the exact opposite of what AI doomists at Citrini had expected. 

Firms in the top third of AI spending increased headcount by 10.2% on average in the two years after adoption.

What about firms that didn't spend on AI, and might actually need more people?

No change at all.

Look at the entry-level panel.

Entry-level headcount at high-adoption firms rose 12%, more than the all-jobs figure.

Entry level is precisely where the displacement fear is loudest – it is the cohort every op-ed says AI eliminates first.

But, we are seeing the opposite.

So why does hiring go up?

Because the mental model most doomists carry is wrong.

People assume AI replaces a worker. The unit of analysis is the task, so implementing AI means less labor required for the task, and the conclusion is subtraction.

But, the inverse of it happens. 

AI raises the output of the worker sitting next to it. A more productive employee is a more valuable one to hire.

When output per person goes up, the return on adding a person goes up with it.

Companies can hire more, produce more at better margins, and increase their earnings. 

That is why the firms leaning hardest into AI are the ones expanding payrolls, and the firms sitting it out are flat.

We keep seeing it firsthand and in our conversations with operators. 

The backlog projects are getting done because new AI-equipped workers are more productive.

I did an FSD stream earlier this year on how “AI is creating more work, not less”.

Dario reportedly said that Anthropic might be the only company in the world. 

Dario's vision is one company + governments which is about as dystopian a picture one can paint.

Maybe his thought process is ‘We will steamroll industries like Amazon did.‘

That’s the wrong conclusion. 

Amazon opened up capabilities it built for itself and turned them into markets for everyone else. 

AWS is another example. Third party shopping and fulfillment is another.

What happened when Amazon, Walmart, Home Depot, and Lowe’s disrupted the local retailer who was cost inefficient?

The labor and land was repurposed to service industries, via the invisible hand (eg, incentives and price signals) to more productive uses. 

Over-priced retail stores were re-purposed into wellness spas, daycares, restaurants, cafe, flower shop, or peptide clinic. 

The point is productivity growth expands the pie.

Ronald Coase’s ‘The Nature of the Firm’ predicts smaller, more nimble firms proliferating, and incumbent profit pools that are at greater risk. 

What Anthropic has achieved is truly impressive. 

But, it's comments like this that lower American public confidence in AI.

84% of Chinese are optimistic about AI, 38% of Americans feel that way.

Dario needs to stop with the hyperbole.

We flagged Collegium last week as dislocated value. It is part of the discounted health-care basket that markets seem to have forgotten about despite growth and solid fundamentals.

Their ADHD drug- Jornay PM- is the only ADHD stimulant dosed in the evening, with patent runways into late 2030s.

The market took notice this week. 

The stock is up roughly 11% in the last two days.

The catalyst was management buying its own stock in an accelerated program – a clear expression of ‘dislocated value’.

On Thursday, Collegium entered a $50 million accelerated share repurchase agreement with Jefferies. The initial delivery was 1.5M shares.

Collegium has about 32.5 million shares outstanding. That single tranche retires roughly 5% of the company in one transaction.

It is part of a $150 million authorization, with $100 million still remaining once this settles.

Against a market cap under $1 billion, the full program is about 16% of the shares outstanding.

A management team only commits a sixth of its own market cap to buybacks when it knows the stock is trading at a discount. 

Collegium trades at 4.0x forward earnings.

The free cash flow yield is 36%.

Read that second number again.

At a 36% FCF yield, this business generates enough cash to retire its entire equity base in under three years.

You are not being asked to underwrite a turnaround, a pipeline, or a story.

You are being handed a company whose annual cash generation is more than a third of its market value.

And the fundamentals are not deteriorating. They are improving.

The chart is the whole argument.

EPS has ground higher for years and now sits around $3.

The share price went the other way, and fell hard into this month.

Earnings up. Price down.

That is what dislocation looks like when you can see it.

Price and fundamentals have detached, and only one of the two is anchored to anything real.

We continue to hold Collegium.

We bought Western Digital this week.

The buy idea came from the mean reversion strategy in the Lumida Invest app.

The mean reversion strategy has outperformed SPY over all timeframes. (But, it is higher turnover and tax efficient and best for non-taxable accounts.)

It added Western Digital on Aug 10th, as the stock began recovering from a drawdown of roughly 47% off its June peak.

Since our addition, WDC is up about 15%.

This is the edge AI in financial services can produce. 

We have multiple other strategies in our app. We are adding a feature where Lumida identifies which strategies we believe are the best to use given the market regime.

You can download the app here.

Now, let’s talk about why WDC is an interesting investment.

Western Digital reported its latest quarter on August 5. Revenue was $3.75 billion, up 44% year over year, ahead of management’s guidance.

Gross margin came in at 54.4%, an expansion of 1,310 bps YoY.

Operating income was $1.66 billion, up 126%, at a 44.2% margin.

EPS was $3.56, more than double last year.

Free cash flow was $1.3 billion, a 34% free cash flow margin.

If we look in the weeds, the drivers are working as well.

Price per terabyte rose high teens percent.

Cost per terabyte fell about 8%.

That spread drove the gross margins.

Future revenue visibility also came in strong.

Western Digital now has long-term agreements with hyperscalers running through 2029, with discussions underway for 2030 and 2031.

Those LTAs cover the majority of the cloud business, with pricing escalators built in.

Demand is also broadening beyond the hyperscalers.

Neoclouds, sovereigns, frontier AI labs, and autonomous vehicle companies are all showing up as incremental buyers.

Enterprise OEMs are shifting back toward hybrid storage from all-flash, which is a quiet tailwind markets haven’t realized yet. 

WDC also has higher capital returns. $3.1 billion was returned to shareholders in 2026. 

The drawdown in WDC wasn’t fundamental-driven. It came down with unwind of semis trade.

And, that drawdown has made it a compelling opportunity. 

P/E NTM is now at 25.3x, about 50% below its peak.

With mean reversion factor in-play, this looks like an over-weight name. 

People complain about AI slop. 

But, how many conversations with humans are a kind of Human AI slop?

Brainstorming session? AI slop. 

What to order for dinner? AI slop. 

What to name your kids? AI slop. 

Your college essay? AI slop. 

What gift to give for that occasion? AI slop. 

Jerry Springer drama? AI slop. 

Danielle Steel novels? AI slop. 

TMZ? AI slop. 

AI slop is actually a coming of age milestone. 

‘And he created AI in his own image’

Aswath Damodaran argues that companies like Meta should separate their AI investments into standalone divisions, giving investors clearer visibility into spending, losses, and the uncertain path to returns.

View Curation

Travis Kalanick explains how his “food computer” is only the starting point for a broader vision of treating the physical world like software—digitizing manufacturing, real estate, logistics, and other atom-based industries.

View Curation

Follow Lumida Memes for more.

Not Subscribed Yet? Don’t miss out on future insights—subscribe to the newsletter now!

For real-time updates, follow us on: 

X | Telegram | Youtube | TikTok | News | Ram’s X | Lumida Health | Lumida Tax | Lumida Memes

As Featured In

Disclaimer: Lumida Wealth Management LLC (‘Lumida”) is located in New York, NY, and is an SEC registered investment adviser. SEC registration does not constitute an endorsement of the firm by the Commission nor does it indicate that the adviser has attained a particular level of skill or ability. Lumida only transact business in those states in which it is registered, or qualifies for an exemption or exclusion from registration requirements. Any direct communication by Lumida with a prospective client will be conducted by a representative that is either registered or qualifies for an exemption or exclusion from registration in the state where the prospective client resides.

The information in this material has been obtained from sources believed to be reliable. While all reasonable care has been taken to ensure that the facts stated in this material are accurate and that the forecasts, opinions and expectations contained herein are fair and reasonable, Lumida, Inc. and Lumida Wealth Management LLC (collectively Lumida) make no representations or warranties whatsoever the completeness or accuracy of the material provided, except with respect to any disclosures relative to Lumida. Accordingly, no reliance should be placed on the accuracy, fairness or completeness of the information contained in this material. Any data discrepancies in this material could be the result of different calculations and/or adjustments. Lumida accepts no liability whatsoever for any loss arising from any use of this material or its contents, and neither Lumida nor any of its respective directors, officers or employees, shall be in any way responsible for the contents hereof, apart from the liabilities and responsibilities that may be imposed on them by the relevant regulatory authority in the jurisdiction in question, or the regulatory regime thereunder. Opinions,forecasts or projections contained in this material represent Lumida’s current opinions or judgment as of the day of the material only and are therefore subject to change without notice. Periodic updates may be provided on companies/industries based on company-specific developments or announcements, market conditions or any other publicly available information. There can be no assurance that future results or events will be consistent with any such opinions, forecasts or projections, which represent only one possible outcome. Furthermore, such opinions, forecasts or projections are subject to certain risks, uncertainties and assumptions that have not been verified, and future actual results or events could differ materially. The value of, or income from, any investments referred to in this material may fluctuate and/or be affected by changes in exchange rates. All pricing is indicative as of the close of market for the securities discussed, unless otherwise stated. Past performance is not indicative of future results. Accordingly, investors may receive back less than originally invested. This material is not intended as an offer or solicitation for the purchase or sale of any financial instrument. The opinions and recommendations herein do not take into account individual client circumstances, objectives, or needs and are not intended as recommendations of particular securities, financial instruments or strategies to particular clients. The recipients of this material must make their own independent decisions regarding any securities or financial instruments mentioned herein and should seek advice from such independent financial, legal, tax or other adviser as they deem necessary. Lumida may trade as a principal on the basis of its views and research, and it may also engage in transactions for its own account or for its clients’ accounts in a manner inconsistent with the views taken in this material, and Lumida is under no obligation to ensure that such other communication is brought to the attention of any recipient of this material. Others within Lumida may take views that are inconsistent with those taken in this material. Employees of Lumida not involved in the preparation of this material may have investments in the financial instruments or securities (or derivatives of such financial instruments or securities) mentioned in this material and may trade them in ways different from those discussed in this material. This material is not an advertisement for or marketing of any issuer, its products or services, or its securities in any jurisdiction.