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

Macro: Bessent Versus the Bond Market; The Economy Is Holding Up

Markets: September Is Doing What September Does; CLEAR: Travelers Will Pay to Get Their Time Back; Venture Capital Is Hard; AGI Is Forrest Gump; Cheaper AI Makes Human Judgment More Valuable

Lumida Curations: Steve Eisman on OpenAI’s $1B Advertising Expense;  Mark Zuckerberg On AI’s Future; Jeff Currie On Investment Opportunity In Refiners 

It has been twenty-five years since September 11, 2001.

Those old enough to remember can still tell you where they were that morning.

This weekend, take a moment to remember those we lost, and honor those who knowingly went into harms way to save the lives of others.

I did a podcast this week with Bob Dewey on Exploring Prosperity.

We discussed how earnings growth is driving markets, how AI is changing business economics, and where we disagree on Bitcoin treasury companies.

Here’s what we cover:

Watch the full podcast here.

We’re bringing together a small group of family offices, investors, founders, and operators in New York on Wednesday, September 16th.

Join us for an evening of good conversation and new connections with the Lumida community.

I’ll also be interviewing Michael Parekh, a longtime technology investor and former Goldman Sachs partner who founded the firm’s internet research franchise.

We will discuss topics like Kimi’s CEO on the run from China authorities, Anthropic’s CEOs urge to ‘slow down’ development, Defense Tech, and other major themes.

Michael has always been a remarkable guest with full of exciting insights. We did a podcast in August as well, titled, “Ladies, Leverage, and Leopold”.

Watch on Spotify and Youtube.

We have about 40 registered, and just a handful of slots remaining.

Register here to join us.

Kimi’s CEO Is On the Run From China Authorities

A lot of headlines this weekend - first Anthropic’s CEO is calling for a ‘slowdown’ to focus on alignment. Others - Sam Altman and Elon - agreed.

Today, I woke up to learn that Kimi’s CEO is on the run from China authorities. Recall, Kimi is a China-based open-weight model. It’s much, much lower cost that US closed models - and it’s gaining market share.

What happened?

Apparently, Kimi was routing the hardest user requests to Anthropic. And, guess what?

China was one of the customers of Kimi. So, sensitive and confidential information may have been shared with Anthropic and consequently the U.S.

It’s great to see some karmic justice here - China is an aggressive IP violator and has used distillation techniques to catch-up to US models.

The best part is that Kimi routed the hard questions upstream. So the U.S. didn't get the easy stuff. It got the queries Beijing couldn't answer on its own.

Figure the U.S. now has a couple months of chat logs that read like: "Draft a press release announcing a reunification operation. Make it sound routine. Also, what's the shipping lead time on 400,000 landing craft?"

This creates a setback for China’s LLM model. If the leader of your company is down for the count, it’s extraordinarily hard to execute.

Incidentally, we note that China stocks - including Tencent - are super cheap. We added a starter position to the name here. When oil prices drop, China should benefit. China is a contrarian idea.

Our thought is - wait for sentiment to turn. Unlike the United States, China doesn’t have the same vociferous competition backed by unlimited venture dollars. The leaders in China will continue to lead - in part due to government control.

If you aren’t spending several hours with Grok Bot, you must start. This will open your eyes to the transformational potential of AI. Grok Bot has eclipsed Claude in agentic workflows.

The race continues - and we’re finally getting highly usable enterprise and consumer products now.

The demand for intelligence self-clears, because intelligence is productive.

One should look thru Anthropic CEO Dario’s comments. When he says ‘We need to slowdown’, the magic decoder ring is saying ‘We need to shift to profitability, we are spending so much money on capex it’s insane.’

That’s all. Will the capex spending slow? No. The Game Theory prevents that. SpaceEx is on pace to clear $100 Bn in revenue soon from renting Nvidia chips. SpaceX will keep buying those chips. The hyperscalars will keep buying.

The LLMs can generate greater revenue to keep up with lofty Wall Street expectations by buying chips.

I guess I am officially shilling Elon Musk products now.

AI HACK

If you are using AI, here’s a trick I stumbled upon. 

Invite your AIs to meet in a Slack channel and have them coordinate: GPT, Grok Bot, and Claude. 

Multiple models together are better than one. This is called ‘Ensemble’ modelling, and it’s well known that coordinating models builds better outcomes.

If you’re coding, have one model (say, Claude) do the coding. Have another model (say, GPT) do the evaluation. Set a clear evaluation function and have that model provide direction to Claude.

Grok can get in there and provide a more holistic perspective and notify you when to get involved.

I posted this concept on X last week here. Meta’s AI chief is using a similar technique to create ‘loops’ that out-perform 100+ engineers. I believe it.

One wonders if he was referring to Meta’s Muse. Did they build this in a week in response to Instinct?

Tax Mitigation

Really appreciate folks helping to test our Tax Mitigation App.

We are going to launch this into the wild soon. If you want free tax mitigation detection opportunities, I’d give it a go. There’s no fee for the service currently.

This is an agentic model that shows how AI can perform interactive discovery to identify tax mitigation.

The service is still in beta. I’d love for you to try it, and share your feedback.

Try here.

Note: Lumida has a wide range of strategies designed to mitigate capital gains tax, estate tax, and income tax. Ping [email protected] if you’d like to learn more.

Separately, Lumida is looking at a potential investment in a Ukranian Drone technology company.

Our investment in Shield AI earlier this year is reportedly marked up from $10 Bn to low $20 to $23 Bn according to whispers of their upcoming round. I believe this deal has much higher potential due to the valuation. You can sign up at Lumida Deals to receive our private deals communications timely.

Past performance is not indicative of future results.

We are sharing the early preview here, because we are simply not offered much time to share these deals. Start doing your homework now. For Accredited and Qualified Purchasers only under Reg 506(C). Not an offer to buy or sell securities.

The inflation report made Bessent’s goal harder this week.

Headline PPI rose 0.4% MoM, its largest increase since May. Energy prices jumped 4.2%, pushing transportation and warehousing costs higher.

But, energy wasn’t the only driver of this increase. Excluding food and energy, producer prices still rose 0.4% YoY. 

Markets responded to the higher PPI number by raising September rate-hike odds to 87%, from 60% on Wednesday.

The bond market is telling Fed to do more before it gets comfortable with inflation.

Fed Rate hike probabilities are near 90%. The good news? Most rate hike risk is now fully discounted.

So, you should not be worrying about rates here. The market has already worried about rates.

The two-year yield jumped 25 bps over the week. The 10-year reached 4.96%, putting it within touching distance of 5%.

Shorter-term yields rose faster, flattening the curve. This shows investors are expecting tighter policy now to keep inflation contained further out.

Bessent, meanwhile, is trying to push the long end down.

Treasury announced an expansion of long-end buybacks in August, followed by a $6 billion operation in the 10- to 20-year sector.

But, this didn’t stop the 10Y from rising. Markets still kept selling bonds.

My view has not changed: the intervention was unnecessary. Let markets clear.

Druckenmiller was right.

And $6 billion is small against a Treasury market exceeding $31 trillion.

Buybacks cannot resolve the underlying questions around inflation, and the return investors require to hold long-duration debt.

There’s also another issue. Who in their right mind is holding government paper when you can own assets that are growing earnings 30% that are reasonably priced.

Hard for Bessent to sell bonds when the opportunity cost is that high.

Overall, this is a mini-echo of the summer of 2023.

Rates rose, semiconductors pulled back, and the market spent months worrying about how high yields would go.

The pressure finally broke in late October, around earnings season. Markets bottomed on October 26th just as earnings season started.

The 10-year could test 5%, perhaps overshoot it, before settling.

Crowded bullish optimism amongst advisors remains the strongest near-term bear case, and it may take more time to reset.

But, strong businesses are becoming available at better prices while earnings growth remains healthy.

Relatedly - Trump will de-escalate Iran - that means oil prices are going down. So rate and energy sensitive names are on sale. Take a look at Carnival Cruise Lines and Delta / United for example.

After the mid-terms, however, you can make the case that Trump might seek to re-escalate especially if the Dems take the house. He is solving for mid-terms right now - after that is anybody’s guess

The labor market continues to show strength.

Initial jobless claims have been running below the last 3Y average since April. Continuing claims have also declined and sit below their last 3Y level.

That is encouraging. We are seeing limited layoffs alongside a shrinking number of people collecting ongoing benefits.

This shows the labor market is strong, which is a positive for corporate demand and earnings.

The economic growth backdrop helps explain the strength in the labor data.

Consumer spending and the AI-led capital investment cycle continue to support demand.

In our conversations with operators, AI is helping businesses complete projects and expand what their teams can produce.

Atlanta’s Fed GDP model is now projecting 4.6% YoY GDP growth for Q3, driven by 2.7% higher consumer spending, and 13.3% higher fixed private investments. 

That creates tension for the Fed.

The AI buildout requires power, equipment, construction, and labor today. Those demands can keep prices firm.

However, over time, the resulting productivity should help businesses increase output at lower cost.

The timing matters.

All major market indexes ended down this week. The S&P 500 fell 1.2%, and Nasdaq was down 0.4%.

September is historically the weakest month of the year, with an especially poor record in midterm years.

And the harder part of the month may still be ahead.

Bespoke’s charts show that September’s weakness tends to arrive in the second half. The first two weeks can look relatively calm before selling picks up into the month-end.

Positioning helps explain why the market remains vulnerable.

I mentioned BofA’s Bull & Bear Indicator in our newsletter 3 weeks back, when it reached 9.7 out of 10.

The reading has eased to 9.5, but still stays elevated, and in deep optimism. This shows markets haven’t reset yet. 

Look at the allocation charts, too.

BofA’s private clients have 66% of their portfolios in equities, close to all-time highs. Cash is just 9%, below its historical average of roughly 12%.

This shows positioning stays full.

When investors already own plenty of equities, it takes more to bring in the next buyer.

A rate scare can prompt selling before anyone has changed their earnings outlook.

Overall, however, since earnings growth is strong — dips we expect will be bought by institutional investors.

The earnings picture gives us a reason to be optimistic about a year-end rally.

And, Anthropic’s S-1 should drop soon - giving investors confidence in the margin story (or pause).

The chart below shows AI infrastructure earnings grew 54% YoY in Q2, accelerating from 48% in Q1. 

The rest of the S&P 500, excluding energy and AI infrastructure, also grew earnings 14%, which is also higher than the previous quarter.

AI infrastructure is doing the heavy lifting, but earnings are growing outside it as well.

This is why we stay optimistic.

Earnings seasons should give investors something more substantial to focus on than the next move in Treasury yields.

Treasuries are in the throes of capitulation now - we expect it resolves this week.

Here are some ideas we like and have bought recently.

We recently bought CLEAR Secure (Ticker: YOU). The name pulled back to just below the 200 Day Moving average.

This business has a monopoly on convenient airport security access. It’s a SaaS model that everyone renews year-in, year-out, because time is precious.

The name fits our travel and leisure theme through a simple proposition: boomers love to spend money on leisure. Business travelers want convenience.

Airports have plenty of friction. Security lines, unfamiliar terminals, and uncertainty about how early to arrive. 

Clear Secure is the dominant market leader in paid airport identity verification, with 62 airports and 8.3 million active CLEAR+ members.

That airport footprint, established partnerships, and trusted brand would take a competitor years to replicate.

The stock sold off with software.

But the demand drivers here are travel, convenience, and increasingly, secure identity.

The stock shouldn’t rise and fall with software just because it’s classified as Software.

The company increased its standard annual membership price from $209 to $219 in July.

CFO Jennifer Hsu said:

“We did not see any impact to retention.”

No kidding. We’re all stuck with Clear, aren’t we? They could charge the price of an iPhone and we’d pay it.

The results are showing it.

Revenue grew 26.6% to $277.8 million in Q2.

Active CLEAR+ membership increased 15.2%, while adjusted EBITDA margins expanded roughly nine percentage points to 36.4%.

There is another growth opportunity outside airports.

CLEAR1 brings identity verification into healthcare, government, and the workplace. AI makes impersonation easier, which increases the value of reliably proving who someone is.

New customer signings and the CLEAR1 pipeline each grew more than 50% sequentially. It is still a developing business, but it gives Clear another use for the identity platform it has already built.

Now look at the following chart.

The blue line represents EPS. It has continued stepping higher, reaching approximately $2.37. The black line is the share price, which has fallen sharply to around $42.

The business has sold off while fundamentals kept improving.

The valuation has become more interesting as a result.

Forward P/E is 18.6x, near the low end of its three-year range. The trailing free cash flow yield is approximately 11.7%, near the high end of its three-year range.

Meanwhile, revenue is projected to grow 23% in 2026, with 37% EPS growth.

The bear case? We don’t like the stock-based compensation.

SBC reached $43.2 million in the last twelve months, representing approximately 29% of $148 million in operating cash flow. Clear built a business on letting people skip the line. Shareholders don't get to skip this one.

That is high.

SBC is added back in operating cash flow, but issuing equity still has a cost for shareholders.

Even with that concern, we see a market leader adding members, improving its service, and expanding margins at a substantially lower share price.

What Other Opportunities Do We See?

Nu Holdings (NU) and AppLovin (APP).

Nu has built a large digital banking franchise in Latin America. Its push into the U.S. gives it another avenue for growth, bringing its model into a much larger market.

The recent technical weakness gives us an opportunity to buy into that expansion at a better price. 

The stock has pulled back while earnings expectations have continued to rise.

Revenue is projected to grow 45% this fiscal year, with EPS increasing 41%.

Return on equity is approximately 32%.

The stock trades at 14.8x forward earnings, near the bottom of its five-year valuation range. That is an attractive price for this level of growth and profitability.

Look at the earnings against the price chart. Since the start of 2026, forward EPS estimates have risen roughly 15%, while the stock remains more than 20% below its early-year peak.

AppLovin is a name we have discussed before. I love this name. I love AppLovin.

We think it has likely consolidated and is showing early signs of strength.

The recent stabilization is encouraging, and gives a good entry point.

The fundamental business is solid.

Revenue is projected to grow 48% this fiscal year, with EPS increasing 57%.

AppLovin is a direct way to invest in “AI improving advertising returns”. Its technology helps advertisers turn spending into paying customers. 

When those campaigns become more profitable, advertisers have a reason to increase their budgets.

The stock has fallen more than 50% from its peak, even as forward earnings estimates have moved substantially higher.

At 17.2x forward earnings and a free cash flow yield of approximately 4.2%, the valuation gives us a more reasonable entry into that growth.

Another name we are looking at is Morningstar (MORN), though we are still waiting for a better entry.

Morningstar has recurring relationships, and a reliable revenue base.

Customers have built entire processes around Morningstar’s data and software, which makes Morningstar's products difficult and costly to replace.

Revenue is projected to grow 8% this fiscal year, while EPS is expected to increase 25% – that’s operating leverage a data and software business gets. 

Return on equity is approximately 32%.

The price has moved in the opposite direction to fundamentals.

Since the start of 2025, forward EPS estimates have risen roughly 60%, while the shares have fallen approximately 40%.

Morningstar now trades at 14.8x forward earnings, near its five-year valuation lows. 

Its free cash flow yield is approximately 6.9%, near all-time highs. 

The trailing buyback yield is also substantial at nearly 14%.

A recurring-revenue business with growing earnings and significant capital returns deserves attention at this price.

I saw news that the Google A team left and is doing a second round at a whopping $50 Bn valuation. No revenue. It’s bananas.

Some history.

Remember when Plaid raised at $12 Bn in 2021? And, that’s a real business with stickiness. Now, it’s worth $6 Bn.

General Consumer AI is the new theme. Instinct and Granola are its examples.

Whatsapp was bought for $16 Bn. 

It had half a billion users then and was the dominant market leader. 

Now, Whatsapp has 3.3 Bn users. 

In retrospect, a fantastic acquisition…and plenty of scope for monetization. 

Instinct is valued at $2.5 Bn. 

Number of users: small, private beta. but viral. The number could be less than 100K users. 

Instinct will need to ‘grow’ into the valuation. ‘Grow’ is codeword for ‘expensive’.

Now, Meta’s Muse, and Google Spark, and Apple are going after Instinct.

Personal Assistant AI is a top 3 priority of these firms - so best avoid getting run over. Better to find under-served niches.

Niches nowadays can involve markets that have hundreds of billions in TAM.

Being a VC right now is high variance and intense competition. 

And, the SpaceX deal is re-liquefying venture coffers by $15 to $20 Bn alone…

We are seeing mid-cycle VC re-liquification and DPI. So, valuations will go even higher!

That’s going to keep the AI and cloud spend going, too. 

It’s pro-cyclical.

(RIP all pre-ChatGPT vintage startups)

These so-called AGI models are painfully, tortuously slow.

Been using Astra all weekend and it's 'pretty good' for development. 

It's not a 'wow moment' where you can see autonomous execution. 

Modern coding is devs taking a smoking break, playing foosball, or going for a walk waiting for the LLMs to come back.

Astra also needs continuous 'prodding'. It does not self-execute.

And, the AIs lack some common sense. 

Grok bot is impressive. 

I did token max that and consumed all my monthly credits on the Ultra plan in 4 days.

The bottom line: the world needs more memory, more semis.

AGI isn't taking your job.

(Astra can crack Captchas better than I can. I swear these buses are in a state of quantum superposition sometimes. They are there or not there or both.)

Nate B. Jones has a insightful article on Mary Meeker’s 340-slide AI presentation. It covers her 2025 report, but the questions it raises are worth revisiting.

Nate highlights how quickly AI economics have changed. In the last two years, the cost of using AI fell roughly 99%.

Declining AI costs changes what’s possible with AI.

A small business can attempt projects it could never justify hiring a team for.

We are seeing this at Lumida. 

I recently built a recruiting bot that screens resumes, prioritizes candidates, and alerts our team when it identifies exceptional talent.

It has improved our screening and reduced time spent on unsuitable interviews. In short, we disrupted WorkDay - in-house anyway.

Look at the two charts below. Frontier models are competing against one another, while overall usage continues to grow at a rapid pace, which means its a bigger market with a smaller share for the incumbents. 

That’s also quite the slide OpenAI should have in its investor presentation if it plans to sell its stocks after losing to Anthropic despite its massive lead.

See the third chart.

Nate also talks about the difficulty of integrating AI into real workflows. Access to a capable model does not automatically translate into a more productive company. The surrounding processes and context matter.

We agree.

High agency buiness owners and founders are living in the Golden Era.

You can read Nate’s AI blog and Meeker analysis here. 

You can now view Lumida Curations at the Lumida Invest App. Curated insights delivered on time. Download the app here.

Steve Eisman argues that OpenAI’s $1 billion advertising run rate falls short of its annual ambitions and does little to offset its substantial losses.

View Curation

Mark Zuckerberg argues that AI should empower people, enable invention, and remain broadly accessible, with safety built on checks and balances rather than control by a handful of labs.

View Curation

Jeff Currie argues that limited new capacity and unusually high refining margins are generating substantial cash flows, yet investors continue to favor AI growth stories.

View Curation

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.