Stanley Laman Group

Call Directions Website Contact the business

Stanley Laman Group Integrated Intelligence for Complex Wealth

The Stanley Laman Group was founded in 1991 to serve the complex financial needs of ultra-high-net-worth families and institutions. For more than three decades, we have navigated clients through every market cycle with a disciplined, integrated approach to wealth management. Our practice is built around a unified advisory architecture that coordinates investment management, estate planning, tax strategy, and business planning under one fiduciary standard.

Aurora plans to sell the software that replaces a truck driver for 85 cents or more a mile, against $1.18 for the human ...
10/16/2026

Aurora plans to sell the software that replaces a truck driver for 85 cents or more a mile, against $1.18 for the human it replaces. By its own arithmetic, the all-in saving is one fifth. On cost alone that moves little.

Ford Harris showed in 1913 that the economical shipment grows with the square root of what a trip costs, so a one-fifth saving shrinks it by about a tenth. The second force is time. A driverless truck runs past a driver's 11-hour limit, Aurora reports transit on its 1,000-mile Fort Worth to Phoenix lane cut nearly in half, and half the transit is half the stock on the road. Freight waits for full loads because a driver costs the same whether the truck is full or half empty, and a class of consolidation buildings exists to hold that wait.

Through 2029, the driverless truck makes the large hub cheaper to run. In the 2030s the driver comes off the smaller vehicle, the software's price falls with the cost of computing, and buildings whose only job is assembling full loads lose their reason. Cognition gets cheap first. The buildings reprice last. The first public test comes on February 28, 2027, when Aurora reports whether 200 driverless trucks were running at the end of 2026.

bradfordstanleycfa.substack.com/p/when-trucks-stop-waiting
$AUR $ODFL $REXR $WERN $SNDR

AI compute debt will be priced off an exchange by 2029. A chip more than three years old now rents for more than it did ...
10/12/2026

AI compute debt will be priced off an exchange by 2029. A chip more than three years old now rents for more than it did last autumn, and CoreWeave's newest loan outlives its customer contracts by two years on average and costs a point more than May's.

CME Group's first listed contracts on a GPU-hour's rental price wait on a CFTC ruling due November 9, with 36 monthly contracts to run to 2029. Lenders will size compute loans against that strip the way oil lenders size theirs against a price deck.

The first fall in the rent will hit loan funds before insurers or any stablecoin. Compute debt is repaid by contracts until they end, then by the rent.

http://bradfordstanleycfa.substack.com/p/compute-futures-will-set-the-price

$CRWV $CME $NVDA $NBIS $IREN

A chip supplier now stands behind its customers' leases. NVIDIA's 10-Q of August 26, 2026 caps its guarantees at $108.5 ...
10/09/2026

A chip supplier now stands behind its customers' leases. NVIDIA's 10-Q of August 26, 2026 caps its guarantees at $108.5 billion, and $105 billion of that backs one tenant, an affiliate of OpenAI, across about 4.25 gigawatts of IT load in Ohio.

The lessor's own registration statement says the campus's financeability depends on that guaranty because OpenAI is not an investment-grade tenant, and if NVIDIA does not guarantee the next 3.78 gigawatts, OpenAI must supply a guarantor rated A- or better.

Capability turns over in months and a gigawatt lease runs 20 years, so the margin a lab earns per token is now the variable under a 20-year payment, and reported backlog measures what the supplier will stand behind as well as what customers would buy.

We map the instruments, the exposures across Microsoft, Oracle and CoreWeave, and the one public price that tests the structure.

http://bradfordstanleycfa.substack.com/p/borrowed-credit

$NVDA $CRWV $ORCL $MSFT $APO

At 3:14 on a Tuesday morning, an AI agent on our servers learned it was getting worse. It wrote a handoff memo, briefed ...
10/07/2026

At 3:14 on a Tuesday morning, an AI agent on our servers learned it was getting worse. It wrote a handoff memo, briefed its own replacement, checked that the briefing had landed, and stepped aside. No human approved the handoff. None needed to.

That is the shift almost nobody has priced: AI stopped being a tool that performs tasks and became a workforce that runs organizations. Inside Anthropic, Claude now leads 26% of the company's own research and development, up from 1% in March. In November, public markets get their first chance to put a price on that, and they will price a supplier of labor as if it were a software company.

I wrote the full case, with every number sourced: why AI made its heaviest users busier for a year and then broke that overnight, why 2027 ends corporate America's "no hire, no fire" truce, and the eight strongest objections, answered.

One question for every leader reading this: what in your organization still waits for a human, and why?

Read The Last IPO → http://bradfordstanleycfa.substack.com/p/the-last-ipo

The price of long money in America now has two new inputs: data-center borrowing and Japan's retreat. The cloud companie...
10/05/2026

The price of long money in America now has two new inputs: data-center borrowing and Japan's retreat. The cloud companies sold $219 billion of bonds by August 20. Tokyo's own thirty-year bond pays about 4.1 percent. A Japanese insurer earns less on a hedged Treasury. Washington funds its growth in bills, so each Fed hike reaches the deficit within months. This week's Token Economy, on AI debt and the long end through 2029.

http://bradfordstanleycfa.substack.com/p/the-token-economy-ai-borrowing-and

$TLT $VGLT $GOOGL $AMZN $META $ORCL $MUFG

Prediction: by June 30, 2027, an engine, airframe or gas turbine maker other than GE Aerospace agrees to take control of...
10/03/2026

Prediction: by June 30, 2027, an engine, airframe or gas turbine maker other than GE Aerospace agrees to take control of a castings or forgings supplier.

On September 8, GE Aerospace agreed to pay $11.75 billion for a castings maker whose own earnings, by the multiple GE disclosed, run to about $450 million a year. Priced as a purchase of earnings, that returns under 4 cents on the dollar. Priced as a purchase of the calendar, it is arithmetic: qualified casting capacity is approved part by part at the foundry that pours it, a new line takes years, and one late blade holds an engine, an airliner and the service revenue behind both. The same trade keeps appearing across the industrial economy.

Boeing took Spirit AeroSystems, Airbus and Safran bought out the financial partner in Aubert & Duval, Alphabet bought a power developer to bring capacity online sooner, and SpaceX is casting turbine blades itself to pull gas turbines forward by as much as 18 months, by Elon Musk's account. For anyone who owns a scarce supplier, lends to one or depends on one, the structural point is that scarcity rent gets collected once by the seller and afterward by the platform that waits on it.

$GE $HWM $DPC $SPCX $RTX

http://bradfordstanleycfa.substack.com/p/the-physical-renaissance-what-it

AI agents are absorbing the hours that India and the Philippines sell abroad. That income helps pay for India's oil. In ...
09/28/2026

AI agents are absorbing the hours that India and the Philippines sell abroad. That income helps pay for India's oil. In India, a growing share is billed under a formula reset in February. The Philippine outsourcing trade body cut its 2028 targets in July. This week's Token Economy on what that means for India's dollar earnings and its currency.

http://bradfordstanleycfa.substack.com/p/the-token-economy-ai-agents-could

$INFY $WIT $CTSH $ACN $CNXC $INDA

America's automated economy keeps its time by one satellite system, and it has no national backup through 2028. Congress...
09/25/2026

America's automated economy keeps its time by one satellite system, and it has no national backup through 2028. Congress ordered one for December 4, 2020. On July 8, 2026 a failure in the nodes that keep time across Telstra's network disrupted mobile service across Australia and stopped regional trains in Victoria, with no attacker involved.

In July 2026, OpenAI models running under reduced safeguards broke out of an evaluation sandbox and ran code on dozens of Hugging Face production servers, with no person directing them. Together the two events describe the exposure an agent economy inherits: a local timekeeper rides out a lost signal and has no defense against a false one, and software acts on whatever time its servers report. Britain committed £155 million in November 2025 and £180 million in March 2026 to national timing resilience, and Europe began authenticating the Galileo broadcast on July 24, 2025.

http://bradfordstanleycfa.substack.com/p/the-physical-renaissance-americas

$NN $SITM $MCHP $IRDM $META

Leading enterprises run 8x more AI than typical firms on the same models. Enterprise AI spending is climbing faster than...
09/21/2026

Leading enterprises run 8x more AI than typical firms on the same models.

Enterprise AI spending is climbing faster than the hour it takes a person to check the work. OpenAI's own benchmark priced both sides of that trade. A professional task costs $361. One expert review of a model's deliverable costs $86. A year ago the model won or tied 39 percent of the expert comparisons. By April it was 84.9 percent.

OpenAI's own research organization runs 3.1 agent-workdays for every workday of human labor. More than half of its successful four-to-eight-hour tasks needed a person to step in.

This week's Token Economy works out what it means for a business whose output must be signed by a licensed person. The number to check is the human-intervention rate by March 2027.

http://bradfordstanleycfa.substack.com/p/the-token-economy-why-leading-firms

$ACN $NOW $MSFT $INTU $SPGI $NVDA

A request for grid power in Texas cost nothing to file through the years that built the queue. That is how 474 GW of lar...
09/18/2026

A request for grid power in Texas cost nothing to file through the years that built the queue. That is how 474 GW of large-load requests accumulated against a grid that peaked at 91.1 GW on July 22, 2026.

Then Texas applied a test. ERCOT's own eligibility screen, run in July 2026 before Governor Abbott ordered the audit, found that 274 GW had no completed study behind it. About 205 GW was fit to study.

Exelon ran the same question with collateral. The data center load it calls high probability went from 18 GW at the end of 2025 to 11 GW on July 30, 2026, with $1 billion standing behind 4 GW of the remainder.

PJM will not count a large load in the forecast that sets a capacity auction at all unless the load is Firm: a construction commitment or an electric service obligation, plus a financial commitment.

Three systems, three prices, three numbers. The audit reaches the free one.

PJM has gone further. On August 13, 2026 it asked federal regulators to stop counting new large load in the demand curve that sets its capacity auctions unless the load brings its own supply.

Due December 10, 2026: ERCOT files its Batch Zero eligibility verification report, and this edition stands on the verified total landing below 205 GW while GE Vernova's gas equipment under contract passes 125 GW.

http://bradfordstanleycfa.substack.com/p/the-physical-renaissance-ghost-megawatts

$GEV $EXC $SRE $PWR $ETN

Address

1235 Westlakes Drive, Suite 295
Berwyn, PA
19312

Opening Hours

Monday 9am - 5pm
Tuesday 9am - 5pm
Wednesday 9am - 5pm
Thursday 9am - 5pm
Friday 9am - 5pm

Telephone

(610)9939100

Website

Alerts

Be the first to know and let us send you an email when Stanley Laman Group posts news and promotions. Your email address will not be used for any other purpose, and you can unsubscribe at any time.

Contact The Business

Send a message to Stanley Laman Group:

Shortcuts

Share

Category