Growth is on Sale: From Models to Meaning
Markets have swung sharply this year. Washington added fresh uncertainty. Geopolitics returned to the forefront. Investors oscillated between conviction in an AI-led expansion and fear that valuations had run too far, too fast.
We’re in the second year of the presidential term. Stock Trader’s Almanac data show this is typically the weakest and most volatile period for equities — and the third year is historically the strongest. Calendars don’t cause earnings, but the pattern explains why turbulence can mask a deeper shift underway.
Thomas Kuhn, in The Structure of Scientific Revolutions, argued that progress advances through paradigm shifts rather than the steady accumulation of normal science. Computing is now undergoing that shift. Software once did only what a programmer specified, line by line. AI systems find patterns in data and generate outputs no one explicitly coded.
The real question isn’t whether a model produces a fluent answer. It’s whether a company can integrate that intelligence into its data and people.
Shyam Sankar, Palantir’s CTO, made the point during the company’s Q2 earnings call. He described Jonah, a submarine-parts worker who still turns wrenches for a living:
He built an AI application that reduced production planning from 30 to 40 days to less than one day. AI alone cannot do that. It takes AI in the hands of the American worker, the tribal knowledge earned through success and failure on the line, and the insights only they have. The models are commodities, but the American worker is not.
Jonah’s advantage wasn’t the model. It was thirteen years on the factory floor. Give AI to a novice, and you get mediocrity sooner. Give it to someone who already knows what a good answer looks like, and you get leverage.
That’s the application layer. Models commoditize. Factory knowledge, hospital protocols, and drilling data remain proprietary. Nvidia’s Jensen Huang calls AI a five-layer cake: energy, chips, infrastructure, models, and applications. The first four layers make intelligence available. The fifth determines who profits from it.
Four layers make intelligence possible. One layer makes it profitable.
The Economy Outrunning the Noise
Geopolitical headlines say one thing, but the data says another.
The July ISM Manufacturing Index came in at 55.6 — the seventh straight month of expansion and the strongest reading since May 2022. All five components, from new orders to employment, moved into expansion territory. That breadth matters. It’s not a single number that flatters a headline.
The Atlanta Fed’s GDPNow model now pegs third-quarter growth at 5.8%. Earnings back it up — S&P 500 profit growth is running above 30%, unusual outside a post-recession rebound.
Skeptics call stocks expensive. Multiples tell a different story. A price-to-earnings (P/E) ratio alone can’t tell you whether a stock is expensive — it ignores how fast earnings are growing. The PEG ratio fixes that: forward P/E divided by expected earnings growth. Economist Ed Yardeni’s tracking shows the S&P 500’s PEG ratio at 0.83, near a 30-year low. Below 1.0, growth is on sale.
Yardeni Research: S&P 500 PEG ratio, near a 30-year low.
Morgan Stanley’s July research explains where that growth is coming from. Its analysts estimate incremental ROIC of 25% to 50% on generative AI investment, ranging from GPU rentals (31%) to proprietary model APIs (46%). Amazon’s second-quarter call put the thesis to the test — AWS grew 37%, AI revenue crossed a $25 billion run rate, and the stock posted its sharpest single-day gain in years. The market had priced in doubt, but the numbers said otherwise.
That has implications for the Fed. New Chair Kevin Warsh has resisted pressure to raise rates even as growth accelerates, arguing that supply-side, productivity-driven growth carries less inflation risk than demand-driven growth. July’s soft inflation reading supports that view: output can rise without prices following.
Where the Tax Code Meets the Portfolio
Tax preparation and tax strategy are distinct disciplines. A return reports what already happened. Strategy decides what happens next — and that work goes best when it’s coordinated with the portfolio, not done in isolation. That coordination is a core part of what BlueBird does.
For real estate owners facing a sale, 1031 exchanges and Delaware Statutory Trusts (DSTs) defer capital gains and facilitate a transition out of active management — both useful for succession and estate planning.
The 2025 tax bill’s permanent restoration of 100% bonus depreciation has opened up innovative strategies worth exploring. Certain leveraged real estate funds can generate front-loaded paper losses that offset other passive income, while oil and gas investments allow drilling costs to be deducted against active income and capital gains alike — each suited to a different situation.
Solo 401(k)s and defined benefit plans do similar work for high-earning business owners.
The fourth quarter is the window. Once the calendar turns, most of these strategies close.
Most require accredited investor status. I’m not a CPA — I work alongside each client’s tax professional to confirm that these strategies fit their full financial picture. If year-end tax planning is on your mind, reply and let’s talk before December.
The Long Walk from Mont Blanc to the Matterhorn
Colorado hikers know the shape of a long trail. The Colorado Trail runs 486 miles from Denver to Durango, and most people complete it in segments over a month. I’ve covered the first 106 in pieces — a weekend here, a weekend there, spread across a summer — trading the Waterton Canyon trailhead for a different river crossing and campsite each time, until the miles added up to Breckenridge.
The Haute Route works on a similar principle, compressed. In one week, I leave for Chamonix with my sister and two of her friends to walk the Haute Route myself — roughly ten days, eleven passes, ending at the Matterhorn. It has no single trail, just a corridor of paths and passes running roughly 110 to 130 miles between Chamonix, France, and Zermatt, Switzerland — Mont Blanc at one end, the Matterhorn at the other. Eleven passes, with most of the route above 6,500 feet, threading beneath the Grand Combin, the Dent Blanche, and the Weisshorn.
Somewhere past Champex, the route skirts Verbier — better known in January for the Bec des Rosses and the Freeride World Tour than for August foot traffic. I’ll take the summer view this time and save the powder for later.
The Haute Route works on a similar principle, compressed. It has no single trail, just a corridor of paths and passes running roughly 110 to 130 miles between Chamonix, France, and Zermatt, Switzerland — Mont Blanc at one end, the Matterhorn at the other. Eleven passes, with most of the route above 6,500 feet, threading beneath the Grand Combin, the Dent Blanche, and the Weisshorn.
Somewhere past Champex, the route skirts Verbier — better known in January for the Bec des Rosses and the Freeride World Tour than for August foot traffic. I’ll take the summer view this time and save the powder for later.
The Haute Route: Chamonix to Zermatt, Mont Blanc to the Matterhorn.
Bella, my eight-pound co-pilot, walked the entire 8-hour round-trip to the summit of Torreys Peak last week under her own power. I picked her up only once — for the photo at the top.
Training day: Torreys Peak, 14,275 feet, with Bella.
In a week, I start walking. I’ll post video updates from the trail on LinkedIn. Here is a post on the first half of the journey
See you on the trail.
Peter Ruh