The New York Times’ (NYT) inescapable Wordle puzzle game turned five this week.

730 million games delivered a confusing verdict: hard mode is easier than easy.

Hard-mode players, boxed in by every revealed letter, statistically solve faster and crash out less than the standard crowd.

The market version of that lesson: those who get creative within constraints keep beating the ones who think optionality is a strategy.

Turns out there are limits to diversification. And it may actually be better to act on a handful of high-conviction ideas than swim in an expanse of decision paralysis.

Which, come to think of it, is our exact pitch for Street Sheet Research…

(Word to Wordle.)

— Brooks & Cas

What: Wall Street is feasting on 20%-plus earnings growth and toasting the AI buildout. But reading the fine print, the party looks catered on credit. Said another way, the cost of the AI spending spree has not hit the income statement yet.

Why: When Nvidia (NVDA) sells a chip, it books revenue fast. Its hyperscaler customers — Meta (META), Microsoft (MSFT), Alphabet (GOOGL), Amazon (AMZN), and Oracle (ORCL) — treat the same purchase as a capital asset and spread the cost over years as depreciation.

What Else: The window has an expiry date nobody can pin down. Those five firms spent an estimated $412B in 2025 and may hit $760B in 2026, against just $211B in depreciation. The estimates for that future hit are all over the place, with Meta's 2028 D&A forecasts scattering 6x wider than its revenue forecasts. Combined free cash flow at the five is set to crater 91% this year, even as net income climbs 25%.

Watch: The forward S&P multiple sits near 22x, rich before depreciation even ramps. The bull case needs revenue to show up before the bills do, and nobody has seen the invoice.

What: The Magnificent Seven spent this decade as the only club worth joining. Then a rocket company crashed the gate.

Why: SpaceX’s (SPCX) run over its first five sessions gave it a $2.45T market cap, making it the sixth-most valuable firm in the US by that metric, worth more than both Tesla (TSLA) and Meta. Awkward, since both are charter Mag 7 names. The label was supposed to mark the market's untouchables, not a leaderboard that a single firm could leapfrog in a week.

What Else: SpaceX is not the only one making the grouping look dated. Microsoft has slumped nearly 20% this year on AI-disruption fears, while Meta and Tesla also trade in the red. Meanwhile Micron (MU) and Broadcom (AVGO) have muscled into the trillion-dollar tier on the same AI trade.

Watch: Wall Street is already floating a replacement, the "MANGOS," though two of its members, Anthropic and OpenAI, have not even listed. Let’s see if any new acronym sticks before the next IPO scrambles it again.

What: Luxury brands spent decades convincing shoppers a four-figure purse signaled they had arrived. The signal is getting jammed. Bain data shows luxury bag sales down nearly 10% from their 2023 peak, an $8B hole in annual spending.

Why: The money hasn’t left the category, just changed counters. Sales of luxury handbags on The RealReal (REAL) are up 20% since 2023, and May searches for vintage bags jumped 131% year over year.

What Else: That shift cuts straight to the bone for brands built on it. Handbags drove 44% of Hermès group sales last year, 65% at Saint Laurent, and 77% at Bottega Veneta. The European luxury names most exposed are down 27% on average since 2024.

Watch: Chanel's Matthieu Blazy makeover, which has shoppers stripping shelves bare of new designs. Luxury names can’t race to the bottom on price, but they can try to drive demand for products that have yet to hit the vintage market.