AI wants the sale. Where does the profit land?

Tuesday brief · September 29

• Follow the money: Google’s new advertising workflow targets more efficient reach. The useful next question is whether advertisers earn more after all costs. [1]
• A new competitor: Meta is building an enterprise AI business. A confirmed executive hire strengthens execution capacity; it does not establish customer revenue. [2,3]
• Wheel desk: Alphabet, Shopify and Amazon offer different ownership cases. Today’s three visual studies show cash at risk, stale-quote limitations and reasons to wait.

The lead · AI must earn the next advertising dollar

Google introduced unified vertical-video buying in Display & Video 360 on September 28, with Gemini supporting execution and measurement. Google reports that an early Hellmann’s/WPP Media test increased unique reach by 24% and reduced cost per unique user by 25%. These are vendor-reported pilot outcomes; they do not establish higher sales or profit. [1]

Our investment read: advertisers buy access to customers, but shareholders ultimately need profitable transactions. Reaching someone more cheaply is valuable only if that audience converts at an acceptable margin. A campaign can look efficient while attracting low-value orders, increasing returns or merely taking credit for purchases that would have happened anyway.

That distinction connects today’s AI story to the watchlist. Alphabet competes for marketing budgets; Shopify helps merchants complete purchases; Amazon combines commerce with cloud infrastructure. Each can benefit from better software, yet the economic benefit lands in different accounts. Strong adoption can coexist with heavy investment or disappointing cash generation.

The counterargument is straightforward: simpler buying tools may reduce operational friction even before a clean profit uplift is measurable. The next evidence should be customer retention, repeat campaign spending and incremental contribution profit after discounts, fulfillment and computing costs. A successful pilot opens an investigation; it does not settle valuation.

Eddie’s watchlist · where the evidence changes

• SHOP — checkout infrastructure: technical documentation supports delegated Shop Pay payments using tokens. That substantiates the mechanism behind agent shopping, without proving its scale. Track completed orders and merchant economics; conversational product discovery alone is insufficient. [4]
• AMZN — growth needs funding: June-quarter AWS sales grew 37%, while consolidated trailing-year free cash flow was negative $7.6 billion. AI investment is central to that cash burden. The next financial update must connect expansion to funded demand. [7]
• MU — tomorrow’s checkpoint: the September 30 earnings call can sharpen the memory-demand picture. Watch capacity spending and customer commitments together. Micron is outside today’s wheel studies because the scheduled event adds uncertainty. [9]

AI buildout & adoption · two economic tests

Meta enters the enterprise contest. Its September 28 announcement puts Muse and business-agent products into a new Enterprise Platform effort led by CJ Desai. MongoDB separately confirms his departure for Meta. Our interpretation: enterprise selling needs implementation, support and credible controls, not just consumer distribution. Product packaging, customer commitments and renewal evidence will determine whether this becomes a durable revenue stream. [2,3]

Compute has to pay for itself. Alphabet generated $39.1 billion of operating cash in Q2 and bought $44.9 billion of property and equipment. Those are quarterly cash figures, not announced future spending. Our interpretation: improving inference efficiency matters because a useful transaction must cover serving costs as well as attract demand. More activity without sufficient monetization can deepen funding needs. [5]

For AI adoption, the practical bottleneck can move from model intelligence to integration: reliable product data, payment authorization and handling exceptions. Our reading is that transaction completion and customer trust deserve as much attention as model benchmarks. Spending on those systems can delay the profit investors expect.

Macro lens: tomorrow’s inflation release is a risk checkpoint, not a predicted direction for stocks. A hotter reading could pressure rate-sensitive valuations; a softer reading could ease discount rates while also raising questions about demand. For commerce businesses, distinguish easier financing from healthy consumers. Neither outcome makes a weak option quote acceptable.

One comparison · revenue is not equal to profit

Text alternative: Alphabet retained $34.0 of operating profit per $100 of Q2 revenue; Shopify $13.6; Amazon $13.7. Their advertising, payments, retail and cloud mixes differ, so this is not a quality ranking. Accounting margins include depreciation rather than deducting the full equipment purchase when cash is paid. Use it alongside investment needs and valuation, never as a substitute for them.

The valuation ruler: our sales multiples divide each reference breakeven by Q2 revenue per diluted weighted-average share, multiplied by four. This keeps the denominator explicit, but assumes a quarter repeats. It ignores seasonality, changing shares, debt and business mix. A low sales multiple can still accompany weak margins; a high one demands more durable growth. It is a comparison tool, not fair value.

3 Wheel Trade Ideas

WATCH-ONLY / PASS · October 16 puts, 17 calendar days. Stocks: September 28 completed session. Cboe files retrieved together September 29 at 06:48 Toronto retain September 28 last trades; exact bid/ask time, timezone, delay, open-interest date and IV rank/percentile are unconfirmed. U.S. options are closed at 08:00 Toronto. Reprice and revalidate after the 09:30 open; no fill is assumed.

Each study assumes one standard 100-share contract; confirm deliverables with your broker. Cash = strike × 100; credit = prior-session bid × 100; breakeven = strike − bid. Daily Wilder RSI14 and SMA20/50/200 use Yahoo adjusted closes; relative volume compares Monday with the preceding 20 full sessions. None is oversold. Delta is not assignment probability. All amounts are USD. [11,13–18]

After assignment: the wheel does not repair a broken investment thesis. Owning 100 shares creates ongoing stock risk; a later covered call limits upside and may not recover a large decline. Size the cash commitment before comparing credits. The loss panels show one downside scenario, while the separate maximum-loss figure shows the much larger exposure if the business fails.

GOOGL · Alphabet

Watch-only · prove support first · Prior-session study: the $320 put, expiring October 16, 2026. One contract means agreeing to buy 100 shares at that price if assigned. Reference arithmetic; not a current offer.

• Why own the shares: Advertising and cloud support ownership, but rates and marketing budgets matter. Reference breakeven is 8.18× annualized Q2 sales per diluted share. This seasonal proxy is not forward P/E; investment gains distort GAAP earnings. [5]
• Trend and levels: Monday $342.75; RSI 48.9; SMA20/50/200 $341.94/$344.06/$337.95; relative volume 0.72×. Near the long average; weak participation. [13]
• Contract quality: Bid/ask $1.68/$1.87; spread $0.19; delta −0.1449; IV 31.00%; volume 586/OI 10,190. Last trade September 28, 15:46:28, provider clock. [14]

SHOP · Shopify

Pass · wide spread, demanding price · Prior-session study: the $130 put, expiring October 16, 2026. One contract means agreeing to buy 100 shares at that price if assigned. Reference arithmetic; not a current offer.

• Why own the shares: Merchant growth supports the business, but softer consumption can hurt. Q2 revenue grew 34%, with 18% non-GAAP FCF margin. Breakeven is 11.63× annualized Q2 sales per diluted share: demanding, not a bargain simply because the put is lower. [6,8]
• Trend and levels: Monday $144.01; RSI 54.9; SMA20/50/200 $137.33/$138.36/$129.85; relative volume 0.78×. Above all averages, without strong volume confirmation. [15]
• Contract quality: Bid/ask $1.53/$2.18; spread $0.65; delta −0.1782; IV 55.47%; volume 104/OI 1,530. Last trade September 28, 15:52:04, provider clock. [16]

AMZN · Amazon

Watch-only · the short trend is weak · Prior-session study: the $230 put, expiring October 16, 2026. One contract means agreeing to buy 100 shares at that price if assigned. Reference arithmetic; not a current offer.

• Why own the shares: Retail and AWS offer multiple demand sources, but consumption and financing risks overlap. Breakeven is 3.11× annualized Q2 sales per diluted share. That lower sales multiple does not erase heavy investment needs or imply cheap earnings. [7]
• Trend and levels: Monday $246.15; RSI 41.2; SMA20/50/200 $253.29/$256.15/$240.94; relative volume 0.90×. Above the long average, below both shorter averages. [17]
• Contract quality: Bid/ask $1.23/$1.29; spread $0.06; delta −0.1442; IV 30.73%; volume 1,354/OI 13,796. Last trade September 28, 15:59:31, provider clock. [18]

Next on the radar

• September 29, 11:00–15:15 Toronto: MongoDB investor day after the leadership change; watch enterprise AI priorities. [3]
• September 30, 08:30 Toronto: BEA’s GDP third estimate and August income/outlays, including PCE inflation. Rate expectations matter to all three studies. [10]
• September 30, 16:30 Toronto: Micron’s call tests memory demand and investment plans. [9]

From YieldCove · make ownership the first decision

Before any premium, record why you would hold the shares and what would change your mind. These are correlated technology exposures, not a diversified portfolio. Visit YieldCove to organize your own assumptions. Keeping cash available is also a decision.

AI-assisted editorial by Codex. Educational information, not personalized investment advice. Early assignment is possible; scenarios exclude fees and taxes.