McKinsey's 11 million number just turned AI's labor story from Wall Street anecdote into a P&L input

JPMorgan Chase's Q2 2026 earnings, released July 14, 2026, are the proof that McKinsey's 11 million is not a future tense. CEO Jamie Dimon told analysts that AI had already reduced headcount by 30% to 40% in some divisions, with the firm's KYC unit cost down 40% since 2022 and about 1,000 AI use cases live across risk, marketing, hedging, note-taking and idea generation. The firm is committing ~$2 billion a year to AI within a $19.8 billion technology budget that grew 10% year over year. The financial print rewarded it: net income of $21.2 billion — a U.S. bank record, up 41% year over year — on revenue of $82.5 billion in the quarter, with EPS of $7.70 versus a $5.55 consensus (a 38% beat).

The read-through is uncomfortable for labor and comfortable for shareholders. Banks like JPMorgan are turning the McKinsey 11 million into margin: as long as trading, payments and IB fees keep scaling, every displaced analyst is a direct lift to the operating-leverage line. The risk is that the same playbook triggers the consumer-side damage downstream — 7.6× more displacement among lower-wage workers implies a thinner checking account at the receiving end of every IPO, mortgage and auto loan JPM books.

If workers are the seats, fewer workers mean fewer seats — and the SaaS market is repricing the math. Atlassian reported its first-ever decline in enterprise seat counts in early 2026, and despite a Q4 FY2026 revenue print of $1.766 billion (+28% year over year) and a Q1 CY2026 beat at $1.79 billion (+31.7%), management guided FY2027 revenue growth to only ~13% — a brutal deceleration. The market responded: shares are around $494 a share today versus a 52-week high of $200... wait, that's wrong.

Correction: Atlassian trades near $187.89 against a 52-week range of $56.01–$200, with a market cap around $49 billion. The market is paying for the seat-count risk, not the revenue print. Workday is the second shoe. Q2 FY2027 revenue came in at $2.65 billion (+12.8% year over year) and AI products drove more than 25% of new ACV, but management has now formally introduced Flex Credits — a usage-based pricing layer — because per-seat economics no longer scale when a customer can replace three HR generalists with one administrator running AI agents.

Salesforce has been the most explicit. The company has shifted to consumption credits as its primary monetization vehicle, a structural change to its per-seat DNA. Salesforce's Q2 FY2027 revenue was $11.345 billion (+10.8% year over year) on operating income of $2.331 billion — a healthy margin profile at 20.5% operating income — but the read is in the forward P/E compression. Salesforce trades at 13.99x forward earnings versus Workday's 17.67x and Atlassian's 36.23x, suggesting investors are giving Salesforce credit for having cracked the agent-economy pricing problem while still doubting Workday and Atlassian's transitions.

If 11 million workers need credentialed reskilling, the obvious winners are the online learning platforms. The reality is messier. Coursera — which closed its acquisition of Udemy in May 2026 at a roughly $1.7 billion enterprise value (0.8 COUR shares per UDMY share) — reported Q2 2026 revenue of $299 million, +60% year over year as reported, but down 1% normalized, excluding Udemy. Margins expanded 620 basis points to 62% on the GAAP line, but the stock dropped 17% the next session because the underlying organic business is shrinking. The merger solves scale; it doesn't solve demand intensity.

Skillsoft is the more troubling read. Q1 FY2027 (reported June 9, 2026) revenue came in at $94.5 million, down 4.7% year over year, dragged by a 21% decline in the consumer business and softer government bookings. Yet management trumpeted that AI skill benchmark completions rose 994% year over year and AI content completions were up 261%. The mismatch is the story: AI training engagement is exploding, but enterprise training budgets are tightening because the customers themselves are cutting headcount. Skillsoft trades at $5.71 with a market cap of just $51 million against an analyst target of $50 — a stock that has nearly priced itself to zero while running an AI product people actually want.

Chegg is the obverse. The company that built itself as a subscription homework-helper is a $0.74 stock trading at a $83 million market cap on revenue of $265 million TTM, with a -20% net margin. Generative AI erased its core value proposition faster than the company could pivot, and management has not articulated a credible reskilling answer. McKinsey's 11 million is the demand Chegg could have served; it is instead a textbook case study in what happens when incumbents underinvest in the long tail of the transition.

McKinsey's most actionable line is the asymmetry: lower-wage workers are 7.6× as likely to need a new occupation, and the shrinking sectors — office and administrative support, retail, transportation — are precisely the customer base of discount retail. Dollar General reported Q2 FY2026 (ended July 31, 2026) revenue of $43.6 billion, EPS of $7.51, and a 3.9% net margin — a thin operating buffer even in a healthy environment. The stock has fallen roughly 25% from its 52-week high of $158 to about $119, and analysts carry a $140.93 price target with 19 Holds, 7 Buys and 3 Strong Buys — a market that knows consumer stress is real but disagrees on magnitude.

McKinsey's 11 million puts a number on that stress. If 7% of the labor force needs new occupations by 2035 and the impact is concentrated in the lower three wage quintiles, the consumer-discretionary wallet shrinks faster than headline retail data suggests. Dollar General is the cleanest public expression because its 194,000 employees operate 18,190 stores serving exactly the income brackets McKinsey flags. The bear case: as automation absorbs office/admin and retail roles, traffic and ticket size both decay. The bull case: the same displacement pushes them into Dollar General's core trade-down channel, which explains why management is still expanding store count despite the share-price weakness.

This is the second-order transmission of McKinsey's number: the same11 million that creates the SaaS seat-count compression at the front office also creates the discount-retail traffic compression at the checkout. Investors pricing Dollar General at 15.84x trailing earnings versus a multi-year average of roughly 21x are starting to discount this; investors pricing JPMorgan at 14.18x trailing earnings versus its historical ~12x mean reversion premium have already priced the offset.

Short term (days to quarters). Three near-term catalysts matter: (1) Q3 2026 earnings from the SaaS cohort, where management commentary on seat-count trends will tell us whether Atlassian's decline was idiosyncratic or systemic; (2) the next Skillsoft print, which will show whether the 994% AI-completion surge is converting into enterprise renewal revenue; and (3) any federal or state retraining-policy response to the McKinsey number, given that 85% of growing jobs require a credential — a level of public-sector co-investment that historically shows up in education services M&A. The next 90 days are about whether the market treats McKinsey's 11M as a forecast or as an earnings driver.

Long term (one to three years). Two structural questions dominate. First, does per-seat SaaS revenue model survive the transition, or do consumption-credit pricing structures (Salesforce, Workday Flex) absorb the entire ~$300 billion SaaS TAM? Second, does the 85% credential requirement harden into a public retraining pipeline — community colleges, apprenticeships, credentialing-as-a-service — and which listed players (Stride, 2U if it survives, Skillsoft if it stabilizes) capture it? The bears point to Chegg's collapse as proof that education incumbents cannot adapt; the bulls point to Coursera's Udemy purchase as proof that scale will sort winners. Both can be right at the same time.