Morning Watchlist: Sunday Edition
A quick note from Behind the Markets
AI can lift productivity and still destroy the economics of a vulnerable software vendor. Electronics demand can rebound and still punish a distributor with bloated inventory. Fertilizer can benefit from food-security spending and still get crushed by feedstock costs.
The independent edge is not cheering for a theme. It is locating the cash flow inside the theme.
This week handed us a perfect laboratory: four companies in four different industries all reported in the same 48 hours, all posted record or near-record numbers, and the market treated them very differently. Today we take them apart.
1) AI's Next Cost Is the Labor Bill
AI is being sold as a productivity miracle. Investors should also model the disruption bill.
A bank analysis highlighted this week by Reuters' Mike Dolan found that, after adjusting for cyclical differences, unemployment in occupations highly exposed to AI is now running roughly 0.5 percentage points above what broader labor-market conditions would normally predict — up from about 0.3 point in April. Those occupations account for roughly 30% of employment.
But read the analysis to its own conclusion, because this is where headlines usually stop and the math keeps going: the strategists themselves calculate that AI-related disruption is contributing at most about 15 basis points to the aggregate unemployment rate. Their words for it: small but increasingly visible. Not "jobsmageddon." A trend that is real, growing, and — for now — modest.
And here's the other side, because on this question the evidence is genuinely contested. Stanford researchers looking at the same period found that unemployment among the most AI-exposed workers has actually risen slightly less than among the least exposed — 0.77 points versus 0.85 since 2022 — which reads as a broadly softening labor market, not an AI-driven one. Anthropic's own labor research found no systematic increase in unemployment for highly exposed workers, with only suggestive evidence that hiring of younger workers has slowed. The Dallas Fed reached the same place: the effect that exists is concentrated in workers in their early twenties entering exposed occupations, not in layoffs of people already employed.
Put the studies side by side and a consistent picture emerges: the disruption is showing up at the entry level first. Companies are not firing their way into AI. They are hiring less underneath — fewer junior seats, fewer manual reviews, fewer entry-level analysts. Output up, headcount flat.
That sharpens the investment question. Not "Will AI take every job?" — that's a cartoon question. Ask who captures the savings, who loses the pricing unit, and who owns the verification, security, workflow, and compliance layer around the new system. The winners may be the firms that turn labor cost into measurable throughput. The losers may be companies whose revenue depends on more human seats, more manual reviews, or more junior billable hours. If your software vendor charges per seat and its customers are quietly not filling seats, that is a revenue model with a slow leak — even in a "strong" economy.
Bottom line: AI productivity is bullish for output and dangerous for weak business models. Model the labor substitution before you pay for the growth story — and watch the entry level, because that is where the bill is arriving first.
2) The Electronics Cycle Is Back. Don't Confuse Sales With Power.
The electronics supply chain is sending a stronger signal than the index. Two companies proved it this week — and the contrast between them is the important takeaway.
Avnet, one of the world's largest electronics distributors, reported fiscal fourth-quarter revenue of $8.3 billion on Wednesday — up 47.7% year over year, a record, and nearly a billion dollars above the $7.3 billion the Street expected. Adjusted EPS of $2.28 crushed the $1.67 estimate. Growth came from every region and every end market, and management guided the next quarter up another 10% sequentially. When the pipes of the electronics industry post numbers like that, the demand is real.
Now look at what that demand is worth. Avnet's adjusted operating margin, after expanding 129 basis points, is 3.8%. Its core Electronic Components business runs at 4.1%. That is distribution: enormous revenue, wafer-thin economics, and working capital that swells with every dollar of growth. A year ago — same company, same franchise — quarterly net income was $6.1 million. Six million, on billions of revenue. That is what the bottom of a cycle looks like when you have no pricing power. The stock knows all this, by the way: at roughly $89 it has doubled off its $44 twelve-month low and sits within shouting distance of its $95 high.
Avnet is the read-through, not the pick. The pick is the company that owns the qualification barrier.
The Stock: TTM Technologies (NASDAQ: TTMI)
TTM makes printed circuit boards, RF components, and mission systems — and it sits at the intersection of the two most durable demand streams in electronics. In the second quarter, reported Wednesday: record net sales of $1.0 billion, up 37% year over year. Data Center and Networking — the AI buildout — grew 91% and is now 40% of sales. Aerospace and Defense grew 14%, is 37% of sales, and carries a program backlog above $1.7 billion. Roughly 80% of the company's revenue is tied to those two megatrends. Total book-to-bill was 1.49 — for every dollar shipped, a dollar and a half was ordered.
And here is why it belongs in a section about cash conversion rather than revenue: non-GAAP EPS hit an all-time record $0.99, up 71% — profit growing twice as fast as sales. Adjusted EBITDA margin reached 16.6%, also an all-time high. Cash from operations was $96.4 million, or 9.6% of sales, with $46 million of free cash flow — and that operating cash figure is up from just 2.6% of sales in the first quarter, meaning the growth is starting to convert rather than merely absorb. Management guided the third quarter to $1.10–$1.14 billion and now expects roughly $4.4 billion for the full year with EPS approaching $5.00.
This is what pricing power looks like next to a distributor: TTM's boards are designed into specific programs — defense platforms, AI networking hardware — where qualification takes years and switching suppliers mid-program is somewhere between painful and impossible. Avnet moves boxes. TTM is in the box.
The other side of the ledger, because there always is one. First, concentration is the mirror image of focus: 80% of sales tied to two megatrends means a pause in either one hits hard, and hyperscaler capex plans can change with a single earnings call. Second, that first-quarter cash figure — 2.6% of sales — is a reminder that this business can absorb cash when it ramps; one good conversion quarter is progress, not proof. Third, TTM has two European acquisitions (STG and ILFA) expected to close this quarter, which adds integration work, and its guidance excludes them entirely. Fourth, after a year like this, expectations are demanding — the stock has re-rated substantially with the AI trade, so verify the current price and analyst spread before committing capital, and treat any broad AI-sentiment wobble as something this name will feel immediately.
Bottom line: Electronics demand is improving, but revenue is not the same as economic power. A 48% grower earning four cents on the dollar and a 37% grower earning record margins on a $1.7 billion qualified backlog are not the same trade. Follow cash conversion and qualification barriers.
3) Fertilizer Is a Cash-Flow Trade, Not a Commodity Story
Agriculture investors often chase the crop price and forget the input side.
CF Industries, the largest nitrogen producer in North America, reported first-half net earnings of $1.34 billion, or $8.71 per diluted share, with adjusted EBITDA of $2.18 billion. Second-quarter earnings alone were $727 million, or $4.73 per share. Average selling prices rose across every segment on a tight global nitrogen market. Trailing-twelve-month free cash flow was $1.82 billion. The company bought back 2.0 million shares for $230 million during the quarter and raised its dividend 20% earlier this year.
Before you extrapolate, three details from inside the release that the headline hides.
First, about $170 million of those first-half earnings came from a litigation settlement. One-time. Strip it out mentally before you value anything.
Second, the spread is compressing exactly as the framework predicts. CF's realized natural gas cost rose to $4.01 per MMBtu in the first half from $3.52 a year ago — a 14% feedstock increase. Nitrogen producers convert gas into fertilizer; when the input rises faster than the output, record headlines can coexist with shrinking unit economics. This half, prices rose enough to win the race. That is a market condition, not a business model.
Third, the company ran an extended outage at its Yazoo City complex, which raised maintenance costs and cut volumes. Even the best operators spend real money keeping thirty-year-old plants running.
Low-cost plants, reliable feedstock, disciplined maintenance capital, and a balance sheet that survives a bad crop year... but here is the uncomfortable part. CF Industries passes the test almost perfectly, and that is exactly why it's not a buy right now. CF sits on cheap US natural gas, making it one of the lowest-cost nitrogen producers on earth. The balance sheet is strong, the buyback is real, the payout ratio is under 20%.
And the market has noticed all of it. The stock is up roughly 55% this year to around $126 — which is above the analyst consensus target of roughly $120–$123. All 19 covering analysts, taken together, land at a consensus of Hold. When a cyclical commodity producer trades above the Street's target after a 55% run, with a tight-supply narrative on every screen, you are being offered the top of the framework's own checklist at the top of its own price curve. New nitrogen capacity is being built globally, trade flows reroute quickly, and the gas-cost creep is already visible in this very report.
So CF goes on the watchlist, not the buy list. The time to own the low-cost producer is when the spot-price narrative is boring, not when it is on the front page. If nitrogen prices normalize or gas keeps climbing, this exact company — with this exact cost position — will go on sale, and the framework will still be true.
Bottom line: Food security creates demand, not guaranteed margins. In fertilizer, cost position and cash conversion matter more than the spot-price narrative — and even the right producer at the wrong price is the wrong trade.
4) The Real Small-Cap Edge Is Reading the Cash Conversion
The current earnings season is giving investors a rare gift: plenty of revenue growth and plenty of reasons to distrust it. Ducommun gave us a live specimen this week — so let's actually run the checklist.
The headline: record second-quarter revenue of $224.5 million, up 12% year over year, in aerospace, defense, and space products. Commercial aerospace rose 16% on 737 MAX and A320 rate increases; military and space grew on missile programs. Gross margin hit an all-time record 28.0%, up 160 basis points. CEO Steve Oswald: "I could not be happier."
Now the questions that matter.
Does the growth become operating cash flow? Yes. Operating cash flow rose to $33.5 million in the quarter — roughly 15% of revenue, a genuinely strong conversion rate for a mid-cap manufacturer.
Is the company funding its customers, or are orders funding the company? Orders are winning. Bookings were $309.7 million against $224.5 million of revenue — a 1.4x book-to-bill — and remaining performance obligations hit an all-time high of $1.2 billion. The backlog is growing faster than the P&L.
What about the balance sheet? Improving. Long-term debt fell to $271.4 million from $298.8 million at year-end. The company is paying debt down during a growth ramp, which is the opposite of funding growth with the balance sheet.
Did management issue stock after calling the quarter "record"? No equity issuance disclosed. Adjusted earnings grew 35% — and note that the flashier 60% GAAP net-income growth got a boost from a one-time compensation clawback, so use the 35% figure. That distinction between the GAAP number and the operating number is itself the discipline this section preaches.
So Ducommun passes the cash-conversion test about as cleanly as a company can. Which brings us to the part that makes this a teaching example rather than a table-pounding buy: the stock already knows. Heading into this report, DCO traded around $192 — against an average analyst target near $181. The shares trade above the consensus of the analysts who follow it, after the market spent months rewarding exactly the missile-program visibility and margin execution this quarter confirmed. Management also flagged that destocking headwinds continue through the rest of 2026, easing only gradually.
That is the whole lesson in one company. The checklist tells you whether the business deserves a premium. The price tells you whether you are being asked to pay two premiums. A boring company that converts orders into cash can beat a glamorous company that converts orders into dilution — but only if you buy the cash converter before everyone agrees with you. Put DCO on the watchlist with CF: right businesses, prices that have already taken the credit.
Bottom line: Small-cap value is not just a low multiple. It is a cash-conversion advantage the market has not fully priced — and this week's tape shows the market is getting better at pricing it, which means you have to get there earlier.
Before You Go
The strongest themes are not the loudest ones. They are the ones that keep working after the narrative changes.
AI must earn its power and labor bill — and so far it is sending the bill to the entry level. Electronics suppliers must convert backlog into cash — one did it at record margins this quarter, and one did it at four cents on the dollar. Fertilizer producers must defend their spread — even the best one just showed you gas costs rising 14%. Small caps must finance growth without handing the upside to new shareholders — and the one that managed it is already trading above its own analysts' targets.
Four record quarters. One buy candidate, two watchlist names, and a distributor that proves the cycle without earning the premium. That ratio is what a filter looks like when it is working.
Found this helpful? Share it with others.
Written by Behind the Markets