Texas Instruments TXN

Reference table

1. The quarter, and the operating leverage

From the 10-Q filed 2026-07-24 and the quarterly income series:

Two observations that matter more than the headline.

First, the year-on-year growth rate has accelerated for three consecutive quarters — 10.4%, 18.6%, 22.8% — and the 10-Q attributes the June quarter to "broad growth led by industrial, data center and automotive." Industrial is the largest and most cyclical end market for this company and its recovery is the single most important demand signal in the file.

Second, and this is the mechanical point: gross margin rose 550 basis points and operating margin 900 basis points in two quarters. From 55.9% and 33.3% in the December quarter to 61.4% and 42.3% in June. That is not a pricing story; it is fixed-cost absorption. Texas Instruments owns its fabs and has just built three new ones, so incremental volume falls through at a very high rate — and the same mechanic runs in reverse when volumes fall.

The five-year context, which is the discipline the quarterly numbers require:

Read the revenue column and the whole valuation question appears. $18.3B, $20.0B, $17.5B, $15.6B, $17.7B — a 3.6% DECLINE across five years, through the largest industrial capital cycle in decades. Peak-to-trough was −21.9%. Trailing revenue of $19.453 billion has still not reclaimed the 2022 peak of $20.028 billion, and gross margin at 58.3% trailing remains 1,050 basis points below the 68.8% of 2022.

Segment and geographic detail. seg_prod for FY2025 reads Analog $14.006 billion and Embedded Processing $2.697 billion, summing to $16.703 billion against reported revenue of $17.682 billion — a $979 million shortfall, because the "Other" line present in FY2023 and earlier has been dropped from the two most recent years (see Section 6). What can be said is that Analog is roughly 79% of the company and Embedded roughly 15%.

seg_geo is clean and ties exactly:

China at 21.4% and growing 25.5% is the fastest-growing major geography — and it is also the source of the one bearish claim in the knowledge base, which asserts that Chinese domestic microcontroller and power-chip makers are "now matching and undercutting" Texas Instruments. The threat is not yet visible in the revenue line, which is precisely when it is worth watching.

2. The capital-expenditure inflection — the central finding

This is the part the vendor file cannot show and the filing states outright.

> "Consistent with our capital management strategy, we expect between $2 billion to $3 billion of capital expenditures in 2026. Beyond 2026, capital expenditures will be dependent on revenue and growth expectations." — 10-Q, 2026-07-24

> "Capital expenditures were $1.19 billion compared with $2.43 billion in the year-ago period..." — same filing, first half of 2026

Capital expenditure has gone $2.797B (2022) → $5.071B (2023) → $4.820B (2024) → $4.550B (2025) → a guided $2-3B (2026). At the guided midpoint of $2.5 billion that is a 45% reduction on 2025 and 51% below the 2023 peak, and the first half is already tracking to it.

The consequence for cash generation is arithmetic. The company reports it directly:

> "Our cash flow from operations of $8.7 billion for the trailing 12 months again underscored the strength of our business model, the quality of our product portfolio and the benefit of 300mm production. Free cash flow for the same period was $6.5 billion. Over the past 12 months we invested $3.9 billion in R&D and SG&A, invested $3.3 billion in capital expenditures and returned $5.8 billion to shareholders."

Three notes on those figures. First, TI's free-cash-flow definition adds back CHIPS Act proceeds — the 10-Q says so explicitly: "Free cash flow is calculated as cash flows from operating activities... less capital expenditures, plus proceeds from CHIPS Act incentives." The vendor-derived figure, which does not add them back, is $5.355 billion. We use both and label them: $6.5 billion on the company's definition (a 2.51% yield), $5.355 billion on ours (2.07%). Second, trailing free cash flow has therefore already more than doubled from FY2025's $2.603 billion without any further improvement in the capex line. Third, $5.8 billion returned to shareholders against $6.5 billion of free cash flow is a 89% payout — and it is almost entirely dividend.

The CHIPS Act contribution is larger than most readers will expect and is disclosed precisely:

> "We expect to continue benefiting from the CHIPS Act. This includes the 35% ITC on qualifying manufacturing investments as well as direct funding of up to $1.6 billion, of which we have received $630 million, for our three large-scale 300mm wafer fabs located in Sherman, Texas, and Lehi, Utah."

In the first half of 2026 CHIPS Act incentives provided $1.10 billion of cash proceeds against $260 million a year earlier, plus a $301 million investment-tax-credit benefit reducing income taxes payable against $203 million. That is roughly $1.4 billion of government support in six months, on a company generating $4.22 billion of operating cash flow in the same period. It is real, it is disclosed, and it is a policy dependency.

And the reason for all this spending, in the company's own words, half-weighted but quantified:

> "We have focused on creating a competitive structural cost advantage by investing in our 300mm wafer production, which describes the diameter of the wafer on which our chips are produced, and costs about 40% less than a chip built on a 200mm wafer... These investments have uniquely positioned TI to deliver dependable, low-cost 300mm capacity with the ability to equip and ramp factories to support customer demand. We believe this approach supports free cash flow per share growth across a range of market conditions."

A 40% unit-cost advantage on the majority of production, in an industry where most analog competitors run 200mm or use foundries, is the strongest structural claim any company in this batch makes about itself — and it is why gross margin can reach 61.4% on revenue that has not yet reclaimed its 2022 peak.

3. Balance sheet, the acquisition, and the buyback nobody has mentioned

From the 10-Q at 2026-06-30, against the vendor's 2025-12-31 balance sheet:

Correction 1 — netDebt omits short-term investments, again. The vendor's $12.166 billion nets total debt against cash and equivalents only, ignoring $1.656 billion of short-term investments on the same balance sheet. The correct year-end figure is $10.510 billion. By 2026-06-30 liquidity had risen to $7.00 billion while $1.20 billion of long-term debt was issued, putting net debt near $9.6 billion. We note that the vendor's enterpriseValueTTM of $269.416 billion implies net debt of $10.392 billion — close to our corrected year-end figure and therefore usable, even though the netDebt field itself is wrong by $1.656 billion.

Correction 2 — a $7.5 billion acquisition and a $5 billion credit facility appear in NO vendor field. From the 10-Q:

> "As announced on February 4, 2026, we have entered into a definitive agreement to acquire Silicon Labs for $231.00 per share in an all-cash transaction, representing a total enterprise value of approximately $7.5 billion... which is expected in the first half of 2027, subject to receipt of regulatory approvals... We expect to fund the transaction with a combination of cash on hand and debt financing."

> "In June 2026, we entered into a 364-day delayed draw term loan credit facility for borrowings up to $5 billion to support the Silicon Labs acquisition consideration and related transaction expenses. As of June 30, 2026, there were no outstanding borrowings on the delayed draw term loan credit facility."

A $7.5 billion enterprise-value acquisition — roughly 2.9% of Texas Instruments' market capitalisation and more than its entire trailing free cash flow — is pending, and neither the transaction nor the $5 billion facility supporting it appears anywhere in the vendor payload. This is the "off-balance-sheet items in no vendor field" defect class. Post-closing net debt will be materially above the ~$9.6 billion figure used here, and no forward estimate in this file can be assumed to include Silicon Labs' contribution.

Correction 3 — the buyback has been all but suspended and nobody appears to have said so.

> "We used $185 million to repurchase 0.9 million shares of our common stock compared with $955 million to repurchase 5.4 million shares in the year-ago period."

First-half repurchases fell 81% by value and 83% by share count. Full-year 2025 repurchases were $1.477 billion. At the first-half pace, 2026 repurchases will be roughly $370 million — 0.14% of market capitalisation. The filing does not explain the reduction, but the timing relative to the Silicon Labs announcement in February is difficult to read any other way. Total shareholder yield is therefore approximately 2.14% — a 2.00% dividend plus a negligible buyback — against $5.8 billion returned over the trailing twelve months (2.24% of market capitalisation) which is not a forward run-rate.

Working capital is the genuine balance-sheet risk. daysOfInventoryOutstandingTTM reads 207.3 days — the highest in this batch — on $4.804 billion of inventory, with a 224.0-day cash conversion cycle. Some of that is structural, because analog products have decade-long lives and TI deliberately holds finished goods to service design-ins. But inventory rose from $3.999 billion (2023) to $4.804 billion (2025) while revenue fell and then partially recovered, and a 207-day balance is a lot of capital to have committed if the industrial recovery stalls.

Interest coverage is 12.9x and debtToEquityRatioTTM is 0.78x. The vendor's debtToEquityScore of 1 out of 5 is harsh on those figures and we note it, but unlike the negative-equity cases in this batch the input is defined and the score is merely aggressive.

4. Valuation — priced in or room?

At $283.63 (market cap $259.02B, ~913.2M shares, rebuilt enterprise value approximately $269.5B):

Estimate coverage is the best in this batch — 23-25 analysts on revenue and 18-22 on EPS through FY2028. The FY2029 and FY2030 rows carry 10 and 8 EPS analysts and are excluded from conclusions, particularly because FY2029 epsAvg of $11.32 is BELOW FY2028's $11.927 while revenue rises from $27.517B to $29.130B — a non-monotonic shape we do not use.

Per the data contract we tested est.ebitAvg and est.ebitdaAvg and they PASS every screen — ebitAvg below ebitdaAvg in every year, netIncomeAvg below ebitAvg in every year, no negative rows, no fabrication signature, and the FY2025 estimate of $5.472 came in against an actual $5.45 (0.4% high) with FY2024's $5.097 against $5.20 (2.0% low). This is the cleanest estimate block in this batch and we say so. One caveat: the FY2028 EPS dispersion is extreme — a mean of $11.927 against a high of $18.176 and a low of $9.971, a 82% spread — which reflects genuine disagreement about the shape of the recovery rather than a data error.

Peer context. The vendor peer set is appropriate for once: Analog Devices ($185.23B), Broadcom ($1,989.43B), Intel ($509.14B), Microchip ($43.80B), Monolithic Power ($65.59B), Marvell ($191.43B), Micron ($1,008.17B), NXP ($59.88B), ON Semiconductor and Lattice. The genuine analog comparables are Analog Devices, Microchip, NXP and ON Semiconductor. No peer multiples are supplied in the file, so no peer-multiple comparison is drawn. The scale fact: Texas Instruments' $259.02 billion market capitalisation is 1.40x Analog Devices' $185.23 billion, and Micron — a memory company in the same batch's supply chain — is now larger than both combined at $1,008.17 billion.

4a. What today's price assumes (the inversion)

At $283.63 — 43.1x trailing and 27.9x FY2027 consensus — the price embeds:

• Revenue reaches $24.892 billion in FY2027 and $27.517 billion in FY2028. (Consensus; 25 analysts each.) That would be 24% and 37% above the 2022 peak of $20.028 billion, from a company whose revenue has DECLINED 3.6% over the last five years. The nearest check is 2026-10-27, where consensus revenue of $5.916 billion implies +8.3% sequentially on a quarter that already grew 13%.
• Gross margin keeps expanding from 61.4% toward the high sixties. (Our inference from the consensus EPS path.) FY2027 consensus EPS of $10.155 on revenue of $24.892 billion implies a net margin of roughly 37% against 31.1% trailing — which on this cost structure requires gross margin near 65-67%, i.e. most of the way back to the 68.8% of 2022. This is the most fragile assumption in the price and it depends on volume absorbing the new 300mm capacity.
• Capital expenditure stays near $2-3 billion. (Company guidance for 2026; explicitly NOT guided beyond.) The 10-Q says "Beyond 2026, capital expenditures will be dependent on revenue and growth expectations." This company has twice chosen to build ahead of demand, and a decision to build again defers the free-cash-flow inflection the whole medium-term case rests on.
• The market keeps paying 25-30x forward earnings. (Our number.) At 20x FY2027E the stock is $203; at 35x it is $355. The multiple band is worth $152 — 54% of the current price.
• Silicon Labs is free. (Filing-confirmed transaction; our observation.) A $7.5 billion acquisition closing in the first half of 2027 appears in no forward estimate we can identify, and neither does the debt to fund it. That cuts both ways and it is unmodelled in either direction.

4b. The return bridge (why the multiple moves)

Expected return over the next twelve months decomposes as: EPS growth (+17.4%, from FY2027E $10.155 to FY2028E $11.927) + multiple drift (COMPRESSION, from 27.9x to roughly 25x, −10.4%) + shareholder yield (+2.14%) ≈ +7% to +9%.

Almost all the return is earnings recovery and the multiple is a headwind. Our base of $295 assumes modest multiple compression, and the reason is the base rate rather than pessimism: 43x trailing and 28x forward is a growth multiple on a business whose five-year revenue change is negative. As the recovery matures and the year-on-year comparisons harden, the multiple should converge toward where a high-quality cyclical analog franchise trades.

Sensitivity, stated openly, on the FY2027 consensus EPS of $10.155: at 20x the shares are $203; at 25x, $254; at 29.1x, $295; at 32x, $325 — exactly the street's target; at 37.4x, $380. On FY2028E of $11.927: 20x gives $239, 25x gives $298, 30x gives $358.

The bull case at $380 requires multiple expansion from an already-full level and that is its weak leg. The bear case at $203 requires no earnings miss at all — only a return to 20x forward, which is roughly where Texas Instruments traded for much of the last decade.

4c. Variant perception (where we differ, what would surprise)

• We differ from the street on level, and the street itself is unusually divided. Consensus $325 across 65 analysts — but 31 buy, 28 hold and SIX sell, the largest sell count in this batch, with targets from $225 to $405. Our base of $295 is 9.2% below the mean and 31% above the low target. On a name with this much analyst disagreement, our contribution is not a directional edge.
• Where we do think we add something is the capital-expenditure inflection, and we suspect it is under-modelled. Capital expenditure guided from $4.550 billion to $2-3 billion, with first-half spend down 51%, on a company whose free cash flow was crushed for four years by exactly this line. Trailing free cash flow has already gone from $2.603 billion (FY2025) to $6.5 billion on the company's own definition — a 2.5x increase — with the revenue recovery still incomplete. Watchable number: full-year 2026 capital expenditure against the guided $2-3 billion, and the 2027 capex commentary at the October or January call. If capex stays at $2.5 billion on FY2027 revenue of $24.9 billion, free cash flow runs near $8 billion and the yield at today's price is 3.1%.
• We think two disclosed items are being ignored and both are in the same paragraph of the 10-Q. The $7.5 billion Silicon Labs acquisition with a $5 billion delayed-draw term loan, and the 81% reduction in the buyback from $955 million to $185 million in a half year. The second is almost certainly a consequence of the first, and together they mean that Texas Instruments has redirected roughly $8 billion of capital away from shareholders and toward a microcontroller acquisition — which, given the knowledge base's one bearish claim about Chinese competition in exactly that category, may be a defensive move. Nobody in this file connects those three facts and we do.
• We surface the bearish claim the entity lane missed. The knowledge base's three entity claims are all one bullish voice. The single most consequential claim about Texas Instruments in the whole knowledge base is in the TEXT lane, carries a named independent speaker, and asserts that China has built a semiconductor industry with "microcontrollers and power chips now matching and undercutting" the company. China is 21.4% of revenue and grew 25.5% last year. We weight it because it is the only structural bear argument available and because Silicon Labs is a microcontroller business.
• Positive surprise that would force a re-rate: third-quarter revenue above the $5.916 billion consensus with gross margin above 62%; full-year capital expenditure at the $2 billion low end of guidance; a formal capital-return framework restored after the acquisition closes; or industrial revenue disclosed as growing faster than the company average.
• Negative surprise that would break the thesis: any sequential revenue decline; gross margin failing to hold above 60%; a 2027 capital-expenditure guide above $4 billion, which would defer the free-cash-flow inflection; an inventory build above 220 days; regulatory delay or repricing of the Silicon Labs transaction; or the CHIPS Act framework being altered, given roughly $1.4 billion of first-half benefit.

Synthos fair values

All three anchors are multiples of the FY2027 consensus EPS of $10.155 (22 analysts — strong coverage), cross-checked against FY2026E and FY2028E.

• Bear ~$203 — 20.0x FY2027E. Cross-checks: 24.2x FY2026E; 17.0x FY2028E; 23.0x the FY2027 consensus LOW of $8.822; 10.8% below the 200-day average of $227.56; a 2.80% dividend yield. The scenario: the industrial recovery stalls, gross margin plateaus near 60%, capital expenditure is guided up again for 2027, Chinese competition becomes visible in the China revenue line, and the multiple returns to roughly where this business traded for most of the last decade. −28.4%. Note this requires no miss against FY2027 consensus — only a normal multiple.
• Base ~$295 — 29.1x FY2027E. Cross-checks: 35.1x FY2026E; 24.7x FY2028E; enterprise value of $279.0B at 11.2x FY2027E revenue; a 2.7% free-cash-flow yield on roughly $8 billion of normalised free cash flow. Sensitivity: 25x gives $254, 32x gives $325 — exactly the street's target. The scenario: the recovery continues on the consensus path, capital expenditure stays inside the guided band, free cash flow inflects toward $8 billion, and the multiple compresses modestly as the year-on-year comparisons harden. +4.0%, plus 2.14% of shareholder yield.
• Bull ~$380 — 37.4x FY2027E. Cross-check: 31.9x FY2028E; 25 dollars below the street's high target of $405. This bull case REQUIRES multiple expansion from an already-full level and that is its weak leg. The scenario: gross margin returns toward the 68.8% of 2022 as 300mm volume absorbs the new fabs, capital expenditure lands at $2 billion, free cash flow exceeds $9 billion, Silicon Labs is accretive, and the market re-rates the manufacturing cost advantage as structural rather than cyclical. +34.0%.

Base is 4.0% above spot; asymmetry roughly 1.20:1 to the upside (28.4% down, 34.0% up). That is a thin ratio on a full multiple. What makes this a Hold rather than a Watch is that the free-cash-flow inflection is real, disclosed and in progress, and a holder is being paid a 2.00% dividend to see it through — but the entry offers no discount and our number sits below the street's.

5. Knowledge base — three claims, one voice, two days

Raw entity hits: 3. Text matches: 1. Independent voice-days: 1.

The search covered TXN and Texas Instruments across the entity field of all 51,928 distilled claims. Stance: 3 bullish, 0 neutral, 0 bearish.

All three entity claims come from a single speaker — Jordi Visser — across a two-day window on 2026-04-25 and 2026-04-26. Two carry speaker: jordi_visser and one speaker: Jordi Visser, all with speaker_role: independent and skill 2.0. Breadth is one. There is no second voice anywhere in the entity lane.

> 2026-04-25 · bullish · conviction 85 · entities NVDA, AVGO, MU, MRVL, TXN, QCOM

> "The shift to agentic/inference AI (east-west traffic) broadens the semi trade beyond GPUs to memory, CPUs, networking and packaging — Micron, Marvell, TXN benefit; compute shortage runs the rest of the year."

> 2026-04-25 · bullish · conviction 80 · entities ON, TXN

> "Power semiconductors (On Semi, Texas Instruments) are going through the roof now that Nvidia is chasing Korean power/shipbuilding partners for power."

> 2026-04-26 · bullish · conviction 75 · entities TXN

> "Texas Instruments, an industrial-demand barometer, reported industrial demand improving while data-center flourishes; AI sensing/power penetrating edge form factors — robots, EVs, test equipment, not just cloud GPUs."

Only the third is a claim about this company rather than a basket. It is a good one, and it is corroborated by the June quarter: the 10-Q attributes growth to "industrial, data center and automotive", which is exactly the sequence the claim describes. A single independent voice reading the industrial cycle correctly three months before the print is worth recording — and one voice is not a lane.

The most consequential claim about Texas Instruments in the entire knowledge base is not in the entity lane at all. It appears as a text match, carries a named independent speaker, and is bearish:

> 2026-02-05 · bullish-stanced but bearish-for-TXN · conviction 60 · thesis · channel dylan_patel · speaker Dylan Patel · speaker_role independent

> "China has built a real semiconductor industry, insourcing more, with microcontrollers and power chips now matching and undercutting T[exas Instruments]..."

We surface it, we weight it, and we make it a long-horizon watch item, for three reasons. First, microcontrollers and power chips are precisely Embedded Processing and a large part of Analog — the company's two segments. Second, China was $3.781 billion of 2025 revenue, 21.4% of the total, and grew 25.5% — the fastest-growing major geography. Third, Texas Instruments is spending $7.5 billion to buy a microcontroller business, which is at least consistent with a defensive read.

Conclusion, stated without inflation. The Synthos knowledge base has essentially no view on Texas Instruments as a security. It has one voice, expressed over two days, reading the industrial cycle correctly, and one named independent bearish claim about structural Chinese competition that the entity tagger did not attach to the ticker. Conviction is Very Low and breadth is 1. That thinness is a direct input to the Hold verdict — there is no knowledge-base reason to prefer this name to any other semiconductor exposure in this batch.

6. Data integrity — what we rejected from the vendor file and why

Eight findings, and the largest is an omission rather than an error.

1. A $7.5 billion pending acquisition and a $5 billion credit facility appear in NO vendor field. The 10-Q discloses the agreement to acquire Silicon Labs at $231.00 per share, an enterprise value of approximately $7.5 billion, announced 2026-02-04 and expected to close in the first half of 2027, to be funded "with a combination of cash on hand and debt financing", plus a $5 billion 364-day delayed-draw term loan facility entered in June 2026 and undrawn at the balance-sheet date. Neither the transaction nor the facility is reflected in bal_a, km_ttm, ratios_ttm or est. This is the off-balance-sheet defect class: post-closing net debt will be materially above the figure used here, and no forward estimate can be assumed to include the acquired business.

2. The buyback has been cut 81% and the vendor file cannot show it. The 10-Q: "We used $185 million to repurchase 0.9 million shares of our common stock compared with $955 million to repurchase 5.4 million shares in the year-ago period." The vendor's most recent repurchase figure is the FY2025 annual of $1.477 billion. At the first-half pace, 2026 repurchases will be roughly $370 million — a quarter of last year — and the shareholder yield falls to approximately 2.14%, essentially all dividend.

3. netDebt omits $1.656 billion of short-term investments — corrected. The vendor's $12.166 billion is total debt less cash and equivalents only. The corrected year-end figure is $10.510 billion, and by 2026-06-30 liquidity had risen to $7.00 billion (cash $3.66B plus short-term investments $3.34B) against $1.20 billion of new debt issued, putting net debt near $9.6 billion. Note that enterpriseValueTTM of $269.416 billion implies net debt of $10.392 billion — close to the corrected year-end figure — so the EV is usable even though the netDebt field is not.

4. seg_prod omits the "Other" line in the two most recent years — flagged. FY2025 reads Analog $14.006B and Embedded Processing $2.697B, summing to $16.703B against reported revenue of $17.682B — a $979 million, 5.5% shortfall. FY2024 has the same problem ($14.694B against $15.641B). FY2023 and earlier carry an explicit "All Other Segments" line of $1.111B. We use the block only to establish the approximate Analog/Embedded split and draw no precise conclusion from it.

5. seg_geo FY2022 contains two identical values — flagged, recent years clean. The FY2022 entry reads EMEA: 4,807,000,000 and CHINA: 4,807,000,000 — the same figure to the dollar, which is almost certainly a duplication. FY2025 and FY2024 tie to reported revenue EXACTLY ($17.682B and $15.641B) and are used without qualification.

6. Two free-cash-flow definitions are in play and they differ by $1.1 billion — both reported. The 10-Q states plainly: "Free cash flow is calculated as cash flows from operating activities... less capital expenditures, plus proceeds from CHIPS Act incentives." On that basis trailing free cash flow is $6.5 billion; the vendor's conventional calculation gives $5.355 billion. We quote both, label both, and use the conventional figure for the yield in our fair-value cross-checks while noting the company's figure in the narrative.

7. The FY2029 and FY2030 estimate rows are non-monotonic — excluded. FY2028 epsAvg of $11.927 falls to $11.32 in FY2029 and recovers to $12.16 in FY2030, while revenue rises in every year ($27.517B → $29.130B → $33.594B). EPS declining while revenue grows 5.9% is not a shape a bottom-up model produces, and the rows carry 10 and 8 EPS analysts against 18 for FY2028. Neither informs any conclusion.

8. est.ebitAvg and est.ebitdaAvg PASS every screen — a verified-clean check worth recording. ebitAvg is below ebitdaAvg in every year; netIncomeAvg is below ebitAvg in every year; no negative rows; no fixed-percentage fabrication signature; and the historical rows reconcile well to actuals (FY2025 estimate $5.472 against a delivered $5.45 — 0.4% high; FY2024 estimate $5.097 against $5.20 — 2.0% low). This is the cleanest estimate block in this batch and, against a 33.5% defect rate across the universe, it is a real finding. One caveat: the FY2028 EPS dispersion is extreme — mean $11.927, high $18.176, low $9.971, an 82% spread — reflecting genuine analyst disagreement rather than a data error.

9. quote.yearHigh/yearLow disagree with tech.hi52/lo52 — we use tech. The quote block reports $334.03 / $152.73; the computed block reports $332.28 / $153.33. Discrepancies of 0.5% and 0.4% — the smallest in this batch. We use tech for consistency with the moving averages.

Not defects, correctly reported and independently confirmed: the June-quarter income statement matches the 10-Q ($5.463 billion of revenue, $1.980 billion of net income, $2.14 of EPS); the implied share count of 913.2 million is consistent with the 920 million diluted weighted average with no share-class complication; interestCoverageRatioTTM of 12.9x is consistent with the debt and income figures; capexToRevenueTTM of 17.0% correctly reflects the trailing capital-expenditure decline; and the FY2025 and FY2024 seg_geo blocks tie to reported revenue to the dollar.

Vendor rating — mixed, and mostly defensible. The composite reads B− / 3 with priceToEarningsScore 1 and priceToBookScore 1 against returnOnEquityScore and returnOnAssetsScore of 5. At 43.1x trailing earnings and 14.4x book, the two low scores are correct signals rather than artefacts and we agree with them. debtToEquityScore of 1 on a 0.78x ratio with 12.9x interest coverage is harsh, and discountedCashFlowScore of 2 reflects the depressed free cash flow that is precisely what is now reversing. We accept the valuation signals and reject the leverage one.

Non-equity tripwire — checked and passed. TXN is common stock, NASDAQ-listed, one class. Beta 1.325; a 52-week range of $153.33 to $332.28 — a 117% spread; a variable, rising dividend; volume of 8.31M shares (roughly $2.36B of turnover). This is common equity.

7. Technicals

• Price $283.63. −14.64% from the 52-week high of $332.28; +84.98% above the 52-week low of $153.33. Position within the annual range: 73rd percentile.
• Moving averages, split: −4.5% BELOW a 50-day average of $297.14, which is declining; +24.6% above a 200-day average of $227.56, which is rising. The long-term uptrend is intact; the intermediate trend is broken.
• RSI 39.3 and MACD −6.06 — weak, and among the weakest readings in this batch alongside Goldman Sachs.
• Maximum drawdown from peak over the trailing year: −14.64% — the current price is the low point of the drawdown.
• Relative performance, and it inverts: 3-month +0.98% vs SPY +7.59% and QQQ +7.67% — a 6.6-point underperformance; 6-month +26.05% vs SPY +11.09% and QQQ +15.71%; 12-month +56.82% vs SPY +24.26% and QQQ +30.80%. The stock has done nothing for three months while its year-on-year growth rate accelerated from 10.4% to 22.8% and gross margin rose 550 basis points.
• Sentiment: 31 buy, 28 hold, 6 sell across 65 analysts — the largest sell count in this batch. Consensus target $325.00 (+14.6%), median $322.50, high $405 (+42.8%), low $225 (−20.7%).

Today's move

TXN closed 2026-08-04 at $283.63, up 5.42% or $14.59 from a $269.04 close. It opened at $276.92, traded $275.12 to $285.52, and closed within 0.7% of the day's high on 8.31M shares against an 8.90M average. No company-specific news is in this file for the date; the last company event was the 2026-07-22 second-quarter release, thirteen days earlier. This was a sector session — the entire semiconductor complex rose sharply — and Texas Instruments participated from a position 14.6% below its own high with an RSI of 39, which is a bounce inside a consolidation rather than a breakout.

The honest read. Three months of going nowhere while the operating numbers improved sharply is either the market anticipating that the recovery is priced, or an opportunity. Our arithmetic says the former: at 43x trailing and 28x forward, three months of flat performance has not created a discount.

8. Insiders — an officer stock award

The most recent transaction in the file is an A-Award of 7,154 shares at a price of $0 to Julie C. Knecht, Vice President and Chief Accounting Officer, filed 2026-07-31 for a transaction dated 2026-07-24, leaving 16,294 shares held.

The reading. This is a routine equity grant, awarded at a zero cost basis shortly after the second-quarter release, and it is a compensation event rather than a decision. The file contains zero open-market purchases and zero open-market sales. Neutral, and reported as neutral. For a company that has just announced a $7.5 billion acquisition and cut its buyback 81%, the absence of any discretionary insider activity is unremarkable but also uninformative.

9. Verdict, kill-criteria and flip conditions

Hold.

What is genuinely good, and all of it is filing-verified: second-quarter revenue of $5.463 billion, up 22.8% year on year and 13% sequentially, with growth "led by industrial, data center and automotive"; gross margin recovering 550 basis points to 61.4% and operating margin 900 basis points to 42.3% in two quarters; three consecutive quarters of accelerating year-on-year growth (10.4%, 18.6%, 22.8%); a capital-expenditure programme ending — guided from $4.550 billion to $2-3 billion, with first-half spend down 51% — after $17.2 billion invested over four years; trailing free cash flow of $6.5 billion on the company's definition against $2.603 billion in FY2025; a 35% CHIPS Act investment tax credit plus up to $1.6 billion of grants that delivered roughly $1.4 billion of benefit in the first half alone; a stated and quantified 40% unit-cost advantage from 300mm production; 12.9x interest coverage; and a 2.00% dividend that has grown every year in the data we hold.

What we are declining to pay for: 43.1x trailing earnings, 27.9x FY2027 consensus, 13.3x sales and 29.6x trailing EBITDA, on a company whose 2025 revenue of $17.682 billion was still 11.7% below its 2022 peak and whose five-year revenue change is minus 3.6%; 207.3 days of inventory, the highest in this batch; a $7.5 billion pending acquisition with a $5 billion delayed-draw facility, neither of which appears in any vendor field or forward estimate; a buyback cut 81% to fund it; a base fair value of $295 that is 4.0% above spot and 9.2% below the street's $325; six sell ratings, the most of the twelve names here; and a knowledge-base lane of three claims from one speaker across two days.

The distinction that matters. Texas Instruments is doing exactly what a well-run cyclical manufacturer should: it built capacity through the downturn, funded it partly with government support, and is now harvesting operating leverage as volumes return while capital expenditure falls away. The free-cash-flow inflection is real, disclosed, in progress and — we suspect — under-modelled. But the price already reflects a recovery that has not yet reclaimed the last peak, and the multiple is a growth multiple on a business with no five-year growth. A holder is being paid 2% to watch the inflection happen. A buyer at 43x trailing is paying for the whole of it in advance.

Pre-registered KILL criteria — what would take this to Avoid:

• Gross margin failing to hold above 60% in any quarter, which would mean the operating leverage has stalled short of the target.
• Any sequential revenue decline, breaking the three-quarter acceleration.
• A 2027 capital-expenditure guide above $4 billion, which would defer the free-cash-flow inflection the medium-term case rests on. The 10-Q explicitly declines to guide beyond 2026.
• Inventory above 220 days outstanding on a rising revenue base, which would indicate the recovery is a channel build rather than end demand.
• China revenue declining, particularly if attributed to domestic competition — the specific mechanism the knowledge base's one bearish claim identifies.
• Silicon Labs repriced, delayed or blocked, or the associated debt taking net debt above roughly $18 billion.

Pre-registered FLIP conditions — what would take this to Buy — Tactical:

• Price below approximately $240 — 23.6x FY2027E and near the 200-day average of $227.56 — at which the base case is +23% and the dividend yield 2.37%. This is 15% below today's close and inside the range this stock has traversed in six months.
• Full-year 2026 capital expenditure at the $2 billion low end of guidance with a 2027 guide at or below $3 billion, which would put normalised free cash flow near $8-9 billion and the yield above 3%.
• Gross margin above 63% in any quarter, which would put the FY2027 consensus EPS of $10.155 comfortably in reach and validate the 300mm cost claim.
• A restored capital-return framework after the Silicon Labs closing, particularly a buyback returning toward $1.5 billion a year.
• Genuine knowledge-base breadth appearing. At present the lane is one voice across two days, and the only structural bear argument had to be recovered from the discarded text lane. That thinness alone caps the conviction contribution at zero.

Where TXN fits in the Synthos Framework Portfolio. The semiconductor / industrial-cycle sleeve, held rather than added to, at a 1.5% position with a 3% target on a fill near $240 or a confirmed 2027 capex guide. On batch overlap: Texas Instruments is the ANALOG expression of the semiconductor cycle and is genuinely different from the others here — its demand is industrial, automotive and personal electronics rather than data-centre, its products are decade-lived rather than annually refreshed, and it manufactures rather than designs. It is therefore the least correlated semiconductor exposure in this batch and the one whose fortunes turn on industrial production rather than artificial-intelligence capital expenditure — though the knowledge base's one company-specific claim argues that distinction is eroding as AI sensing and power move into edge form factors. Logged as a tracked Synthos call (Hold) as of 2026-08-04 at $283.63.

Single biggest risk: this is a recovery, not growth, and the multiple says otherwise. Revenue was $18.344 billion in 2021, $20.028 billion in 2022, $17.519 billion in 2023, $15.641 billion in 2024 and $17.682 billion in 2025 — down 3.6% across five years, with a 21.9% peak-to-trough decline in the middle. Trailing revenue of $19.453 billion has still not reclaimed the 2022 peak. At 43.1x trailing earnings the price assumes that the current 22.8% growth rate is the beginning of a trend rather than the recovery half of a cycle. If revenue simply returns to the $20 billion area and stops — which is what the last five years would predict — then FY2027 consensus of $24.892 billion is 20% too high, EPS is nearer $8 than $10.15, and at 25x that is $200. That is the $203 bear case, and it requires nothing to go wrong operationally.

Most fragile assumption in the price: that gross margin returns toward the high sixties. The FY2027 consensus EPS of $10.155 on revenue of $24.892 billion implies a net margin near 37% against 31.1% trailing, which on this cost structure needs gross margin around 65-67% — most of the way back to the 68.8% of 2022. The June quarter delivered 61.4%, up from 55.9% two quarters earlier, so the direction is right and the distance remaining is large. It depends entirely on volume absorbing three newly-built 300mm fabs. If gross margin plateaus at 62-63%, FY2027 EPS is nearer $8.75 than $10.15, and at 28x that is $245 — 14% below today's price with every other number in this dive unchanged.

Provenance & disclosures

• Traceability: THREE tagged knowledge-base claims name Texas Instruments by entity, plus one text match. All three entity claims come from a SINGLE speaker — Jordi Visser — across a two-day window on 2026-04-25 and 2026-04-26 (two as speaker: jordi_visser, one as speaker: Jordi Visser, all speaker_role: independent, skill 2.0). Breadth is one. There is no second voice in the entity lane at all, which makes this the thinnest lane in this batch alongside Arista. Two of the three are sector-basket claims listing TXN among five or six semiconductor names; only the 2026-04-26 claim — reading the company as "an industrial-demand barometer" reporting improving industrial demand — is company-specific, and it is corroborated by the June quarter's disclosed growth "led by industrial, data center and automotive." The single most consequential claim about Texas Instruments in the entire knowledge base is in the TEXT lane rather than the entity lane, carries a named independent speaker (Dylan Patel, 2026-02-05), and is bearish: that China "has built a real semiconductor industry, insourcing more, with microcontrollers and power chips now matching and undercutting" the company. We surface and weight it because microcontrollers and power chips are the company's two segments, because China is 21.4% of revenue, and because Texas Instruments is spending $7.5 billion to buy a microcontroller business. Stance on the entity lane: 3 bullish, 0 neutral, 0 bearish. Breadth 1, net conviction positive-low. No management voice appears. All quotes verbatim from stored claim text.
• Data as-of: income statement, balance-sheet liquidity, capital-expenditure guidance, the Silicon Labs disclosure and the CHIPS Act detail for the quarter and six months ended 2026-06-30, from the Form 10-Q filed 2026-07-24; fiscal 2025 from the 10-K filed 2026-02-06 · estimates 2026-08-04 · prices 2026-08-04, quote timestamp 1785873600 = 2026-08-04T20:00:00Z ($283.63, +5.42%; 50-DMA $297.14; 200-DMA $227.56; RSI 39.3; MACD −6.06) · knowledge-base claims 2026-08-04. Texas Instruments' fiscal year is the calendar year. All figures come from the Synthos vendor data file for TXN or from the SEC filings in the TXN archive; no figure comes from memory, recall or external retrieval.
• Filing archive contents: 10-K filed 2026-02-06 (fiscal 2025, tables preserved); 10-Q filed 2026-04-24 (March 2026 quarter); 10-Q filed 2026-07-24 (June 2026 quarter — the primary source for this dive); 8-Ks filed 2026-04-22, 2026-06-02 and 2026-07-22. Tables were available on all substantive documents.
• Where the filings corrected or extended the vendor (detailed in Section 6): a $7.5 billion pending acquisition of Silicon Labs at $231.00 per share, expected to close in the first half of 2027, and a $5 billion 364-day delayed-draw term loan facility entered in June 2026, neither of which appears in ANY vendor field — the off-balance-sheet defect class, meaning post-closing net debt will be materially above the figure used here and no forward estimate can be assumed to include the acquired business; a buyback cut 81% by value and 83% by share count in the first half ($185 million on 0.9 million shares against $955 million on 5.4 million), invisible in a vendor file whose most recent repurchase figure is the FY2025 annual; capital-expenditure guidance of $2-3 billion for 2026 against $4.550 billion spent in 2025, with first-half spend down 51% — the central finding of this dive and unobtainable from the payload; netDebt corrected from $12.166 billion to $10.510 billion for the omission of $1.656 billion of short-term investments, and to approximately $9.6 billion on the 2026-06-30 liquidity position of $7.00 billion; two competing free-cash-flow definitions reconciled — the company's $6.5 billion adds back CHIPS Act proceeds, the conventional figure is $5.355 billion, and both are quoted and labelled; the CHIPS Act contribution quantified at a 35% investment tax credit plus up to $1.6 billion of grants, with $1.10 billion of cash proceeds and a $301 million tax benefit in the first half alone; seg_prod flagged for omitting the "Other" line in the two most recent years, a 5.5% shortfall; the FY2022 seg_geo row flagged for containing two identical values; and the FY2029 and FY2030 estimate rows excluded as non-monotonic. Where vendor and filing AGREED — worth recording: the June-quarter income statement matches the 10-Q exactly; the FY2025 and FY2024 seg_geo blocks tie to reported revenue to the dollar; and, notably, est.ebitAvg and est.ebitdaAvg PASS every arithmetic defect screen and reconcile to actuals within 0.4% and 2.0% for FY2025 and FY2024 — the cleanest estimate block in this batch, and against a 33.5% universe-wide defect rate that is a real finding.
• Basis note: Texas Instruments reports GAAP results only and publishes no adjusted earnings measure; the earn_cal actuals match GAAP diluted EPS. Trailing and forward multiples in this dive are therefore all GAAP and no basis reconciliation is required — the only name in this batch for which that is true. The one non-GAAP measure the company does use is free cash flow, which it explicitly defines as including CHIPS Act proceeds, and both definitions are reported separately above.
• Estimate coverage: 23 analysts on FY2026 revenue and 20 on FY2026 EPS; 25 and 22 on FY2027; 25 and 18 on FY2028 — the best coverage in this batch. The FY2029 and FY2030 rows carry 10 and 8 EPS analysts and are excluded, both for coverage and because FY2029 EPS DECLINES while revenue rises. The FY2028 EPS dispersion is extreme — mean $11.927, high $18.176, low $9.971, an 82% spread — reflecting genuine analyst disagreement rather than a data error, and it is disclosed rather than smoothed.
• Peer note: the vendor peer set is appropriate — Analog Devices, Broadcom, Intel, Lattice, Microchip, Monolithic Power, Marvell, Micron, NXP and ON Semiconductor. The genuine analog and embedded comparables are Analog Devices ($185.23B), Microchip ($43.80B), NXP ($59.88B) and ON Semiconductor. No peer multiples are supplied in the file, so no peer-multiple comparison is drawn. Texas Instruments' $259.02 billion market capitalisation is 1.40x Analog Devices'.
• Fair-value caveat: the $203 / $295 / $380 anchors are multiples of the FY2027 consensus EPS of $10.155 — 20.0x, 29.1x and 37.4x — cross-checked against FY2026E and FY2028E. Stated arithmetic, not a discounted cash flow. Sensitivity: 25x gives $254, 32x gives $325 — exactly the street's target. The bear case requires no earnings miss at all, only a return to a 20x forward multiple; the bull case REQUIRES multiple expansion from an already-full level and that is named as its weak leg. No case incorporates the Silicon Labs acquisition, in either direction, because neither the transaction economics nor the associated debt appears in any forward estimate in this file.
• Timing: second-quarter 2026 results were released 2026-07-22, thirteen days before this dive, beating consensus EPS by 12.0%, and the 10-Q was filed 2026-07-24. The next print is 2026-10-27, 84 days away, with consensus of $2.42 EPS on $5.916 billion of revenue — implying 8.3% sequential revenue growth on a quarter that already grew 13%.
• Accessibility note: no information in this dive is conveyed by colour. All emphasis is carried by bold text, table structure and explicit labelling.
• Not investment advice. Independent research, educational and informational only, never personalised. No recommendation to buy, sell or hold any security is made to any person.
• Version: 2026-08-04-full.