TI's capex supercycle is ending and FCF is set to ramp toward $8+/share — but at $324, you're paying for the recovery before it arrives.
TI spent three years building fabs nobody else would build. Now capex falls, the Sherman fab is running, CHIPS Act subsidies are flowing, and FCF is guided to $8+/share. At $324 — 14% above Street consensus and probability-weighted fair value ~$341 — this is a hold, not a buy. The number that changes the setup is below $290.
Full thesis
Texas Instruments' three-year capex supercycle was a bet that industrial and automotive analog demand would recover and that building CHIPS Act-eligible U.S. fabs would create long-term cost advantages. The bet is paying off: the Sherman fab is running, CHIPS Act cash is flowing, and FCF is set to ramp from trough to $8+/share. The problem is the market already knows this — at $324 and 35x forward P/E, you're paying for the recovery before it arrives. The re-entry zone below $290 offers asymmetric risk/reward; at current prices it's a hold with margin of safety for existing holders.
Texas Instruments: The Factory Owner's Inflection
For three years, TI ran an unpopular playbook — burn $5B a year building domestic 300mm fabs while fabless peers printed free cash flow and got bid up on AI narratives. The capex cycle just ended. The FCF is turning on. Data center analog demand nobody was modeling is up 90% year-over-year. And now the stock has doubled off the lows. Running the math before the next move.
The Bottom Line
Three things determine where TXN goes from here: does the FCF guidance hold without CHIPS Act timing games, does the industrial recovery sustain through 2H 2026, and does the data center analog story get a real multiple re-rate. Get all three right and the bull case closes. Miss on industrial — TI's biggest segment at 37% of revenue — and the bear case happens fast regardless of the fab narrative.
Where We've Been
TXN spent most of 2025 trading like a sleepy industrial semi company nobody wanted to own. While the market chased AMD, NVDA, and AI software names, TXN sat at $153 — its 52-week low — absorbing the combined weight of elevated capex, compressed FCF, and an industrial end market that hadn't recovered. Then three things hit at once.
TXN share price · Jan 2025 → May 2026
The catalysts that mattered
| Date | Catalyst | Stock Reaction |
|---|---|---|
| Oct 2025 | Q3 2025 earnings: industrial showed first genuine recovery signs. Data center acknowledged as a growth vector for the first time with specificity. | +8% on the week |
| Dec 20, 2025 | Sherman SM1 fab begins production. TI announces "geopolitically dependable" supply chain narrative starts to land with institutional buyers. | +5% on the news |
| Jan 28, 2026 | Q4 2025 beat: data center revenue at ~$1.2B annual run rate, up 50%+ YoY. Management guides Q1 above consensus. Stock jumps 7% in a single session. | +7% on the day |
| Feb 4, 2026 | Silicon Labs acquisition announced: $7.5B all-cash. Adds wireless connectivity, IoT, edge AI. Stock initially flat; rallied as strategic rationale — edge AI platform — became clearer. | +4% over two weeks |
| Apr 22, 2026 | Q1 2026 blowout: revenue $4.83B (+19% YoY), data center +90% YoY, industrial +30% YoY. EPS $1.68, beat consensus by 23.5%. FCF TTM $4.35B, up 154% YoY. | +18% in two sessions |
| May 6, 2026 | BofA upgrades to Buy ($320 PT). Barclays upgrades to Equalweight from Underweight. Multiple target raises across the Street. | +5% on upgrades |
| May 26, 2026 | Stock at $324, within $2 of all-time high $326.42. 14% above Street consensus target of $284. | Near all-time high |
Q1 2026: The Print That Changed Minds
Revenue of $4.83B on 19% year-over-year growth. Data center up 90% YoY. Industrial up 30% YoY. EPS $1.68 against a consensus of $1.36 — a 23.5% beat. That's not a modest outperformance; it's a regime change in how the Street models TI's end markets.
End market breakdown
| End Market | Q1 2026 YoY | QoQ |
|---|---|---|
| Data Center | +90% | +25% |
| Industrial | +30%+ | +20%+ |
| Automotive | +mid-single digits | ~flat |
| Personal Electronics | recovering | modest |
| Total Revenue | $4,830M (+19%) | +9% |
| GAAP EPS | $1.68 | +23.5% vs est. |
| FCF (TTM) | $4,350M | +154% YoY |
CHIPS Act incentives: $555M received in Q1 (Sherman production start triggered payment). TTM FCF includes $965M total CHIPS Act benefits.
What management said
CFO Rafael Lizardi said data center is now at an "annual revenue run rate approaching $1.5 billion," growing rapidly and broad-based across hyperscalers. He noted that industrial growth was "broad across all sectors and regions — not concentrated in one geography or application."
Q2 2026 guidance: revenue $5.0B–$5.4B, EPS $1.77–$2.05. That's +15-25% YoY revenue growth and ahead of prior consensus on both lines.
On the capex shift: "2026 capital expenditures expected to be $2–3 billion" — down from approximately $5B in 2025. That single sentence is the FCF inflection thesis in eight words. Capex falls by $2-3B; everything else flows to free cash.
FCF per share target reiterated: "at least $8 per share in 2026." At $324/share, that's a 2.5% FCF yield. At the 52-week low of $153, that same $8 FCF target would have been a 5.2% FCF yield — the kind of setup that shows why the stock had to double.
$725 Billion of Power Management Demand
TI doesn't make AI chips. It makes the components that make AI chips work. Every GPU rack in every hyperscaler data center requires power management, voltage regulation, signal conversion, and thermal control — and most of that content comes from TI's analog and embedded catalog. As hyperscaler capex doubles, TI's addressable opportunity expands with it.
Hyperscaler capex commitments · 2024A → 2027E
TI's exposure across hyperscaler buildouts
| Company | 2026 Capex | % AI-Related | TI Analog Exposure |
|---|---|---|---|
| Amazon (AWS) | $200B | ~70% | TI power management across all AWS compute and networking families; Graviton, Trainium, and Inferentia systems all carry TI content |
| Microsoft (Azure) | $190B | ~75% | TI is specified in Azure rack power rails; $25B of Microsoft's capex flagged as memory cost — HBM modules carry TI power ICs |
| Alphabet (Google) | $185B | ~70% | TI content in Google AI server power systems; TPU clusters require high-efficiency VRMs where TI has strong position |
| Meta | $135B | ~80% | TI analog throughout Meta's Grand Teton and Arcturus AI systems; no single-brand deal, but systematic volume at scale |
| Oracle / OCI | $50B | ~85% | OpenAI Stargate partner; TI power management in Stargate cluster rack designs |
Seaport Research called out TI specifically in a May 2026 note: "Rising data center power needs could boost Texas Instruments materially — as AI accelerator wattages increase, the power management complexity (and TI content) per rack increases with it." That's the insight in one sentence. A 200W GPU becomes a 700W GPU; the voltage regulation challenge goes up non-linearly, and TI is the go-to supplier for that complexity.
The Silicon Labs acquisition adds another dimension to this. Silicon Labs' Series 3 platform delivers 100 AI GOPS with a 10x computing uplift — designed for intelligent edge devices in factories, vehicles, and infrastructure. TI isn't just capturing hyperscaler capex; it's positioning for the distributed AI buildout that follows the centralized one. The edge AI market could be as large as the data center market by 2030 by some estimates, and it runs on the same TI analog building blocks.
The Moat Everyone Forgot to Model
While the semiconductor industry spent a decade going fabless — offloading manufacturing risk to TSMC so they could asset-light their way to higher multiples — Texas Instruments did the opposite. They kept their fabs. Then they built more. It looked like a capital allocation mistake for three years. Now it looks like a moat.
What TI actually built
Texas Instruments is one of the last true Integrated Device Manufacturers (IDMs) at scale. AMD, Qualcomm, Apple, Broadcom — all fabless, all on TSMC, all exposed to the same geopolitical tail risk. TI makes 95%+ of its chips internally by 2030. The buildout:
- RFAB2 (Richardson, TX): New 300mm fab alongside the legacy RFAB1. Transfers from 150mm complete. Ramping toward full build-out; will more than double RFAB1 capacity.
- LFAB1 (Lehi, UT): 300mm, adding new products on 45–65nm process nodes. 28nm qualification underway. Targets products currently at external foundries.
- Sherman SM1 (Sherman, TX): Clean room complete. Production underway as of December 2025. First volume output triggered $555M CHIPS Act payment in Q1 2026.
- Sherman SM2 (Sherman, TX): Shell complete, construction paused pending demand ramp. Reduces future construction lead time from years to months.
By 2030: 95%+ of wafers internally sourced, 80%+ on 300mm. Total US chip manufacturing investment: $60 billion — one of the largest in American semiconductor history.
The CHIPS Act economics
This is where the financial engineering of TI's strategy becomes visible. TI secured a $1.6 billion CHIPS Act direct funding award covering three fabs in Texas and Utah. Beyond the grants:
- 35% Investment Tax Credit (ITC) on qualifying semiconductor capex — raised from 25% under the original act. That's 35 cents back on every dollar of capex spent in the US.
- $555M received in Q1 2026 alone, triggered by Sherman SM1 hitting production milestones.
- ~$670M in 2025 from ITC and direct grant payments.
- $965M of the TTM FCF figure came from CHIPS Act benefits — which is why you need to look at normalized FCF, not just the headline number.
The effective cost of TI's US capex, after ITC and grants, is materially below the headline $5B/year. Fabless companies building equivalent capacity through TSMC don't get this subsidy. Over the 10-year horizon of these fabs, the ITC alone represents billions of dollars of structural cost advantage.
Goldman's depreciation concern — real but incomplete
Goldman Sachs double-downgraded TXN to Sell in late 2025 with a $156 price target. Their core thesis: depreciation of ~$2.3–2.7B in FY2026 compresses reported EPS, the company's execution through the cycle has been "lackluster," and TI has "less leverage than peers in the AI upcycle."
Goldman isn't wrong that depreciation is a headwind. It's a real non-cash charge that flows through the income statement and compresses reported EPS relative to cash flow. But the thesis is incomplete for three reasons:
- Depreciation is on sunk capital. The $5B/year has been spent. The cash is gone. What's left is the accounting treatment. The actual cash generation — measured by FCF — improves dramatically as new capex falls to $2-3B.
- The $8/share FCF target is a cash-basis metric. TI is explicitly guiding investors to look at FCF per share, not EPS. The $8 figure doesn't get compressed by depreciation accounting.
- Goldman's $156 target implies 19x FCF at $8/share. For a 21-year dividend grower with owned fabs, domestic supply security, and multi-decade customer relationships, 19x FCF is a liquidation multiple. The stock was at $153 at the 52-week low and didn't stay there long.
The depreciation headwind is real and will pressure EPS through 2026-2027. But investors who focus on EPS will underestimate TI, and investors who focus on FCF might overpay. The right framework is FCF normalized for CHIPS Act timing — probably $6-7/share in 2026 ex-grants, growing to $10+ by 2027-28 as capex stabilizes.
The geopolitical moat nobody is pricing
TI's fabs are in Texas and Utah. AMD's are in Taiwan. In a world where tariffs are policy, reshoring is bipartisan, and "geopolitically dependable supply chains" are a procurement requirement for US defense and government customers — owning domestic analog manufacturing capacity is a structural competitive advantage. It doesn't show up in quarterly EPS. It shows up in 10-year contracts and sole-source designations. That's exactly how analog supply chains work.
When Memory Goes Up, TI Wins (Unlike AMD)
The memory supercycle that squeezed AMD's gross margins on every AI accelerator shipped tells a different story for TI. DRAM prices up 171% YoY. HBM now consuming 23% of all DRAM wafer capacity. Micron, Samsung, and SK Hynix expanding HBM production lines as fast as they can. Each of those dynamics is a demand driver for TI's power management portfolio, not a cost headwind.
DRAM & HBM price index · Jan 2024 = 100
Three ways the memory supercycle helps TI
1. Every HBM stack needs precision power management — that's TI. High-bandwidth memory operates at extremely tight voltage tolerances. Each HBM module requires voltage regulators, power sequencers, and monitoring ICs that TI specializes in. More HBM per server (the trend is toward more stacks, not fewer) means more TI content per server.
2. Memory manufacturers are spending billions on HBM capacity expansion. Micron is spending $20B+ on new fabs. Samsung and SK Hynix are ramping similarly. Every new HBM fab line requires manufacturing equipment with analog control systems, sensors, and power delivery — much of it from TI's industrial catalog. Memory fab capex is TI industrial demand.
3. The shift from commodity DRAM to HBM increases TI dollar content per module. Commodity DDR4 had relatively simple power requirements. HBM with its 3D stacking, tight thermal management, and high-speed interface demands a much more sophisticated analog ecosystem. The move up the memory complexity ladder expands TI's TAM per unit shipped.
The contrast with AMD
AMD's Q1 2026 data point — 53% gross margin that management wants to hold at ~56% in Q2 — reflects the HBM cost pass-through challenge on MI-series accelerators. An MI350X carries ~192GB of HBM3E at roughly $35-40K in memory cost alone. That's a bill-of-materials problem for AMD, not a revenue opportunity.
For TI, the same HBM stack that squeezes AMD's BOM is a unit of demand for TI's power management portfolio. The dynamics are opposite, and most investors covering AI semiconductors have AMD's framing in their heads when they think about memory — not TI's.
Net read
The memory supercycle is a modest but genuine TI tailwind. The bigger story is the structural shift: as AI compute wattages increase — H100 at 700W, B200 at 1,200W, Vera Rubin likely higher — the power management complexity and dollar content per rack increases with it. TI's addressable opportunity expands with every generation of higher-wattage AI accelerator, regardless of who makes the GPU.
Revenue Trajectory & The FCF Machine Turning On
The TXN thesis in one picture: capex ran at $5B/year for three years, FCF got crushed to $1.5/share, and the Street stopped caring. Now capex falls by half, revenue grows, and FCF should more than quintuple from the 2024 trough. That's the setup — if the numbers come in.
TXN revenue by end market · 2023A → 2028E
Capital expenditure vs. free cash flow per share · 2020A → 2027E
| Year | Revenue | YoY | GAAP EPS | FCF/Share | Notes |
|---|---|---|---|---|---|
| FY 2022A | $20.0B | +9% | $9.41 | $6.40 | Peak cycle; pre-capex ramp |
| FY 2023A | $17.5B | −13% | $7.07 | $1.47 | Capex peaked; revenue fell — worst of both |
| FY 2024A | $15.6B | −11% | $5.20 | $1.63 | Revenue trough; capex still $5B |
| FY 2025A | $17.7B | +13% | $5.45 | $2.85 | Recovery begins; CHIPS Act $670M |
| FY 2026E | ~$21B | ~+19% | ~$7–8 | $8.0+ | Capex falls to $2–3B; CHIPS Act $555M Q1 alone |
| FY 2027E | ~$23B | ~+10% | ~$9–10 | $10+ | Silicon Labs closes; depreciation begins rolling off |
| FY 2028E | ~$26B | ~+13% | ~$12+ | $12–13 | 300mm scale economy fully realized; margins recover |
Sources: stockanalysis.com historical financials; Q1 2026 earnings release and guidance; JPMorgan, UBS, BofA estimates for FY26-28; management FCF target of $8+/share for FY2026. Silicon Labs contribution excluded from FY2027 until close confirmed.
The February 2026 acquisition of Silicon Labs adds roughly $1.8B of annual revenue, edge AI processing capability (Silicon Labs Series 3: 100 AI GOPS, 10x computing uplift), and a wireless connectivity platform spanning IoT, smart home, industrial automation, and automotive. Expected synergies: ~$450M annually within three years post-close. EPS accretive in the first full year. The integration challenge — migrating a fabless company's product lines into TI's manufacturing model — is real but manageable given TI's fab capacity headroom. Closing expected 1H 2027.
SWOT
Where TI wins, where it bleeds, where the upside lives, what to watch.
Strengths
- Owned 300mm fabs in Texas and Utah — domestic manufacturing moat fabless competitors cannot replicate in any near-term timeline
- Data center +90% YoY in Q1 2026 — a growth rate nobody modeled 18 months ago, and it's accelerating
- Industrial recovery broad-based — +30% YoY in Q1 across all sectors and regions; TI's largest segment at 37% of revenue
- $1.6B CHIPS Act award + 35% ITC — structural cost advantage vs. any TSMC-dependent peer; effectively subsidized domestic capacity
- 21-year dividend growth streak — $5.68/yr, growing; Dividend Aristocrat designation approaching (25-year threshold in 2030)
- FCF inflection — from $1.6/share (2024 trough) to $8+ guided (2026); capex cliff is the catalyst
- Analog design-in stickiness — once TI is specified into a power rail, it stays for 10–15 years; switching costs are prohibitive
- Silicon Labs — edge AI, IoT, wireless connectivity added to TI's manufacturing and sales platform
Weaknesses
- No "AI chip" narrative — TI is infrastructure for AI, not the AI chip itself; doesn't get the multiple re-rate that GPU companies do
- Depreciation headwind $2.3–2.7B in FY2026 — Goldman's legitimate concern; compresses reported EPS vs. FCF reality
- Silicon Labs adds leverage — $7.5B all-cash deal funded with debt raises financial risk at a moment of macro uncertainty
- Automotive growth muted — mid-single-digit YoY in Q1; EV transition creates demand uncertainty in TI's second-largest end market
- Gross margins compressed at 57% — down from 68% at the 2022 peak; reflects underutilization of new fab capacity
- CHIPS Act FCF is lumpy and non-recurring — the $8/share FCF target includes significant one-time incentive payments; normalized FCF is probably $6–7/share
- Stock ahead of consensus — $324 vs. $284 average PT means you're paying above what most sell-side analysts think it's worth
Opportunities
- FCF compounding — if FCF grows from $8/share (2026E) to $12+/share (2028E), a 35x multiple applied to $12 = $420 stock with no multiple expansion
- Margin recovery — as Sherman and LFAB1 utilization ramps, gross margins recover from 57% toward 65%+; operating leverage is dramatic
- Data center analog TAM — Seaport Research estimates rising AI accelerator wattages could drive TI data center revenue to $3–5B by 2028; that's not in consensus models
- Industrial edge AI — Silicon Labs + TI analog = infrastructure platform for factory automation, smart grid, and industrial AI; trillion-dollar TAM over a 10-year horizon
- Geopolitical reshoring — US government preference for domestically manufactured semiconductors in defense, infrastructure, and critical applications; TI is positioned; fabless competitors are not
- Dividend Aristocrat trajectory — institutional income mandates that automatically purchase Aristocrats would provide ongoing buying support
- 300mm cost leadership — by 2028-2030, TI should have the lowest cost-per-wafer of any analog producer at scale; structural margin tailwind
Threats
- Industrial demand durability — 30%+ YoY recovery is real but the base effect is favorable; risk is Q3/Q4 deceleration if macro softens or China competition accelerates in auto
- Goldman's EPS disappointment scenario — if reported EPS misses consensus due to depreciation, sentiment turns negative even if FCF is fine
- Silicon Labs integration complexity — migrating a fabless product portfolio to TI's manufacturing model takes years; execution risk is real if timelines slip
- Data center deceleration perception — 90% YoY growth won't sustain; when it normalizes to 40-50% YoY, some investors will call it a slowdown regardless of absolute levels
- CHIPS Act political risk — new administration or budget dynamics could alter the ITC rate or grant timing; low probability but material if realized
- China trade exposure — TI has some analog revenue exposed to China; tariff escalation or export restrictions could affect sales in specific product lines
- Automotive: EV demand uncertainty — if EV adoption trails forecasts, automotive semiconductor demand (21% of TI revenue) could disappoint through 2027
- Multiple compression risk — at 35x forward EPS, any earnings miss or macro shock compresses the multiple; not as stretched as AMD at 54x, but not cheap
Bull · Base · Bear
Twelve-month forward scenarios. Probabilities anchored to current analyst dispersion ($156 Goldman to $320 BofA) and the binary execution risks tied to the FCF inflection, industrial durability, and Silicon Labs integration.
$420
FCF prints $9–10/share in 2026, beating the $8 guide. Industrial sustains at 20%+ YoY growth through 2H. Data center re-rates to a recognized growth story — analysts build dedicated models for TI data center instead of burying it in "enterprise systems." Silicon Labs deal closes smoothly. Gross margin recovery visible in Q3/Q4 as fab utilization improves. Multiple expands to 42x on FCF of $10+ = $420. The "domestic fab owner" narrative takes hold as geopolitical risk forces hyperscaler procurement teams to preference US-made analog over TSMC-dependent alternatives.
$355
FCF hits $8/share as guided. Revenue reaches $21B for full-year 2026. Industrial moderates to mid-teens YoY growth in 2H as base effects normalize. Data center continues rapid growth but without a narrative re-rate. Multiple holds at 40–44x forward FCF. Silicon Labs integration on track for 1H 2027 close. This is roughly what BofA ($320 PT), JPMorgan ($280), and UBS ($295) are pricing — the stock earns a modest premium to their targets if the FCF guide holds. $355 reflects $8/share FCF at 44x multiple.
$235
Goldman's thesis plays out. Depreciation bites harder than expected, pushing reported EPS below $7 and disappointing the EPS-focused consensus. Industrial demand stalls in 2H 2026 as macro softens — TI's largest segment decelerates. CHIPS Act payment timing creates a miss vs. FCF guidance. Silicon Labs faces regulatory delay. Multiple compresses to 28x EPS = ~$200, with a floor near $235 where FCF yield becomes compelling again. Goldman's $156 PT requires every bad thing to happen simultaneously and isn't a realistic base — but $235 is a credible bear case if two of three risks materialize.
Price scenarios · May 2026 → May 2027
Time-Horizon Outlook
What to watch across the next 3 months, the rest of 2026, 2027, and the longer arc.
Jun–Aug 2026
Stock has absorbed Q1 and the round of upgrades. Sentiment is constructive but the stock has run hard. Near-term is data-dependent.
- Q2 earnings (late July) — $5.0–5.4B guide is achievable; beat requires data center to sustain and industrial not to decelerate
- CHIPS Act Q2 payment — another $200–300M FCF contribution possible; timing matters for guidance validation
- Silicon Labs shareholder vote expected
- Risk: any commentary on industrial demand softening in 2H triggers a 10–15% pullback regardless of Q2 result
- Watch: automotive — mid-single-digit growth is fine but flat QoQ for two more quarters starts to look structural, not cyclical
Sep–Dec 2026
The capex cliff is the story. Q3 and Q4 should show FCF step-up as the full year's reduced capex flows through. Q4 guide sets 2027 direction.
- Capex cliff in full effect: $2–3B vs $5B prior year — FCF should step up materially in Q3/Q4
- Depreciation peak: 2026 is likely peak depreciation year before it starts rolling off in 2027
- Gross margin trajectory: first visible signs of margin recovery as fab utilization improves
- Q4 earnings + 2027 guide: likely sets the next 6 months of stock direction
- Risk: hyperscaler commentary on data center construction slowing could compress TI's fastest-growing segment expectations
The validation year
Silicon Labs closes. Depreciation starts declining. Margins expand. EPS growth should outpace revenue growth for the first time in years.
- Silicon Labs closes (1H 2027) — first year of accretion; $450M synergy clock starts
- 22nd consecutive year of dividend growth — raising the dividend maintains the Aristocrat trajectory
- Depreciation begins declining vs. 2026 peak — EPS gets less headwind each quarter
- Data center contribution could reach $3B+ if hyperscaler capex sustains through 2027
- Risk: capex digestion at hyperscalers — "we built enough compute for now" commentary hurts TI's fastest-growing segment
- Risk: Silicon Labs integration delays or regulatory friction extends the deal timeline into 2028
The long arc
If the 300mm buildout works the way TI's management has described it for the past three years, 2028 is when the economics become undeniable.
- 95% internal wafer sourcing achieved — structural cost leadership in analog established
- Gross margins recovering toward 65%+ from the 57% trough
- FCF potentially $12–14/share — at 30x multiple, that's $360–420 stock with no revenue growth assumption baked in beyond what's already contracted
- Silicon Labs fully integrated — IoT/edge AI becomes a recognized third growth leg
- Dividend Aristocrat milestone (25-year streak) in sight by 2030 — institutional income demand expands
- Tail risk: AI capex cycle peaks, hyperscalers stop building — data center analog demand normalizes to GDP-ish growth
Risk Matrix
Ranked by what would actually move the stock 15%+ in either direction over the next twelve months.
What the Street Says
~37 analysts covering. Broad Buy consensus with an average target of ~$284 — 14% below current. The stock has run ahead of analyst models, meaning the burden of proof falls on the next two or three earnings prints. Wide dispersion from Goldman's $156 to BofA's $320 tells you the outcome isn't settled.
| Firm | Rating | Price Target | Date | Key Takeaway |
|---|---|---|---|---|
| Goldman Sachs | Sell (double downgrade) | $156 | Dec 2025 | Depreciation burden ($2.3–2.7B in FY26), "lackluster" cycle execution, limited AI leverage vs. peers — the bear thesis in full |
| Bank of America | Buy (upgraded) | $320 | May 2026 | Post-Q1 upgrade from Neutral; industrial + data center + FCF inflection thesis; raised PT from $235 |
| Barclays | Equalweight (upgraded) | $250 | May 2026 | Upgraded from Underweight; flagged data center +90% YoY and industrial +30% YoY as thesis-changers; still cautious on multiple |
| JPMorgan | Overweight | $280 | Apr 2026 | Raised from $227; FCF recovery on track; fab strategy underappreciated; cautious on depreciation timing |
| UBS | Buy | $295 | May 2026 | Raised from $260; data center exposure underappreciated in models; 300mm cost advantage emerging |
| Susquehanna | Positive | $300 | 2026 | Raised from $275; domestic fab strategy as structural moat; CHIPS Act FCF support validates capital allocation |
| Seaport Research | Buy | ~$320 | May 2026 | "Rising data center power needs could boost TXN materially" — most bullish on the analog data center content opportunity |
| TD Cowen / Baird / Stifel | Buy / Outperform | $270–$290 | Post-Q1 | Multiple firms raised targets post-Q1 to $270–290 range; constructive on industrial recovery, cautious on pace of data center normalization |
The read: The dispersion — $156 to $320, a 105% spread — tells you the binary outcomes are real. Goldman's depreciation thesis and the bulls' FCF inflection thesis can't both be right simultaneously. The stock at $324 is actually above BofA's $320 bull target, which means the market is pricing in something even more optimistic than the Street's most bullish model. That's typically where you hold what you have, don't add, and wait for the numbers to confirm or deny the thesis.
One notable data point: Goldman's $156 target was issued in December 2025 when the stock was trading in the $200s. The stock subsequently rallied 60%+ past it. That doesn't mean Goldman is wrong about the depreciation math — it means the market decided the FCF story was more important than the EPS story. Whether that was the right call gets answered by the next three earnings prints.
Translating the Scenarios to Math
The framing below is an analytical exercise, not a personalized recommendation. It walks through how the bull/base/bear scenarios map onto common position-sizing frameworks. None of it is tailored to any specific reader's financial situation, risk tolerance, time horizon, or tax position. Anyone applying this kind of math to real capital should do their own work and consult their own advisors.
The expected-value math
Start with the scenarios: bull $420 at 25% probability, base $355 at 50%, bear $235 at 25%. Probability-weighted fair value is $341 against a $324 current price — roughly +5% over twelve months. That's an expected return that barely covers the dividend yield, let alone the risk of holding a single-name semiconductor with a binary FCF story.
The bar for adding here isn't "is TI a good company." It clearly is. The question is whether the current price adequately compensates for the risks. At $324, the probability-weighted math says it roughly does — which is to say, the market has already done a reasonable job pricing the distribution. That's not what a "buy here" setup looks like. It's what a hold-and-reassess setup looks like.
How the math shifts at different entry prices
| Hypothetical Entry | Bull Upside | Bear Downside | Prob-Weighted Return | Up/Down Ratio |
|---|---|---|---|---|
| $324 (current) | +30% | −27% | +5% | 1.1 : 1 |
| $290 | +45% | −19% | +18% | 2.4 : 1 |
| $260 | +62% | −10% | +29% | 6.2 : 1 |
| $235 (bear floor) | +79% | 0% | +45% | ∞ |
The math that made TXN worth buying at $153 (the 52-week low) is not the same math that applies at $324. At $153, $8/share FCF implied a 5.2% FCF yield on a 21-year dividend grower with owned fabs and a CHIPS Act subsidy. That was obvious. At $324, the same $8/share FCF is a 2.5% yield — not bad, but not obviously cheap for a semi at an inflection with execution risk.
The "quality compounder" framing
TI is a different risk profile than AMD. AMD at 54x forward P/E is a pure execution bet on a product cycle. TI at 35x forward P/E is a slower, more durable business with owned assets, multi-decade customer relationships, and a structural cost advantage coming online. The right position size for TI isn't the same as for AMD — it's a different bucket. Larger, more stable, more appropriate for a core holding than a speculative one. But even quality compounders need to be bought at prices that make the math work.
What changes the setup
Three things would make TI a clear add rather than a hold: (1) a Q2 or Q3 miss that sends the stock below $290, creating a much better up/down ratio; (2) a Goldman downgrade reversal or explicit acknowledgment that the FCF thesis is correct, which could re-rate consensus targets above $320; or (3) a data center analog announcement — a specific customer win, sole-source designation, or analyst note building a dedicated data center TAM model — that changes how the Street values TI's fastest-growing segment. Any one of those would change the math. None of them is guaranteed to happen in the next 90 days.
Glossary of Abbreviations
| YTD | Year to Date — return since January 1 of the current year |
| Fwd P/E | Forward Price-to-Earnings — share price divided by next twelve months' consensus EPS estimate |
| EV/EBITDA | Enterprise Value divided by EBITDA — the most common cross-capital-structure valuation multiple |
| EBITDA | Earnings Before Interest, Taxes, Depreciation, and Amortization — a proxy for operating cash profitability |
| EPS | Earnings Per Share — net income divided by diluted shares outstanding |
| FCF | Free Cash Flow — operating cash flow minus capital expenditures; TI's stated long-term goal is to grow FCF per share at >10% annually |
| capex | Capital Expenditure — TI's 300mm fab investment program is the defining capital allocation story; capex peaked ~$5B in 2023 and is guiding to $2-3B in 2026 |
| ATH | All-Time High — the highest price a stock has ever traded |
| analog | Semiconductor products that process continuous real-world signals (voltage, temperature, current); TI is the world's largest analog chip company by revenue |
| ADAS | Advanced Driver Assistance Systems — automotive safety features (collision avoidance, lane keep, adaptive cruise); one of TI's fastest-growing end markets |
| embedded | Embedded processors — TI's second product category after analog; microcontrollers and processors for real-time industrial and automotive control |
| power management | ICs that regulate, convert, and distribute electrical power in electronic systems; TI's largest analog sub-segment and the primary AI data center opportunity |
| RFAB2 | Richardson Fab 2 — TI's 300mm manufacturing facility in Richardson, TX; part of TI's multi-billion dollar internal fab investment strategy |
| 300mm | 300-millimeter wafer diameter — current industry standard for leading-edge manufacturing; larger wafers yield more chips per run; TI is building out 300mm capacity as a structural cost advantage |
| IDM | Integrated Device Manufacturer — a company that both designs and manufactures its own chips; TI's IDM model gives it cost control that fabless analog competitors cannot match |
| CHIPS Act | U.S. CHIPS and Science Act (2022) — TI received a $1.6B grant for domestic 300mm fab construction in Texas and Utah |
| destocking | The process by which customers reduce excess inventory built up during supply chain disruptions; TI's 2023-2024 revenue decline was primarily driven by customer destocking |
| inventory | Channel inventory — the stock of chips held by distributors and customers; high channel inventory suppresses orders and ASPs; TI tracks this closely as a cycle indicator |
| EV | Electric Vehicle — a major end market for TI's analog and embedded products; EVs use 2-3× the semiconductor content of ICE vehicles |
| AI | Artificial Intelligence — driving data center power management demand; each AI server rack requires significantly more power conversion chips than a standard server |
| hyperscaler | Large cloud and AI infrastructure providers (Microsoft, Google, Amazon, Meta); expanding AI data center buildout is creating a new demand wave for TI power management ICs |
| TTM | Trailing Twelve Months — the most recent four reported quarters of financial data |