Diversified global power company: regulated utilities + renewable energy developer + energy infrastructure. The cheapest data center power play you can find. | Analysis date: 2026-03-12
AES scored AGI 8 in our screening -- high for a utility. It trades at P/TB 2.49 with a $10.1B market cap, making it one of the cheapest AGI-exposed utilities. The interesting angle: AES is not a pure regulated utility like PNW or OGE. It is a hybrid -- part regulated utility (AES Indiana, AES Ohio, El Salvador), part renewable energy developer (13.2 GW operating, 11.9 GW backlog), part energy infrastructure (gas plants, LNG, Chile). That diversification has been punished by the market (stock down 50%+ from 2022 highs), but could be a strength if data center power demand plays out.
The question: Is the diversification a feature or a bug? Is $31B of debt a time bomb or a manageable cost of infrastructure buildout?
AES operates through four segments (SBUs), which is the key to understanding the investment case. This is not a simple utility.
| SBU | What It Does | FY2024 Revenue | FY2024 Adj. EBITDA | % of EBITDA | Character |
|---|---|---|---|---|---|
| Renewables | Solar, wind, energy storage, hydro globally. 13.2 GW operating. | $2,510M | $552M | 21% | Growth engine. Data center PPAs. |
| Utilities | AES Indiana (integrated), AES Ohio (T&D), El Salvador (4 utilities). 2.7M customers. | $3,608M | $792M | 30% | Regulated. Stable. 9.9% ROE allowed. Data center load growth in territory. |
| Energy Infrastructure | Natural gas, LNG, coal/pet coke plants. Chile operations. Merchant + contract. | $6,238M | $1,366M | 51% | Cash cow. Includes declining coal assets being exited. |
| New Energy Technologies | Fluence (energy storage tech), Uplight (energy management), Maximo | $1M | -$38M | -1% | Venture-stage. Call option. |
| Corporate & Eliminations | -$79M | -$33M | |||
| Total | $12,278M | $2,639M | 100% |
Adjusted EBITDA was $2,639M in FY2024. But Adjusted EBITDA with Tax Attributes was $3,952M -- a $1,313M uplift from IRA tax credits on renewables. The Renewables SBU earned $552M in Adj. EBITDA but $1,845M when tax attributes are included. This makes renewables the true #1 earner, not Energy Infrastructure. The IRA is worth ~$1.3B/year to AES. If it gets repealed or modified, that is a massive hit.
| Fuel/Technology | % of Capacity | Notes |
|---|---|---|
| Renewables (solar, hydro, wind, storage, landfill gas) | 50% | ~16,000 MW. Fastest-growing segment. |
| Natural Gas | 32% | ~10,300 MW. LNG in Dominican Republic, Panama. Local gas elsewhere. |
| Coal | 16% | ~5,100 MW. Exiting. 13.4 GW coal exits since 2017. |
| Pet Coke / Oil | 2% | ~640 MW. Minor. |
| Total Operating Capacity | 100% | ~32,109 MW operating |
AES explicitly projects data center U.S. demand growing from ~30 GW today to 90 GW by 2030 (a 60 GW increase). They are targeting a significant share of this growth through both their renewables platform (selling clean power via PPAs) and their utility territories (load growth in Indiana and Ohio).
| Balance Sheet Item | FY2025 | Notes |
|---|---|---|
| Total Assets | $51.8B | Massive physical asset base |
| PP&E (Net) | $37.8B | Power plants, T&D infrastructure, construction in progress ($7.9B) |
| Gross PP&E | $47.6B | Before $9.8B accumulated depreciation |
| Goodwill | $342M | Minimal -- good sign. Written down from $1.2B in 2021. |
| Intangibles | $2.0B | PPAs, concession rights, development assets |
| Total Equity (incl. minority interest) | $9.1B | |
| AES Stockholders' Equity | $4.1B | After minority interests of ~$5.0B |
| Tangible Book (AES only) | $1.7B | $4.1B - $342M goodwill - $2.0B intangibles |
| Total Debt | $31.0B | Per yfinance. ~$22.7B non-recourse, ~$5.7B recourse. |
| Cash | $1.6B | Plus $515M restricted cash |
| Net Debt | $28.5B | |
| Debt / Equity | 259% | Extremely high. But typical for capital-intensive utilities. |
| Operating Cash Flow | $4.3B | Comfortable debt service |
| Interest Expense | $1.4B | ~33% of OCF goes to interest |
$31B in total debt sounds terrifying, but the structure matters enormously:
Bankruptcy risk is LOW. The non-recourse structure means even worst-case project failures are ring-fenced. The parent has $5.7B recourse debt against $4.3B operating cash flow. AES has survived multiple crises (2001 Enron era, 2008 financial crisis, Argentina defaults) without going bankrupt. The company structure is designed for resilience.
Result: Proceed to next question.
Three things need to align:
Evidence this is working: 4.4 GW of new PPAs signed in 2024, 3.0 GW completed construction, 2.1 GW data center load at AES Ohio. The flywheel is turning.
| Scenario | Adj. EPS | P/E Multiple | Implied Price | Market Cap | Multiple vs Today |
|---|---|---|---|---|---|
| Base Case (moderate growth) | $3.50 | 18x | $63 | $45B | 4.4x |
| Bull Case (data center + re-rate) | $5.00 | 20x | $100 | $71B | 7.0x |
| Extreme Bull (full utility multiple) | $6.00 | 22x | $132 | $94B | 9.3x |
| Max Plausible | $7.00 | 22x | $154 | $110B | 10.8x |
Is $5-7 EPS plausible by ~2031?
Unlike BE at $44B market cap where 10x was mathematically impossible, AES at $10.1B market cap makes 10x mathematically plausible. It requires both earnings growth (from backlog conversion and utility rate base growth) AND multiple expansion (from 6x P/E to 18-22x P/E). The multiple expansion alone -- if the market simply valued AES at utility-peer multiples -- would be a 3x. Add in earnings growth and you approach 7-10x. This is the setup we look for: cheap on multiple AND growing earnings.
| Bear Point | Severity | Our Assessment |
|---|---|---|
| $31B debt load | High | Legitimate concern. But mostly non-recourse and well-covered by OCF. Interest coverage 3.1x. |
| IRA repeal/modification risk | High | $1.3B/year from tax attributes. Trump admin uncertainty. But AES has safe-harbored backlog projects. |
| International complexity (Argentina, Chile, El Salvador, etc.) | Medium | FX risk, political risk, drought risk. AES has operated internationally for 40+ years. Selling non-core assets ($3.5B target). |
| Negative FCF (-$1.6B in FY2025, -$4.6B in FY2024) | Medium | Capex-heavy growth phase. OCF is strong ($4.3B). Negative FCF = investing in growth assets. |
| Coal legacy | Medium | Still 16% coal. But actively exiting (13.4 GW retired since 2017). ESG screens exclude AES. |
| Earnings quality -- GAAP EPS swings wildly | Medium | $2.39 EPS in FY2024, $1.31 in FY2025. Impairments, FX, asset sales create noise. Adj. EPS is more stable ($2.14). |
| Conglomerate discount | Low | Market prefers pure-play utilities. AES is a mix of everything. Simplification via asset sales is underway. |
| Share dilution | Low | 666M shares (2021) to 712M (2025) = 7% dilution in 4 years. Not great, not terrible. |
The two biggest risks are debt and IRA dependency. Everything else is manageable or being addressed. If interest rates spike AND the IRA is repealed, AES would be in serious trouble. That dual-risk scenario is unlikely but not impossible.
| Floor Component | Value | Notes |
|---|---|---|
| AES Stockholders' Equity (book) | $4.1B | After minority interests |
| Less: Goodwill | -$342M | |
| Less: Intangibles | -$2.0B | PPA values, concessions -- some real value here |
| Tangible Book Value | $1.7B | = $2.36/share. Very low. |
| PP&E (Net) | $37.8B | Hard physical assets. Replacement cost likely higher. |
| Gross PP&E less Debt | $16.6B | $47.6B gross PP&E - $31B debt. Rough asset value after debt. |
| Trough Earnings Power (Adj. EPS ~$1.50 x 8x P/E) | $12/share | If everything goes wrong but company survives |
| Normalized Earnings Power ($2.14 EPS x 10x P/E) | $21/share | No growth, no re-rating |
| Dividend floor ($0.70/share at 7% yield) | $10/share | Yield support -- dividend has been maintained through crises |
The stock already hit $9.46 (52-week low in April 2025), which is near our trough floor. At $9.46, the market cap was ~$6.7B against $4.3B operating cash flow -- essentially a 1.6x OCF multiple for a company with 32 GW of power generation. That was panic pricing.
The dividend provides a soft floor. AES has maintained dividends through multiple crises. At $0.70/share, a 7% yield implies ~$10. The stock bounced off this zone.
Floor confidence: MODERATE. The non-recourse debt structure provides structural protection. The regulated utility earnings provide a stable base. But international exposure, IRA risk, and the sheer complexity of the business reduce confidence. This is not a simple utility with transparent earnings.
This is the core question: is AES's diversification a strength or a weakness? Let us compare to the pure-play regulated utilities in our AGI 9 universe.
| Metric | AES | PNW | OGE | D |
|---|---|---|---|---|
| AGI Score | 8 | 9 | 9 | 9 |
| Market Cap | $10.1B | $12.2B | $9.7B | $54.7B |
| Trailing P/E | 10.9x | 19.9x | 20.3x | 17.9x |
| Forward P/E | 5.96x | 17.9x | 18.2x | 16.3x |
| P/B | 2.49x | 1.73x | 1.95x | 1.95x |
| EV/EBITDA | 13.6x | 12.9x | 11.9x | 13.6x |
| Dividend Yield | 4.95% | 3.61% | 3.57% | 4.26% |
| Beta | 0.94 | 0.49 | 0.58 | 0.67 |
| Operating Margin | 16.6% | 11.0% | 18.7% | 22.0% |
| ROE | 1.95% | 9.1% | 9.8% | 9.7% |
| Debt/Equity | 259% | ~200% | ~114% | ~150% |
| Revenue | $12.2B | $5.3B | $3.3B | $4.1B |
AES trades at 5.96x forward P/E vs. 16-18x for pure-play utilities. If AES traded at even 12x forward earnings (still a discount), the stock would be ~$29 -- a 2x from current price. At 16x (peer average), it would be ~$38 -- nearly a 3x.
Why the discount?
Our take: the discount is partially warranted (complexity, international risk) but is overdone. AES has more data center exposure than any of these pure-play utilities -- 11.9 GW renewable backlog + 2.1 GW utility data center load -- yet trades at 1/3 the P/E. The market is pricing AES for risk but giving zero credit for growth.
| Metric | 2021 | 2022 | 2023 | 2024 | 2025 |
|---|---|---|---|---|---|
| Revenue ($B) | $2.8B | $3.1B | $12.7B | $12.3B | $12.2B |
| Gross Profit ($B) | $559M | $563M | $2,504M | $2,314M | $2,211M |
| Operating Income ($B) | -- | -- | -- | $2,026M | $1,970M |
| Net Income ($M) | -$1,330M | -$505M | -$182M | $802M | $900M |
| EPS (Diluted) | -$0.95 | -$0.82 | $0.34 | $2.37 | $1.31 |
| Adj. EPS | -- | -- | $1.76 | $2.14 | -- |
| Dividends/Share | $0.61 | $0.32 | $0.67 | $0.69 | $0.70 |
| Shares Out (M) | 666 | 668 | 669 | 706 | 712 |
Note on the revenue jump: Revenue jumped from ~$3B (2021-2022) to $12.7B (2023). This reflects a reporting change -- prior years may have used net reporting for certain pass-through energy costs. The underlying business growth is real but not a 4x revenue jump.
| Metric | 2021 | 2022 | 2023 | 2024 | 2025 |
|---|---|---|---|---|---|
| Operating Cash Flow ($B) | $1.9B | $2.7B | $3.0B | $2.8B | $4.3B |
| CapEx ($B) | -$2.1B | -$4.6B | -$7.7B | -$7.4B | -$5.9B |
| Free Cash Flow ($B) | -$0.2B | -$1.9B | -$4.7B | -$4.6B | -$1.6B |
| Interest Expense ($B) | $0.9B | $1.1B | $1.3B | $1.5B | $1.4B |
| Interest Coverage (OCF/Int) | 2.1x | 2.4x | 2.3x | 1.9x | 3.1x |
CapEx surged from $2.1B (2021) to $7.7B (2023) as AES ramped renewable construction. It dropped to $5.9B in FY2025 while OCF jumped to $4.3B. The gap is narrowing. If CapEx moderates to $4-5B/year while OCF stays at $4B+, AES could approach FCF positive in 2026-2027.
The negative FCF is not a sign of a broken business -- it is a sign of a company investing aggressively in growth. The question is whether the assets being built will generate adequate returns. Given they are largely under long-term PPAs with creditworthy counterparties (Amazon, Microsoft, utilities), the risk is manageable.
| Metric | 2021 | 2022 | 2023 | 2024 | 2025 |
|---|---|---|---|---|---|
| Total Assets ($B) | $33.0B | $38.4B | $44.8B | $47.4B | $51.8B |
| PP&E Net ($B) | $19.9B | $23.0B | $30.0B | $33.2B | $37.8B |
| Goodwill ($B) | $1.2B | $0.4B | $0.3B | $0.3B | $0.3B |
| Total Equity ($B) | $4.6B | $4.5B | $6.0B | $7.7B | $9.1B |
| Cash ($B) | $0.9B | $1.4B | $1.4B | $1.5B | $1.4B |
The key story: PP&E nearly doubled from $19.9B to $37.8B in 4 years. AES has been on a building spree. Total equity also doubled from $4.6B to $9.1B -- the growth is being funded by a mix of non-recourse project debt, tax equity, asset sales, and retained earnings. Goodwill dropped from $1.2B to $0.3B through impairments and dispositions -- the balance sheet is getting cleaner.
| Metric | Value |
|---|---|
| Annual Dividend Rate | $0.70/share |
| Dividend Yield | 4.95% |
| Payout Ratio (GAAP) | 54% |
| Payout Ratio (Adj. EPS $2.14) | 33% |
| Payout Ratio (OCF-based) | 12% ($501M div / $4.3B OCF) |
| Dividend Growth (2023-2025) | $0.67 -> $0.69 -> $0.70 (~2%/yr) |
At a 12% payout ratio on operating cash flow, the dividend has enormous room. AES pays $501M in dividends against $4.3B OCF. Even if OCF halved, the dividend would still be covered. The 4.95% yield provides income while you wait for the re-rating thesis to play out. Dividend growth has been modest (~2%/yr) because AES is prioritizing CapEx for growth -- this is the right capital allocation decision.
AES attacks the data center opportunity from two angles simultaneously, which is unique among our utility-sector holdings:
This dual approach is genuinely differentiated. PNW and OGE benefit from data center load growth in their territories (Angle 2 only). AES gets both the regulated utility upside AND the merchant/contract renewable PPA upside. The market is not giving AES credit for this.
From our AGI scoring analysis:
AGI creates enormous demand for exactly what AES sells. The risk is that AGI eventually solves the energy problem itself (better batteries, fusion), but the investment window is 5-10 years -- more than enough for AES to benefit.
| # | Risk | Probability | Impact | Mitigation |
|---|---|---|---|---|
| 1 | IRA repeal or significant modification | 25-40% | Very High ($1.3B/yr at stake) | Backlog projects are safe-harbored. Pipeline at risk. AES has secured panels/batteries for 2025-2026 projects. |
| 2 | Interest rate spike + debt refinancing risk | 15-25% | High | Most debt is fixed rate. Non-recourse structure limits contagion. Active hedging program. |
| 3 | Tariffs on solar panels/batteries | 30-50% | Medium | Supply chain already shifted outside China. U.S. domestic manufacturing secured for 2026-2030. |
| 4 | Argentina/LatAm macro crisis | 20-30% | Medium | AES has survived multiple Argentine crises. Selling non-core international assets ($3.5B target). |
| 5 | Construction execution risk | 15-20% | Medium | Track record of on-time delivery. But 4.9 GW under construction is a lot. Supply chain disruption possible. |
| 6 | Drought/hydrology at Colombia, Chile, Panama | 10-15% | Medium | Diversified across solar, wind, thermal. But Colombia drought cost $148M in FY2024. |
| 7 | Data center demand disappoints | 10-15% | Medium | Multiple independent forecasters project 60 GW growth. Contracts already signed. |
| 8 | Dividend cut | 5-10% | Medium | 12% payout on OCF. Would take catastrophic decline to trigger. |
| Period | Price Range | Context |
|---|---|---|
| 2021 | $19 - $23 | Clean energy enthusiasm. Biden infrastructure narrative. |
| 2022 | $17 - $25 | Interest rate hikes begin. Utility stocks wobble. |
| 2023 | $13 - $24 | Steep decline. Higher rates + impairments + LatAm concerns. |
| 2024 | $12 - $20 | Continued decline. Negative FCF, coal legacy, complexity discount. |
| 2025 (to date) | $9.46 - $17.28 | Hit bottom at $9.46 (Apr). Recovering. Current ~$14.22. |
The stock is down ~45% from 2021 highs. During this same period, utility peers (PNW, OGE, D) are flat to up. AES has massively underperformed. This is either justified (the risks are real) or a contrarian opportunity (the market is overweighting the negatives and ignoring the data center/renewables growth).
AES is actively simplifying its portfolio, which should reduce the conglomerate discount over time:
| Asset Sale | Proceeds | Status |
|---|---|---|
| AES Brasil (47.3% interest) | ~$630M | Closed October 2024 |
| AES Ohio (30% to CDPQ) | ~$546M | Expected 1H 2025 |
| AES Dominican Renewable (50%) | TBD | Signed, expected 2025 |
| Amman East & IPP4 (Jordan) | Sold down | Completed 2024 |
| Mong Duong (Vietnam coal) | Held for sale | Pending |
| Target total through 2027 | $3.5B | ~75% announced or closed |
The strategy is clear: sell non-core/international assets, concentrate on U.S. renewables + U.S. utilities + high-returning LatAm energy infrastructure. Use proceeds to fund growth CapEx and reduce recourse debt. This should progressively reduce the conglomerate discount.
| Target (by ~2031) | Required Entry for 10x | Entry Price/Share (~712M shares) | Status |
|---|---|---|---|
| $60B (mid bull: $5 EPS x 17x) | $6.0B | ~$8.40 | Just below 52-week low of $9.46 |
| $75B (high bull: $5 EPS x 21x) | $7.5B | ~$10.50 | Near recent lows |
| $94B (extreme bull: $6 EPS x 22x) | $9.4B | ~$13.20 | BELOW current price |
At $14.22, AES is trading just above the 10x entry price under the extreme bull case. This means:
The risk/reward is asymmetric. Downside is limited (floor at $7-10, dividend support at $10), upside is 3-7x in the bull case. The stock has already been beaten down -- it priced in bad news that the earnings do not support.
Ideal entry: $10-12/share ($7-8.5B market cap). At that price, even the mid bull case gives you 5-7x. The stock touched $9.46 in April 2025. If it revisits that zone (market correction, another IRA scare), that is a compelling entry. At current $14.22, it is still interesting but not a table-pounder.
Our assessment: The diversification is a feature at this price. The market is pricing AES as if the complexity will destroy value. But the underlying businesses -- U.S. renewables with long-term PPAs, regulated utilities in data center corridors, gas plants generating steady cash flow -- are individually valuable. AES is doing the right thing by simplifying (selling international assets, exiting coal). As the portfolio concentrates on U.S. assets, the discount should narrow.
The comparison that matters: AES at 6x forward P/E with 11.9 GW renewable backlog + 2.1 GW data center utility load vs. PNW/OGE at 18x forward P/E with no comparable renewable pipeline. That is either a massive value trap or a massive opportunity. We lean toward opportunity.