Axia: This Latam Utility Hit My Target (+40%) in 3 Months. Now What?
The call played out. The model is refreshed. But my updated rating might surprise you.
Back in January, I published a deep dive on a Latin American utility offering 40% upside.
Three months later, it touched my target price. The rerating played out faster than I expected.
The stock has since pulled back a bit, Q4 results just dropped, and I’ve refreshed the model. The new target is 30% higher. But my rating has changed too, and probably not in the direction you’d expect.
Table of Contents:
TLDR
• Trigger: AXIA hit our original $13.40 target, retraced to $13.15, and Q4 FY2025 results provide a natural checkpoint to update the thesis.
• Thesis scorecard: Of my 5 original thesis pillars, 4 are confirmed, and 1 is evolving. The decotization story, efficiency gains, governance improvement, and transmission stability are all tracking. The demand/data centre pillar is evolving as the timeline has stretched.
• Verdict: HOLD with an updated target of $17.45 (up from $13.40). The thesis is working, but the easy money has been made. Risk:reward at $13.15 is balanced, not compelling enough for a BUY.
Let me walk through what happened, what changed in the model, and why I’m staying patient.
Thesis Scorecard
In the original deep dive, I built the bull case around five pillars. Here’s where each one stands after Q4 FY2025.
Pillar #1. Decotization Unlocks Free Market Pricing – Confirmed
Original thesis: Legacy quota contracts were phasing out, freeing 29–43% of Axia’s capacity to sell at market prices instead of cost-based rates. By 2027, quota exposure drops to zero.
What happened: This is playing out exactly as expected. In Q4 FY2025, Axia’s generation margin rose to R$101/MWh, up from R$78/MWh a year earlier. That’s a 29% improvement driven directly by the shift from quota pricing to free market contracts. The company’s ACL (free contracting environment) volumes continue to grow as quota volumes decline. Management confirmed on the Q4 call that the remaining quota exposure is on track to expire on schedule.
Assessment: This pillar is the engine of the entire thesis, and it’s firing. The margin expansion from decotization is showing up in the numbers, not just in the projections.
Pillar #2. Surging Power Demand / Data Centre Boom – Evolving
Original thesis: Brazil’s data centre demand jumped 330% in a single year, with total demand projected to reach 13.2 GW by 2035. Axia, as the only generator with massive uncontracted capacity, was positioned to be the go-to supplier.
What happened: The demand story hasn’t weakened, but the timeline has stretched. Brazil’s data centre buildout is proceeding, with Microsoft, AWS, and Google all expanding campuses. But hyperscale construction timelines are measured in years, not quarters. Axia’s plan to build a data centre adjacent to its Xingó hydro plant in Sergipe is still in early stages. In Q4 FY2025, Axia won 4 lots in a transmission auction (R$1.63 billion in capex, R$140M RAP), which strengthens the infrastructure side but doesn’t directly monetise the data centre angle yet.
Assessment: The structural demand thesis is intact, but the near-term earnings impact is still a 2027+ story. I’m not downgrading this pillar to “challenged” because the direction is right, but I’m flagging that the market may be patient only for so long.
Pillar #3. Post-Privatization Efficiency Gains – Confirmed
Original thesis: Operating expenses had been cut 18%, headcount reduced 17%, and non-core assets divested since privatization. The new management was running Axia like a business, not a bureaucracy.
What happened: PMSO (personnel, materials, services, and other expenses) fell 12.8% y/y to R$6.7B in FY2025. That’s on top of the cuts already made. Axia also completed the disposal of its nuclear subsidiary stake and continued shedding legacy assets like the Cepel research centre. Investments jumped 28% y/y to R$3.9B in Q4, showing that the efficiency gains aren’t just about cutting but about redirecting capital to higher-return projects.
Assessment: The cost discipline is holding. Management isn’t just coasting on early privatization wins. They’re still finding ways to trim while reinvesting.
Pillar #4. Governance Transformation – Confirmed
Original thesis: The Supreme Court settlement confirmed the 10% voting cap, the government gets 3 of 10 board seats, and Axia was on a path to Novo Mercado listing (one share, one vote).
What happened: The governance framework is stable. The Novo Mercado migration remains in progress. More importantly, Axia paid a record R$8.3B in dividends for FY2025, including extraordinary dividends. A company still under government thumb doesn’t pay out R$8.3B to shareholders. This is a tangible signal that capital allocation is driven by returns, not politics.
Assessment: The governance overhang that haunted Eletrobras for years has largely faded. This pillar is confirmed and doesn’t need much monitoring.
Pillar #5. Transmission as Stable Ballast – Confirmed
Original thesis: Axia operates 37% of Brazil’s transmission grid, providing regulated, inflation-adjusted returns that balance the volatility of the generation business.
What happened: Transmission revenue grew 4.3% in FY2025. The segment continues to deliver stable, predictable cash flows. Axia won additional transmission lots in the April 2025 auction, adding R$140M in annual permitted revenue (RAP). However, ANEEL’s reduction in RBSE (Existing Basic Network) revenue is a headwind worth noting. The regulator adjusted the remuneration base downward, which creates some drag on transmission earnings going forward.
Assessment: The ballast is holding. The RBSE reduction is a modest negative, but transmission remains the steady anchor it was meant to be.
FY2025 Earnings Review: Noisy Numbers, Clean Trend
Let me get the headline numbers out of the way: FY2025 reported EBITDA was R$8.5B, down 67.5% y/y. Net income was R$6.6B, down 36.8%. These numbers look terrible on the surface.
They’re misleading.
The EBITDA collapse is almost entirely explained by a swing in regulatory remeasurements. In FY2024, Axia booked R$6.1B in positive remeasurements related to transmission contract revaluations. In FY2025, that line flipped to negative R$4.1B. That’s a R$10.2B swing in a single non-cash accounting line. Add in R$7.2B in losses from the nuclear subsidiary disposal, and you have your EBITDA headline.
Strip those out, and the underlying business looks different.
Net revenue rose 2.7% to R$41.3B. PMSO fell 12.8%. The generation margin jumped from R$78/MWh to R$101/MWh in Q4 alone. These are the metrics that matter for the thesis.
Q4 FY2025 specifically was stronger than it looked. Adjusted regulatory EBITDA came in at R$6.4B when you exclude the R$250M wind reimbursement, R$225M transmission pass-through, R$108M incentive programmes, and R$60M rebranding costs. Net income for the quarter was R$1.25B, up 141% y/y on an adjusted basis.
One unusual item: Axia recognised R$12.6B in deferred tax assets during Q4. This reflects management’s confidence in future taxable income (you can only book a deferred tax asset if you expect to use it), but it also inflated the balance sheet in ways that don’t reflect cash generation.
On the balance sheet, net debt rose 23.4% to R$46.5B. Gross debt stood at R$74.3B. Net Debt/EBITDA hit 5.5x on a reported basis, though that ratio is distorted by the same EBITDA noise. On an adjusted basis, the ratio is more like 2.5–3.0x, which is manageable for a utility with contracted cash flows.
Guidance Comparison
Management hasn’t issued formal forward revenue or EBITDA guidance in the traditional sense. But we can compare the operational targets and strategic commitments from the original deep dive to where they stand now.
Management credibility check: Over the past year, management has delivered on every operational metric they’ve guided to. PMSO came in below target. The nuclear divestiture was completed. Dividend policy was more aggressive than expected. The one area where they haven’t been transparent is leverage, where the reported 5.5x number looks alarming until you adjust for the accounting noise. I’d rate management credibility as solid, with a caveat that they need to provide cleaner adjusted figures going forward.
Valuation Update
DCF Assumption Changes
Here’s what changed in the model, and why.
The EBIT margin trajectory now climbs to 50.9% by the end of the forecast period, up from 41.7% in the original model. This reflects the compounding effect of decotization: as more capacity moves from quota pricing to free market pricing, each megawatt-hour generates a fatter margin.
I actually revised down revenue growth in 2026 from 12.7% to 9.1%. I’m being more conservative on the pace of ACL contract wins than the original model assumed. But that conservatism is more than offset by the margin expansion and lower discount rate.
Implied Expectations Analysis
This is where the HOLD thesis crystallizes.
At $13.15, the market is pricing in roughly 4%–5% revenue growth and EBIT margins in the high 30s through the medium term. My updated model assumes 7%–9% revenue growth in the near term (decelerating to 4.4%) and margins climbing above 45%. The gap between the market-implied scenario and my model represents the upside, and it’s real. But it’s not as wide as it was in January.
In January, at $9.46, the market was pricing in essentially flat revenue and stagnant margins. The implied WACC was north of 14%. The stock was being valued as if decotization didn’t exist. That mispricing was obvious and actionable.
Today, the market has corrected most of that gap. The implied WACC at $13.15 is closer to 12%, which is still above my 10.9% assumption, but the delta is smaller. The market now acknowledges that Axia is a different company post-privatization. It just hasn’t fully priced in the margin trajectory.
That remaining gap, from $13.15 to $17.45, is a 33% upside. Decent for a utility. But to capture it, you need the margin story to unfold over 2–3 years, and you need Brazil’s macro backdrop to cooperate. The risk-to-reward profile has shifted from “mispriced and obvious” to “fair value with upside if execution continues.”
Risk Register Update
Faded: Political Re-Nationalization Risk
This was the biggest fear when I wrote the deep dive. The Lula government was testing the waters, and the Supreme Court settlement was fresh. Since then, the government has accepted its 3-of-10 board seats, Axia paid R$8.3B in dividends (the government got its share), and the political noise has quieted. This risk hasn’t disappeared, but it’s faded from a headline concern to background noise.
Unchanged: BRL/USD Currency Risk
The model uses a fixed R$5.00/USD rate. The real has been volatile, trading between R$4.80 and R$5.80 over the past year.
For a USD-denominated investor in the ADR, currency moves can easily eat 5–10% of returns in either direction. This remains a live risk that the model doesn’t fully capture.
Unchanged: Hydrology / Power Price Volatility
Brazil’s power grid is hydro-dependent. A dry year can cause spot prices to spike (good for Axia’s uncontracted capacity) or force curtailments (bad). A wet year lowers prices. The 2025 hydrological year was relatively normal, but climate patterns are unpredictable. This is structural and won’t go away.
Evolving: Leverage Concerns
Net Debt/EBITDA at 5.5x reported looks uncomfortable. The adjusted figure (stripping out the remeasurement noise) is more like 2.5–3.0x, which is fine. But gross debt at R$74.3B is substantial, and with Brazilian interest rates still elevated (Selic at 15.00%), refinancing costs matter. If EBITDA normalises in FY2026 as I expect, the reported ratio will look much better.
New Risk: RBSE Revenue Reduction
ANEEL reduced the RBSE (Existing Basic Network) remuneration base. This creates a drag on transmission revenue going forward. It’s not thesis-breaking, but it wasn’t in the original risk register and deserves monitoring. Transmission was supposed to be the boring, predictable segment. Regulatory adjustments to the remuneration base remind you that “regulated” doesn’t mean “guaranteed.”
New Risk: EBITDA Noise Masking True Performance
The R$10.2B swing in regulatory remeasurements between FY2024 and FY2025 makes it harder for the market to assess underlying profitability. If Axia can’t present clean adjusted figures that gain market trust, the stock may trade at a wider discount than the fundamentals warrant. This is partly a communication risk.
Relative Performance vs. Comps
Since the original deep dive in January, AXIA is up 37.2% from the entry, making it the top performer in the peer group. Engie Brasil came close at 34.4%, but the rest of the field lagged: the S&P 500 Utilities sector returned 11.2%, Enel Chile managed 7.1%, and Pampa Energia was essentially flat at -0.8%.
On a valuation basis, the picture is more nuanced. AXIA’s forward EV/EBITDA has re-rated to 6.7x, sitting in the middle of the comp set. Enel Chile trades at 7.6x, which makes sense for a stable Chilean utility with lower country risk. Pampa Energia is at 4.3x, but Argentina is Argentina.
The key takeaway: AXIA has re-rated on price performance but the multiple hasn’t stretched to expensive levels. At 6.7x forward EBITDA, it’s still below Enel Chile and roughly in line with where Brazilian utilities trade. The stock isn’t cheap anymore, but it isn’t pricing in the full margin expansion story either. In January, AXIA was a clear relative value play. Today, it’s a hold-and-watch. The multiple gap has narrowed, and further re-rating from here depends on the earnings proving out, not just the narrative.
Verdict: Hold – The Thesis Is Working, the Price Has Caught Up
Axia did exactly what I expected it to do: the market re-rated the stock as decotization showed up in the margins, and now the easy money from that mispricing has been captured. The remaining 33% upside to $17.45 is real, but at $13.15 the risk:reward is balanced, not compelling. I’m holding my position and watching for a catalyst that accelerates the margin story.