Open in interactive viewer → charts, metric popovers & call review

Axtel: Two Bidders at the Door, a Guidance That Hinges on Dark Fiber

Mexico's enterprise-IT and fiber operator posted solid government growth and an AI-driven cost cut — but management spent the call refusing to talk about the only thing that matters: the competing tender offers.
AXTLF · Earnings Call · 2026-10-22

The Bid Is the Story, the Business Is the Distraction

Axtel's Q3 2026 call opened with the CEO acknowledging what every listener already knew: the boards of both Axtel and Controladora Axtel have received separate proposals to acquire 100% of each company via tender offers. Everything after that was, in effect, a business update delivered under an M&A cloud. When pressed on timeline, the CFO was blunt: “We don't have time line. The processes are under the control of the investor -- potential investors not really under the control of the company... at the end of the day, we would have to continue running the business.” — Adrian de los Santos Escobedo, CFO · 2026-10-22 Asked directly whether the offer price reflects the company's real value, he declined entirely — citing the obligation for the board to opine roughly ten days into any live tender offer.

Given the processes and the situation in that respect right now, I wouldn't say much about that. It's an obligation that once there's a tender offer in the market, the Board of Directors will have to comment on the merits of the offer about 10 days into the tender offer.

Adrian de los Santos Escobedo, CFO · 2026-10-22
That is the tension at the heart of this report. For a ~$278M market-cap name with no price tape provided and no curated keyword history of its own, the deal overhang is the investment case; the operating numbers are the floor under it. And note the asymmetry: the analyst questions asked specifically about a deadline from Transtelco Holding, implying at least one bidder is already named. Management offered no deadline because — in the CFO's framing — this was never an open competitive auction with a clock attached.

Guidance Held Hostage to Dark Fiber

The quarter itself was respectable but uneven. Revenue rose 3% year-over-year, with Government up 41% (recurring +25%, nonrecurring +84%) and Enterprise up just 2% — IT and cybersecurity +15%, telecom -1%, voice -11%. Wholesale infrastructure fell 18% against a tough prior-year comparison that included a large upfront dark fiber contract. Gross margin compressed to 67% from 72%, and comparable EBITDA of MXN 917 million rose 1% year-over-year (6% sequentially), a 29% margin versus 30%. The CFO was transparent about why margins slipped: a richer mix of lower-margin nonrecurring revenue, plus roughly MXN 150 million of lost high-margin dark fiber revenue year-to-date. His defense — that nonrecurring work carries little CapEx behind it, so cash conversion is comparable — is fair, but it reframes Axtel as a lumpier, project-tilted business than a pure recurring telecom. That lumpiness shows up most starkly in guidance, which management has effectively tied to a handful of deals it does not control: “Guidance, it's still achievable. As I said, if we are successful in these dark fiber contracts closing in the fourth quarter. We could reach our guidance.” — Adrian de los Santos Escobedo, CFO · 2026-10-22 Then the honest kicker — “When we estimated the guidance for this year, we were expecting a totally different ball game.” — Armando de Pena, CEO · 2026-10-22 That is a CEO telling you the year was re-planned mid-flight. Guidance is not a forecast here; it's a contingent claim on contract closings.

AI Inside, Not AI Hype

The genuinely new operating thread is internal AI deployment. Axtel pushed AI into service assurance and customer support, which enabled an 8% headcount optimization, a MXN 144 million reorganization charge, and a fast payback: “The savings in terms of more efficient expenses or reduced expenses, we expect payback of approximately 1 year.” — Adrian de los Santos Escobedo, CFO · 2026-10-22 The CEO added that revenue per employee has tripled since 2022 — a four-year trend, not a fresh quarter's novelty. The more interesting move is how Axtel positions itself against the market's AI AI data center frenzy. It has invested over 1,000 kilometers of new fiber linking Querétaro data centers to McAllen, Texas, and forged an alliance with Trans American Fiber for a submarine route to the U.S. East Coast and Latin America. Yet the CFO drew a deliberate line: “Axtel does not serve hyperscalers, LLMs, that kind of clients of significant capacity in data centers. We use our cloud and data center capabilities to serve enterprise clients.” — Adrian de los Santos Escobedo, CFO · 2026-10-22 This is adjacency, not the main event — data center interconnect as a wholesale and enterprise connectivity play, not a hyperscale build-out.

We are seeing as a big opportunity, mostly for Axtel and for the IT services that we provide. We are not thinking at this moment to build a big data center or anything like that.

Armando de Pena, CEO · 2026-10-22
That restraint looks well-judged given the tape. The global AI data center theme is one of the widest in the market — dozens of tickers — but it has cooled sharply over the last 90 and 30 days after a strong twelve-month run. Elsewhere, peers lean harder into the narrative: SAP talks up its Business AI platform, Nokia its IP and optical networks, Tele2 its agentic AI roadmap. Axtel's version — AI agent tooling deployed on its own cost base and cybersecurity resale to Mexican enterprises — is smaller and more defensible.

The Verdict on the Numbers

Strip out the noise and Axtel did the harder thing well: expenses fell 8% year-over-year, corporate costs dropped 36% on restructuring and a one-time employee-benefit provision, CapEx fell to $20 million (11% of revenue) from $30 million (18%), and free cash flow came in at positive $23 million. Cash rose to $44 million from $38 million, while net debt of $454 million left leverage at 2.4x comparable EBITDA versus 2.3x a year earlier — stable, if uninspiring. So we have a company generating real cash, cutting costs with AI, growing government and IT/cyber revenue at healthy rates, and yet unable to promise a full-year number without two or three dark fiber contracts landing in the next ten weeks. Add two live tender offers and a board forbidden from opining on price, and you get a stock that trades on event probability rather than earnings power. The business is executing; the shareholder return is being decided elsewhere.