Unidata's "Halved" Profit Is Mostly an Accounting Ghost — the Real Fight Is the Margin Squeeze

Unidata (BIT:UD) reported first-half net profit of €2.0 million against €4.1 million a year earlier, and the "profit halves" headline writes itself. Both camps are looking at the same two numbers. The bear sees a widening margin squeeze and a balance sheet loading up with debt to fix it. The bull sees a one-off accounting ghost from the prior year doing all the damage, with revenue and EBITDA actually up 8%. At roughly €100 million in market value and about 10 times forward earnings, only one of those readings leaves room for error.

The shared record

Before the duel, the facts both sides can sign:

• H1 2026 consolidated revenue of €53.6 million, up 8% from €49.5 million in H1 2025. Service revenue rose 5%, infrastructure revenue 17%. Recurring revenue of €34.5 million is 64.3% of the total.
• EBITDA of €13.1 million, up 8%; adjusted EBITDA of €13.7 million, flat on the year. Adjusted EBITDA margin slipped to 25.6% from 27.7%.
• Net profit of €2.0 million, versus €4.1 million. The year-ago figure contained a non-monetary, non-recurring financial gain; strip it out and H1 2025 profit was about €2.2 million, so the real decline is mild.
• Net financial debt rose to €43.9 million from €37.7 million at the end of 2025, partly funding €10 million of H1 investment, including €4.5 million into UniCenter, a "green" TIER IV colocation data center in Rome.
• Management reconfirmed 2026 guidance: revenue of €114–116 million, adjusted EBITDA of €28–29 million (around a 25% margin), and net debt of €37–39 million.

That record is the whole fight in miniature. Revenue and EBITDA are growing, the nominal profit collapse is largely a comparison artifact, but the margin is compressing and the balance sheet is stretching. The disputes are whether those two trends are temporary and whether the price already pays fair compensation.

Round one: Did profit really halve?

The bear's opening is the headline itself, and it is the strongest thing the bear has. A 50% earnings decline is the kind of number that makes a reader reach for the sell button before reading a word.

But the number is mostly an illusion. The €4.1 million in H1 2025 was inflated by a one-off, non-monetary financial gain, not by operations. Excluding it, the prior-year profit was approximately €2.2 million, so the real fall to €2.0 million is on the order of single digits. In the meantime revenue grew 8%, EBITDA grew 8%, EBIT rose 4% to €7.0 million, and the underlying EBITDA margin (24.5% versus 24.7%) barely moved.

The bear can concede the one-off but answer that "non-monetary gain" flatters the comparison in a second way: the company reports an adjusted EBITDA figure that strips out items, and this year cuts through a nominally higher-margin services line. That is the turn into the business's real question.

Round one to the bull. The headline decline is not evidence of a collapsing business; it is a favorable baseline comparison unwinding.

Round two: Is the margin squeeze structural?

Here the bear gets its cleanest shot. Adjusted EBITDA margin fell two full points, to 25.6% from 27.7%, on flat adjusted EBITDA. That is a real compression, not an artifact. Two forces explain it. First, mix: the fast-growing business is lower-margin infrastructure (up 17%), while the higher-margin services line grew only 5%, so the revenue mix dilutes overall profitability. Second, the company is deliberately spending — €10 million in the half, including the UniCenter data center, plus €5.5 million of tangible and intangible capex.

The bear's worry is that this is the beginning of a slide, not a blip: a telecom-plus-data-center operator borrowing to fund an asset-heavy build that will add depreciation and interest for years while margins settle lower. At roughly 1.7 times the €25 million-odd of net profit-adjacent cash flow, leverage is climbing just as the profit is expected to decelerate.

The bull's answer is that the margin dip is the bill for a future revenue stream, not a loss of pricing power. Recurring revenue is already 64% of the total and growing; the customer base rose 5% to 32,629; the fiber network extended 380 kilometers to about 8,880. The UniCenter build is an investment in a higher-return, more resilient recurring business — cloud, IoT, data center, cybersecurity — that management frames as the medium-term growth engine. Depreciation and interest hit today; the EBITDA those assets throw off lands tomorrow.

Round two is a genuine toss-up, and the evidence decides it by a whisker against the bear. The margin compression is real and mix-driven, but so is the capital being deployed to offset it. Whoever wins this round is whoever is right about whether UniCenter and the new contracts convert investment into earnings rather than into idle capacity.

Round three: Can the balance sheet carry the bet?

Reverse-engineer the price and the question sharpens. At a market cap of about €100 million, an enterprise value near €140 million (netting the €43.9 million debt), and roughly 10 times forward earnings, the market is not paying for a high-growth miracle. It is paying a modest multiple for a business with 64% recurring revenue, plus optionality on a data center.

The bear reads that same structure as the trap: an ordinary telecom multiple that already embeds the UniCenter dream, with net debt guidance of €37–39 million acting as the ceiling that borrowing is racing toward. If the data center disappoints, the multiple compresses to a facilities-operator multiple and the equity loses twice — no growth and heavier leverage.

The bull counters that the price is the cheap part. Ten times forward earnings for a business growing revenue and EBITDA at 8% with a majority-recurring base is not demanding; the stock is up roughly a fifth year to date and about a quarter over twelve months, but largely in line with the FTSE MIB index. The margin of safety is that you are not paying a premium for the data center, so the downside is a business that still grows and pays the ordinary multiple.

Round three to the bull, narrowly. The valuation does not require the UniCenter bet to succeed; it just needs the underlying telecom business to keep compounding. The bear's scenario needs multiple compression on top of a failed build to justify avoiding the stock.

The ruling

The headline is a headline, not a thesis. Unidata's underlying profit did not halve; a one-off, non-monetary gain inflated the year-ago comparison, and the real decline is single digits against 8% revenue and EBITDA growth. The genuine dispute — and the only one that matters — is whether the margin step-down and the rising net debt are the cost of a deliberate, funded data-center bet or the first sign of a low-margin, capex-hungry slide. On the evidence so far, the bet is being paid for with money the business is generating, the backlog of new contracts is building, and the price is not capitalizing the optionality.

So: the business case and the stock case both go to the bull at this price, but only conditionally. The burden of proof sits on management to show the investment converts into earnings, and the tripwire is concrete. When full-year 2026 results arrive in early 2027, the bull needs adjusted EBITDA to land within the €28–29 million guidance and net debt to stay inside €37–39 million. Drift below a ~24% adjusted EBITDA margin while net debt climbs past €39 million would flip this from a funded growth cycle into a genuine margin squeeze — and that is the single observation that would pull the verdict back to the bear.