Computer Modelling Group: Cheap Enough to Buy Back, Not Yet Proven to Buy

Computer Modelling Group Ltd. (TSX: CMG; OTC: CMDXF) makes reservoir-simulation software — the digital models oil and gas producers run to decide where to drill and how much oil a field will yield. Its annual general meeting on September 11 was, on its face, a non-event: the company announced the voting results, its directors were re-elected, and no contested matter changed anything. A routine AGM vote is exactly the kind of press release an investor skips, and this one contained no new number.

But read it against what the same board has been doing for the past month, and the meeting becomes a bookmark on the company's biggest capital decision in years. CMG is offering to buy back roughly 6% of its own shares — up to C$20 million — at C$4.00 to C$4.50 apiece, a premium of 7% to 20% over the stock's close the day before the deal was announced. Management said it believes the shares trade at "a meaningful discount to intrinsic value" and that the market under-weights the durability of the recurring revenue base, and no director or officer plans to tender their own stock.

That is a bold message to send about a company whose top line is presently shrinking. It deserves inspection — and it carries a built-in test that lands within two quarters.

Why revenue is falling, and why the board is buying anyway

The headline numbers read like the opposite of a buy signal. Revenue for the June quarter (fiscal Q1 2027) was $27.8 million, down 6% from a year earlier. Organic revenue fell 16%, masked only by a 10% lift from the acquisitions of SeisWire and Rose & Associates. Net income fell 60%. For all of fiscal 2026, revenue edged down 3% and adjusted EBITDA fell 18% to $36.1 million.

The board's counter-argument is a decomposition of that slide. The steepest organic drop traces to one lost customer contract booked in Q2 fiscal 2026 and lapped ever since — Q1 2027 is the last quarter carrying that handicap. A separate chunk is deliberate: CMG is winding down low-margin professional-services work (a Bluware cleanup and the end of CoFlow development funding), trading revenue for mix. Strip both out, and the core is recurring-software annuities — 73% of total revenue — which fell only 3% in the quarter. Management guides organic recurring revenue to hold steady for the year, adjusted EBITDA to be flat, and free cash flow to improve.

The recurring-software base is the whole thesis, and it has genuine moat characteristics: CMG sits in an oligopoly with the software arms of Schlumberger and Halliburton, its models are embedded in years of engineers' training and reservoir data, and it reinvests well over 20% of revenue in R&D. Software of this kind carries high margins and sticky multi-year agreements; a major international operator signed a multi-year simulation license in November 2025.

The cheap-enough bridge

This is where the valuation does the work. The stock is down about 39% over the past year, from a 12-month high near C$6.54 to roughly C$3.80, near the bottom of its range. But the operating deterioration is much smaller than that price fall — revenue down single digits, EBITDA down 18%. The multiple did the compressing, and it compressed more than the business did.

At the current price, CMG trades at roughly 8.6x enterprise value-to-EBITDA and a ~6% free cash flow yield, against a balance sheet that is nearly net cash: C$17.7 million of cash and an essentially undrawn C$100 million credit line. For a 73%-recurring software franchise with 29% EBITDA margins, that is cheap by the standards of comparable software — if the growth stops falling. The board's premium-price buyback is a way of saying it believes exactly that, and the fact that insiders aren't tendering argues the offer is not a marketing gesture.

The strongest bear fact is that the recent growth has come from acquisitions, not demand: organic revenue has been negative, so the reported numbers are the arithmetic of buying revenue rather than earning it. If CMG's organic decline is not the one-time lapse management describes but the start of share loss — or a capex-constrained energy market consolidating vendors — then the moat does not protect revenue, and "cheap" is the market pricing a real problem. That risk is exactly why the multiple sits where it does.

The proof window

The setup produces a clean, near-term test, and two dates matter. The C$20 million buyback expires September 21; the results of the auction — how many shares tendered and at what price — will show whether holders agree with the board's C$4.00 to C$4.50 view of value. Then fiscal Q2 2027 earnings land around November 12, the first quarter where the lost-contract comparison normalizes. If organic recurring revenue re-stabilizes there, the "one-time lapse" story holds, the dip was overdone, and the buyback floor near C$4 begins to look like the cheap side. If it stays negative, the explanation was wrong and the market's discount was justified.

That makes CMG a wait rather than a rush. The valuation has reset faster than the business has deteriorated, and management's own cash is behind the bet — two reasons the risk/reward now tilts toward the buyer. But the thesis lives or dies on a promise of organic stability that is not yet proven, and the next two quarters will supply the verdict either way. You do not need to own it today to know by Thanksgiving whether the slump was the story or just a chapter.