Bentley's Prepaid Deposits Decide Whether It Is Fairly Priced or Cheap

Summary and thesis

Bentley Systems sells the design, simulation and operations software that civil, structural, utility and mining engineers use to build and run physical infrastructure. Its products sit inside multi-year public and private projects where file formats, trained engineers and specified deliverables make switching slow and costly.

The result is a franchise that grows in the low teens with unusually low stock pay for software and owner cash that rose in each of FY2023, FY2024 and FY2025.

The thesis is that Bentley is a durable system of record whose consumption pricing turns more automated design into more usage rather than fewer seats. Superintelligence tools that generate or check engineering work still need Bentley's physics engines, data formats and project history, and Bentley bills large accounts on daily usage rather than per seat.

The open issue is valuation, and it turns on how to treat the Cloud Services Subscription (CSS) deposits that customers prepay before they consume. On the strictest treatment, the stock trades at 43.3x owner cash; counting all deposit cash lowers that to 33.82x and the growth the price requires to below current ARR growth.

Yet because the historical floor multiple moves with the treatment, the floor value stays between $34.12 and $35.92 per share under all three views. The anchor remains provisional, so all three are shown below.

Business model and products

Products. The portfolio spans MicroStation (design and drafting), OpenRoads (road and civil design), STAAD (structural analysis), SYNCHRO (construction planning) and other discipline tools, plus the Bentley Infrastructure Cloud and Asset Analytics for digital twins of operating assets. In Q2 2026 a new enterprise finance and quote-to-cash platform went live.

Bentley is also exposing its engines to third-party superintelligence assistants, which it says creates value it intends to monetise in due course; no revenue figure is disclosed.

Customers. About 42,000 accounts in 189 countries. Engineering and construction firms generated 52% of FY2025 revenue and owner-operators (public and private infrastructure owners) the remainder.

Resources, which includes mining, led growth in Q2 2026, followed by public works and utilities, including the electric grid.

Pricing and revenue model. Subscriptions were $378.6M of $410.7M in Q2 2026 revenue, about 92%. Large accounts buy E365, a consumption plan billed on daily usage that requires a prepaid CSS deposit drawn down as customers use the software.

Smaller users buy term or perpetual licences through the Virtuosity e-commerce channel. Growth comes from infrastructure spending (including US federal infrastructure funding, grid, water and mining), moving perpetual users onto subscriptions, consumption uplift and price.

Retention and backlog. Last-twelve-month recurring revenue net retention was 109%, flat year on year. Gross retention is not disclosed.

Remaining performance obligations are only $280.6M, 94% due within 12 months, because consumption contracts do not book long backlog; RPO is not a useful gauge here. ARR was $1,536.0M at Jun 30, 2026, up 12% in constant currency.

Concentration. A customer-concentration percentage was not located in the sources reviewed. The 42,000-account base and the split between engineering firms and owner-operators suggest a broad base, but no top-customer figure is stated here.

Go-to-market. Direct enterprise sales for E365 accounts and Virtuosity e-commerce for individual and small-firm users.

Financials

Revenue grew 11% to 14% a year from FY2021 to FY2025, from $965M to $1,502M. Margins compressed slightly in Q2 2026 as Bentley invested, and net interest expense rose to $17.3M in H1 2026 from $7.3M after refinancing.

Company free cash flow of roughly $499M over the trailing twelve months is operating cash flow less capex, before any deposit adjustment.

The balance sheet carries $1,220.9M of debt at face: a $95.9M revolver and a $550.0M term loan, both due October 2029, plus $575.0M of convertible notes due July 2027. Cash was $147.0M, so net debt is $1,073.9M. The 2026 notes ($677.8M) were repaid in H1 2026, mainly with the new term loan and revolver.

Valuation: three ways to treat the deposits

Owner cash here is operating cash flow less the deferred-revenue and deposit build, capex including capitalised software, stock pay and interest income. At the $39.70 close on Oct 23, 2026 and 311.79M basic weighted shares, market value is about $12.38B.

Adding $1,073.9M of net debt gives enterprise value of about $13.45B, or 8.4x trailing revenue. Every treatment below starts from the same trailing-twelve-month cash flow to Jun 30, 2026 and differs only in how much of the CSS deposit build it keeps.

How each anchor is built.

What an investor must believe

Strict ($310.7M). Deposits are customer money, not owner cash. Because the 10-K says unused balances roll over or are refundable and no breakage is disclosed, every dollar of the build could in principle flow back out.

A strict holder accepts that the balance has grown every year from $110.3M in 2020 to $463.3M in 2025 and $496.3M at Jun 30, 2026, but refuses to capitalise it until the refund history is known. This is the conservative house view.

Normalised ($361.6M). Deposits behave like deferred revenue in a growing subscription business: as long as customers keep consuming, a balance proportional to revenue stays on the books permanently, and its growth is real owner cash. This holder must believe that refunds are rare, that rollovers dominate, that the deposit-to-revenue ratio will stop rising once E365 migration matures, and that the build then converges on about balance times revenue growth.

It requires no assumption about breakage, which is zero.

Full ($397.7M). The whole build is repeatable operating cash. This holder must believe that the migration that took the ratio from 13.8% to 30.8% of revenue continues for years, so that $78M to $92M a year of deposit growth persists rather than fading toward about $51M.

It also requires that a downturn in engineering activity does not prompt customers to size smaller deposits, since balances are refundable. The record supports this so far, with positive builds in 2020 and 2022, but the balance compounding at about 33% a year against revenue near 13% cannot persist indefinitely.

Evidence that would settle it at the Nov 5 print

• Refund disclosure. Any statement in the Q3 10-Q or on the call that splits rollovers from refunds, or quantifies refunds, would show whether deposits behave like a liability or like sticky prepaid revenue. Refund history is not available.
• Burn-down. The Q3 balance against the Jun 30 level of $496.3M. In 2025 the balance fell from $447.9M at Mar 31 to $427.7M at Sep 30. A similar seasonal decline is normal; a sharper one would suggest customers are drawing down faster or sizing deposits smaller.
• Deposit growth against revenue. Year-on-year growth in the Sep 30 balance compared with revenue growth near 12%. If the balance grows far faster, migration is still running and the full case gains ground; if growth converges toward revenue growth, the normalised case is confirmed; if the balance grows slower than revenue, even the normalised build is too generous.
• E365 penetration. Any disclosure of the share of ARR on consumption plans would show how much migration remains.

Risks

• Family control. The Bentley family holds all Class A shares at 29 votes each. Directors and officers as a group controlled 55.8% of the vote at Mar 31, 2026, and family trusts hold further Class B stakes. Outside holders have little say over strategy, capital allocation or a sale.
• Pledged shares. The Executive Chair has pledged 2.5M shares to PNC. A sharp fall in the stock could force sales or a margin call, though the pledge is small against family holdings.
• Deposit reversal. Balances are refundable. A downturn in engineering activity could shrink deposits and turn the build negative, which has not happened in any year on record.
• Refinancing and interest. The $575.0M convertible matures July 2027, and net interest expense more than doubled in H1 2026.
• Public budgets and consumption. Infrastructure spending cycles and usage-based billing add volatility.
• New entrants. Superintelligence design-automation tools could compress engineering hours; consumption pricing is the offset, and monetisation of engine access is unproven.
• Geographic exposure. China and Russia exposure was not verified.

Catalysts

The Q3 2026 print is scheduled for Nov 5, 2026, before the open. Points to watch: ARR growth against the 10.5% to 12.5% guide, net retention near 109%, whether the full-year revenue and free cash flow outlook holds, the Q3 deposit balance against the seasonal burn-down seen in 2025, and any update on monetising engine access by third-party assistants.

Over the next year, the treatment of the July 2027 convertible and the pace of buybacks will shape per-share owner cash.