LMAX’s 2024 PYTH disposal gain explains much of the fall in reported 2025 earnings

LMAX Group's reported revenue fell 44% and EBITDA fell 78% in 2025, but the headline collapse is largely a comparison against a one-off digital-asset gain disclosed in the prior year. The group's annual report says it realised a $62.5 million gain from the sale of PYTH tokens in 2024. That single gain was equivalent to about 34% of 2024 reported revenue and 62% of 2024 EBITDA, making it the dominant reason the reported year-on-year comparison looks far weaker than the underlying trading operation.[1]

A reconciliation of the annual report separates three quantities: cash proceeds from token sales, the gain recognised on those disposals and management’s adjusted measures of ongoing trading performance. Confusing those quantities would exaggerate either the strength of 2024 or the weakness of 2025. The company itself identifies the token disposal as a principal reason for the reported decline.[1][2]

The reported comparison

LMAX reported 2025 revenue of approximately $103 million, down from $184 million in 2024. Reported EBITDA fell to about $22 million from $101 million. On those rounded figures, revenue declined 44.0% and EBITDA 78.2%. Those are accurate reported movements, but they are not like-for-like measures of the recurring trading business because 2024 included the realised PYTH gain.[1]

The annual report attributes the movement principally to $62.5 million of gains realised on PYTH inventory in 2024. The financial statements also present the transaction through sterling cash-flow and accounting lines, including proceeds and a gain translated into the group's reporting currency. Those sterling figures should not be mixed mechanically with the dollar performance summary. The clean comparison for the headline calculation uses the annual report's dollar-denominated $62.5 million explanation against the dollar revenue and EBITDA figures.[1]

Dividing $62.5 million by 2024 revenue of $184 million gives 33.97%. Dividing it by 2024 EBITDA of $101 million gives 61.88%. The second ratio does not mean the token transaction generated exactly 61.88% of every component of EBITDA; it shows the scale of the disclosed gain relative to the reported total. Remove the gain arithmetically from the 2024 revenue headline and the result is about $121.5 million, but that is only a bridge, not a formal pro-forma income statement. Tax, inventory valuation, associated costs and classification effects mean it should not be described as audited “revenue excluding PYTH”.[1]

The adjusted business moved in the opposite direction

LMAX’s alternative performance measures show a steadier operation. Its results announcement reported adjusted gross revenue of $112 million, up 4%, and adjusted EBITDA of $25 million, up 11%. The annual report records adjusted net revenue of $108 million against $107 million, an increase of about 1%. Gross revenue, net revenue and EBITDA measure different stages of the earnings calculation and should not be interchanged.[1][2]

The adjusted reconciliation is not a neutral law of nature; it is management's framework for separating trading performance from inventory valuation and other items. It should therefore be presented alongside, not instead of, the reported accounts. In this instance, however, the rationale for the largest adjustment is unusually transparent: a token disposal created a large realised gain in one period and was not repeated in the next.[1]

How the token inventory arose and changed

The annual report records a substantial PYTH position. At the end of 2025, the group held 324,631,904 PYTH tokens, of which 125,150,000 were restricted. It says no PYTH tokens were sold during 2025. In 2024, by contrast, 143,733,062 tokens were sold for $75.6 million, producing the disclosed gain that transformed the prior-year comparison. The remaining inventory continued to expose the accounts to token-price valuation effects even in a year without disposals.[1]

This matters because LMAX is both a market-infrastructure business and a holder of digital assets received through commercial or strategic arrangements. A token inventory can create earnings that do not scale with client volumes, spreads or transaction fees. It can also generate impairment or valuation movements unrelated to the operational quality of the exchange and broker businesses. Investors, counterparties and competitors therefore need two views at once: reported economic results including the token outcome, and a reconciliation of recurring marketplace performance.[1]

A stronger trading year beneath a weaker reported year

LMAX reported record trading volume of $8.2 trillion for 2025 and said cumulative volume had passed $50 trillion. Those operating measures provide context for the earnings comparison, but they do not erase the reported decline or prove a proportional improvement in revenue. The revenue and EBITDA reconciliations are still needed to distinguish the business measures from the token-disposal base effect.[2][1]

The evidence supports a narrower reading than a collapse in the core franchise: an exceptional 2024 disposal gain makes the reported year-on-year comparison much weaker than the adjusted operating comparison. Management’s adjustment is useful, but it remains a defined alternative measure rather than a guarantee of recurring earnings. Its exclusions and reconciliation still matter.[1][2]

What cannot be concluded

The accounts do not support calling the 2024 result artificial. The tokens were sold and the gain was realised. Nor do they support treating adjusted EBITDA as cash available to shareholders, because adjusted measures omit items management considers non-recurring or non-operational and can differ from cash generation. The analysis also cannot predict the value of the remaining PYTH inventory. Token prices are volatile, restrictions affect liquidity and the group's accounting treatment may produce future gains, losses or impairments.[1]

It would also be misleading to describe the $62.5 million as “token revenue” without qualification. The annual report's presentation spans inventory gains, disposal proceeds and adjusted-revenue reconciliations. The safest formulation is the company's own: a realised gain on PYTH inventory that was the principal driver of the year-on-year reduction in reported revenue and EBITDA.[1]

Why the inventory accounting can dominate a venue's earnings

A trading venue normally earns recurring income from transactions, data, connectivity, financing or related services. Token inventory introduces a different return stream: value can arise from receiving an asset, remeasuring it, impairing it or selling it. The timing of those events can make one year's revenue and EBITDA depend more on a token market than on client trading. That is exactly why the 2024 and 2025 reported totals need a bridge.[1]

The remaining 324.6 million-token position also means the exceptional item was not wholly historical at the end of 2025. Restricted tokens could become transferable later, market prices could change and accounting assessments could produce future volatility. No sale in 2025 removed the realised-disposal effect, but it did not remove the balance-sheet exposure.[1]

Volume growth does not translate mechanically into revenue

Trading volume and adjusted revenue do not have to grow at the same rate. Product mix, client mix and the amount charged for different flows can change. The disclosed 28% increase in notional trading volume and 4% rise in adjusted gross revenue therefore cannot establish a change in any particular customer’s fee without a matching product-level breakdown.[2]

The divergence is not necessarily negative. Institutional venues can lower unit pricing to win larger flows, add low-yield products, deepen liquidity or expand relationships that create future revenue. But record volume alone should not be used as proof of proportional earnings growth. The 2025 results show that scale and monetisation moved at different speeds.[2]

Testing both interpretations

The most sceptical interpretation is that removing the token gain flatters a business whose reported revenue still fell and whose adjusted metrics are management-defined. The answer is not to discard the reported result; it is to present both. Even after removing the $62.5 million gain arithmetically, 2024's simplified revenue bridge remains above 2025 reported revenue, and adjusted growth is modest relative to volume growth. The underlying story is steadier, not uniformly strong.[1]

The opposite exaggeration—that the core franchise lost almost four-fifths of its earnings—is also unsupported. The 78% EBITDA decline compares a normal trading year with a year containing a gain equal to nearly two-thirds of EBITDA. A fair headline must identify that base effect at the same prominence as the reported fall.[1]

Restricted inventory makes headline market value an incomplete measure

The annual report's distinction between restricted and unrestricted PYTH matters because a token that cannot yet be transferred is not economically identical to one that can be sold immediately. Restrictions can delay monetisation, increase concentration and make quoted market prices less representative of realisable value. The accounts' impairment and valuation policies therefore matter as much as the spot token price.[1]

A reader should also avoid multiplying all 324.6 million tokens by a single market price and calling the result cash. The position may be subject to release schedules, liquidity limits, tax and execution impact. The 2024 disposal itself shows that realised proceeds and accounting gain are distinct measures. Future reporting should separate restricted quantities, released quantities, disposals, realised gains and valuation movements so the operating business is not obscured again.[1]

The disclosure also creates a repeatable reporting test for 2026. Analysts should reconcile token disposals, realised gains, impairments and restricted balances before comparing revenue or EBITDA with 2025. If another disposal occurs, the same base-effect problem will return; if none occurs, adjusted performance should converge more closely with reported operating trends. Publishing a clear token bridge each year would allow readers to separate marketplace execution from digital-asset inventory without treating either as less real.[1]

This approach also protects against the reverse distortion: excluding every token-related result as though it were irrelevant. LMAX operates in digital assets, so token economics can be part of the business. The editorial task is to distinguish recurring execution income from inventory outcomes, not to erase either category.[1]

LMAX's results demonstrate a wider disclosure problem for trading groups that receive or hold digital tokens. A token grant or inventory position can become large enough to overwhelm conventional measures of operating progress. A reader who stops at the reported percentage change may miss a growing marketplace; a reader who uses only the adjusted figures may ignore a genuine economic exposure. The two must be reconciled explicitly.[1]

The cleanest statement of the 2025 result is therefore not that LMAX's business collapsed. It is that an exceptional 2024 PYTH disposal gain—equal to about 62% of that year's EBITDA—did not recur, causing the reported comparison to fall sharply while adjusted net revenue and adjusted EBITDA increased. That formulation preserves the realised gain, the underlying improvement and the continuing token exposure without conflating them.[1][2]

Report an error or submit an update