A chart-led census

The game supply divide: which studios serve licensed casinos, and which do not

Sweden, Denmark, and Romania have written into law that licensed game studios may supply only licensed operators. This is what that looks like in the supply graph: a census of 320,187 casino-to-studio relationships across 353 studios.

By Antti Virtanen , Lead Analyst Published 31 July 2026 6 min read

In short

A census of 320,187 casino-to-studio relationships. 43.5% of all supply would have to move to make licensed and offshore casinos look alike, and some studios serve hundreds of licensed brands and not one offshore brand.

Segregation

43.5%

of supply relationships would have to move to make the two shelves alike

The dissimilarity index across all studios. It barely moves under any control we applied.

The zero column

Zero

offshore brands served by Light & Wonder or Bet Digital

Between them, they supply 363 licensed brands. Neither supplies a single offshore casino in the dataset.

Catalogue size

72 vs 31

studios carried by the median offshore vs licensed brand

Offshore casinos carry more than twice the catalogue, and a far more interchangeable one.

Commercial effect

Zero

measurable relationship between catalogue size and conversion

Within-stratum correlation of -0.04. A bigger game library does not sell better.

The map

The 54 largest of 353 studios are placed by how much of each shelf they reach. The empty diagonal is the finding.

Marker size is proportional to the number of brands a studio serves. Not every point is labelled.

Supply has split into two shelves. Studios cluster along the edges, not the diagonal. The corners are occupied by companies that appear on hundreds of casinos on one side and almost none on the other.

Show data table Hide data table

Studio	% of licensed brands	% of offshore brands	Brands served
NetEnt	86.8%	78.8%	1304
Pragmatic Play	78.4%	95%	1278
Play’n Go	76.4%	79.3%	1189
Evolution	74.9%	81.2%	1181
Red Tiger	65.8%	82.6%	1085
Hacksaw Gaming	56.7%	92.1%	1023
Big Time Gaming	66.9%	61.9%	1010
Nolimit City	60%	77.9%	1000
Yggdrasil	56.3%	86.9%	996
Thunderkick	58.6%	78.6%	987
Quickspin	58.1%	71.7%	952
Microgaming	64.2%	53.1%	943
Playson	47.5%	91.2%	916
Elk Studios	60.7%	44.8%	869
Relax Gaming	50.6%	69%	857
Betsoft	36.5%	93.3%	801
Booming Games	36.8%	86.7%	777
Blueprint Gaming	50.6%	41.4%	741
Spinomenal	32.5%	89.3%	739
Spribe	31.3%	85%	708
Tom Horn Gaming	28.8%	79.5%	657
Push Gaming	45.1%	34.8%	652
Endorphina	25.4%	86%	646
BGaming	22.7%	89.8%	631
Playtech	30.9%	63.8%	614
RubyPlay	25%	78.6%	610
Fantasma	39.9%	34.8%	593
Novomatic	19%	89.3%	588
Habanero	23.2%	72.6%	565
EvoPlay	14.6%	87.1%	530
Amatic	18.7%	71.7%	511
Inspired Gaming	44.2%	0.7%	499
Foxium	36.8%	17.6%	486
Just For The Win	40.1%	7.1%	479
FUGASO	10%	83.6%	463
Lightning Box	39.5%	4.5%	462
GameArt	9.5%	77.9%	434
NextGen Gaming	36.6%	1.4%	416
Gamzix	3.9%	86%	405
Platipus	4.8%	78.1%	382
3 Oaks Gaming	3.8%	80%	379
Lucky Streak	6.7%	69.5%	367
Mancala	2.2%	71.9%	327
Popiplay	2.4%	71.2%	326
Felix Gaming	1.9%	72.4%	325
Vivo	3.1%	64.3%	305
Barcrest	21.9%	0.2%	246
WMS	19.6%	0.2%	221
Slingo	18.8%	0.5%	213
Light & Wonder	16.9%	0%	190
Bet Digital	15.4%	0%	173
Expanse Studios	0.5%	37.9%	165
Just Slots	1.5%	34%	160
Pascal Gaming	0.2%	36.4%	155

Source: Kasinohai casino dataset. Exclusive licence classification; brands holding both or neither are excluded..

The two corners

• Share of licensed brands
• Share of offshore brands

Penetration of each shelf, licensed-only studios

Nine studios reach between 15% and 44% of licensed brands while reaching under 8% of offshore ones. Two of them reach zero.

Show data table Hide data table

Studio	% of licensed brands	% of offshore brands
Inspired Gaming	44.2%	0.7%
Just For The Win	40.1%	7.1%
Lightning Box	39.5%	4.5%
NextGen Gaming	36.6%	1.4%
Barcrest	21.9%	0.2%
WMS	19.6%	0.2%
Slingo	18.8%	0.5%
Light & Wonder	16.9%	0%
Bet Digital	15.4%	0%

Source: Kasinohai casino dataset.

• Share of licensed brands
• Share of offshore brands

Penetration of each shelf, offshore-only studios

Pascal Gaming reaches 36% of offshore brands and 0.2% of licensed ones (two brands out of 1,121). The asymmetry runs in both directions.

Show data table Hide data table

Studio	% of licensed brands	% of offshore brands
Gamzix	3.9%	86%
FUGASO	10%	83.6%
3 Oaks Gaming	3.8%	80%
Platipus	4.8%	78.1%
GameArt	9.5%	77.9%
Felix Gaming	1.9%	72.4%
Mancala	2.2%	71.9%
Popiplay	2.4%	71.2%
Expanse Studios	0.5%	37.9%
Pascal Gaming	0.2%	36.4%

Source: Kasinohai casino dataset.

• Headline measure
• Under control

Share of supply relationships requiring reallocation to equalise the two shelves Scale runs to 50%.

Matching for catalogue size changes the result by half a point. This is not an artefact of offshore casinos simply carrying more studios.

Show data table Hide data table

Control applied	Dissimilarity index
All studios, no control	43.5%
Matched for catalogue size	42.9%
One brand per operator	36.2%
Brands launched 2020 onwards	39.6%

Source: Kasinohai casino dataset. Operator deduplication keeps one brand per corporate group..

What each shelf looks like

• Licensed brands
• Offshore brands

Scale runs to 100 studios.

Offshore casinos carry more than twice as many studios as licensed ones at every point in the distribution.

Show data table Hide data table

Quartile	Licensed brands	Offshore brands
Lower quartile	16	49
Median	31	72
Upper quartile	57	84

Source: Kasinohai casino dataset. Licensed n=1,115; offshore n=419..

• Licensed brands
• Offshore brands

Two-thirds of an offshore casino's library is shared with most of its competitors, against a quarter for licensed brands. More games, but less differentiation.

Show data table Hide data table

Measure	Licensed brands	Offshore brands
Catalogue drawn from near-ubiquitous studios	24.2%	67.7%
Catalogue overlap between two random peers	0.37	0.54
Distinct studios serving the shelf	341	259

Source: Kasinohai casino dataset. Overlap measured as Jaccard similarity between the studio sets of two randomly chosen brands on the same shelf..

• Licensed shelf
• Offshore shelf

Game supply is one of the least concentrated layers in gambling. The top five studios account for under a tenth of relationships on either shelf, meaning the problem is segregation, not monopoly.

Show data table Hide data table

Measure	Licensed	Offshore
Herfindahl index	82	87
Effective number of suppliers	122	115
Top five studios’ share of relationships	9.6%	6.9%

Source: Kasinohai casino dataset.

Two industries, turning over separately

• Percentage points gained
• Percentage points lost

Movement in reach across licensed brands, by launch cohort

The licensed shelf has its own churn (Hacksaw up 39 points, Microgaming down 48) and it is almost entirely disconnected from what is rising offshore.

Show data table Hide data table

Studio	Gained	Lost
Hacksaw Gaming	+38.9pp	—
RubyPlay	+38.2pp	—
Fantasma	+30.9pp	—
Push Gaming	+21.8pp	—
Microgaming	—	−48.1pp
NextGen Gaming	—	−37.7pp
NYX	—	−32.2pp
iSoftBet	—	−31.2pp

Source: Kasinohai casino dataset. Percentage-point change in reach across licensed brands, earliest versus latest launch cohort..

Rank the studios by growth on each shelf separately, and the two lists barely relate. There was a Spearman correlation of 0.19 across the studios we could score. Knowing who is winning offshore tells you almost nothing about who is winning in the licensed market. The market-wide impression that the old studios are dying is mostly a composition effect: across all brands the classic licensed-market basket falls from 39.9% to 12.1%, but inside the licensed shelf alone it is flat, 51.3% to 50.9%.

Does any of it sell?

The apparent breadth effect, and the reporting artefact that produces it Scale runs to 50.

No relationship. The median within-stratum correlation is −0.04, and seven of 10 strata are negative. A bigger game library does not convert better.

Show data table Hide data table

Measure	Value
Pooled correlation, breadth vs conversion	0.032
Median correlation within market × tier strata	-0.04
Strata tested	10
Of which negative	7
Narrow-catalogue cells reporting zero conversions	33.1%
Broad-catalogue cells reporting zero conversions	11.7%

Source: Kasinohai casino dataset and Kasinohai observed-demand data. Brand × market cells with observed activity in 2025–2026. Conversion is first-time deposits per thousand clicks; no volumes are published..

Why the shelves are separate

The segregation is not an accident of commerce. Several European regulators have legislated it, and one has now enforced it.

Sweden has operated a two-way closed loop since July 2023: a licensed software supplier may serve only licensed operators, and a licensed operator may use only licensed software [5]. Denmark closed its loop in January 2025 [6]. Romania expressly prohibits licensed suppliers from serving unlicensed operators that target Romanian players [7]. The UK requires a software licence to supply gambling software by remote communication, wherever the studio sits [4].

In July 2026, the UK regulator secured a £4.75m payment in lieu of a financial penalty from Evolution because its games reached six unlicensed sites accessible to UK consumers, and used the case to state a general obligation on suppliers: know who you supply, know where your games are played, and monitor it continuously [1]. The company’s own reporting puts a price on compliance: European ring-fencing measures materially affected 2025 performance and, in its words, created a new baseline for growth [3].

A gap sits in the middle of this. Malta, where most studios are domiciled, does not treat its B2B licensees as subject persons under its anti-money-laundering law, which is the precise legal category the UK used to penalise a Maltese-licensed supplier [8]. The target market enforces the obligation, not the one in which the studio is registered.

Read against the figures above, the policy conclusion is unusually clean. Supplier licensing appears to be the most effective structural instrument in European gambling regulation: it has produced a measurable, robust separation in a market layer that no consumer sees and no consumer chooses. It has done so without depending on anyone reading anything, which, on the evidence of our previous study, is the only kind of rule that reliably works.

Method, controls and limitations

Dataset. 320,187 brand-to-studio relationships covering 353 distinct studios, captured on 2026-07-31. Relationships are recorded per brand per market; the analysis is at brand level.

Classification. “Licensed” means a brand holding an EEA or UK national gambling licence and no Anjouan, Costa Rica or Tobique licence (1,121 brands). “Offshore” means the reverse (420 brands). Brands holding both families, or neither, are excluded rather than assigned. That exclusion removes roughly a thousand brands, most of them Curaçao-licensed, and is the single biggest judgement call in the study. Including Curaçao as offshore would enlarge the offshore shelf considerably; we kept it out because Curaçao spans both regulated-adjacent and grey operations and would blur the comparison it is meant to test.

Segregation measure. The dissimilarity index is the share of all supply relationships that would have to be reallocated between studios for the two shelves to have identical supplier profiles. It runs from 0 (identical) to 1 (completely separate). Figure 4 recomputes it under three controls; the catalogue-size control is the important one, because offshore brands carry more studios and that alone could manufacture an apparent difference. It does not.

Naming. Game studios are named throughout. They are third-party suppliers whose presence on a casino is publicly observable by visiting the site, and naming them is necessary for the finding to be checkable. No casino brand or operator is named anywhere, and no monetary total, traffic volume or website-level figure is published.

Limitations

• Presence, not volume. We observe that a studio’s games appear on a casino, not how much they are played or what they earn. A studio on 400 brands may account for a trivial share of play on each.
• Studio names are not corporate entities. Several names in these charts belong to the same parent, with Light & Wonder, Barcrest, WMS and Bet Digital having shared corporate history, for instance. Treating them as independent overstates the number of distinct suppliers behind the licensed shelf, and the concentration figures in Figure 7 should be read with that in mind.
• Aggregators are invisible here. Much offshore content arrives through aggregators and turnkey platforms rather than direct studio contracts. Where a studio’s games reach a casino via an intermediary, our data records the studio, not the chain. This is precisely the chain the Evolution case turned on.
• Direction of causation is not established. We cannot tell from this data whether studios refuse offshore operators, offshore operators cannot afford or access them, or platform integrations simply differ. The regulatory record supports the first, but the data alone does not prove it.
• The null result is not an absence. Figure 9 shows no relationship between catalogue breadth and conversion in our data. That is not evidence that game content does not matter, only that the number of studios carried does not predict commercial outcome once market and licence tier are controlled.

Reuse

Any figure here may be reproduced with attribution and a link to this page. We are not aware of a previously published map of which studios supply which operators, so replication against other datasets would be particularly welcome. For the underlying aggregate tables, contact the author.