Live casino is arguably online gambling's most prized product category – a multi-billion dollar industry built on a deceptively simple proposition: the trust and theatre of a real dealer, delivered through a screen.
But that proposition is now being challenged from two directions simultaneously, and the companies leading each charge have very different ideas about what "the future of live casino" actually looks like.
On one side, a wave of AI-native startups are betting that large language models and real-time digital avatars are ready to replace human croupiers entirely. On the other, a new breed of supplier is arguing that full AI isn't quite there yet – and that the smarter move is to clone the human, not simulate one.
It is a fork in the road that will define the next chapter of live casino supply.
The AI-first camp
Arguably the most high-profile flag-planter in the AI-dealer space right now is BetHog, the crypto casino founded by FanDuel Co-Founders Nigel Eccles and Rob Jones.
Last week, the company announced a $10m Series A funding led by Will Ventures and RockawayX, bringing its total funding to $16m, alongside the launch of Sentient Studios – a B2B platform that allows casino operators to deploy AI-powered live dealers across their own products.
The platform operates on a pure revenue-share model with no setup fees, monthly minimums, or fixed-term contracts, a deliberate effort to lower adoption barriers for operators cautious about committing to an unproven format.
The commercial rationale is compelling on paper. BetHog has been running its own AI blackjack dealer, Sunny, since October 2025, and the internal data is striking: in a Huddle interview earlier this year, Eccles said the company's AI dealer proved 10 times more popular than its live dealer equivalent, with better retention and player satisfaction scores to match.
Whether that reflects genuine player preference for AI or simply the novelty effect of a new product in a relatively niche crypto casino remains an open question – but it is a data point that will attract operator attention.
BetHog is not alone. Playgon Games recently signed a definitive agreement with real-time animation studio Digital Nation Entertainment to co-develop an AI dealer platform, with initial operator deployments targeted for Q3 2026.
Its pitch is similar: AI dealers that operate as real-time, responsive hosts capable of natural conversation, emotional adaptation and continuous interaction, without the scheduling constraints and studio overheads that define traditional live casino supply.
Meanwhile, QTech Games has already partnered with Sentient Gaming Group to bring an AI roulette product – featuring a virtual croupier named Amanda – to operators in emerging markets.
In other words, the market is moving fast. The AI-dealer pitch to operators is essentially a software economics argument: infinite scalability, no studio rental, no staffing, no sick days and the ability to localise into any language at the flip of a switch.
The moment a player perceives something as artificial or uncanny, the trust relationship that makes live casino valuable in the first place can begin to erode
Can players be won over by an AI dealer?
The software economics case for AI dealers is a strong one and, in certain markets, it becomes particularly compelling. For operators eyeing Brazil's newly regulated market – where, as Avanti's Jonas Delin has noted, Portuguese localisation will be legally mandated – the prospect of instantly scalable, multilingual AI dealers is far more attractive than recruiting and training Portuguese-speaking studio staff at pace.
The same logic applies to operators trying to serve low-stakes recreational players who simply cannot justify a $25 minimum blackjack table.
But operator appetite and player engagement are two different things. The harder, and arguably more important, question is whether players will genuinely trust what they are playing against – and that question sits at the heart of the entire debate.
The fact that BetHog is now pivoting its technology into a B2B product, and that competitors such as Playgon and Sentient Gaming Group are building comparable offerings, tells us the supply side is convinced.
Nobody has yet proven, however, that AI dealers can sustain player loyalty at scale outside of BetHog's own – still-nascent – ecosystem.
The counter-argument: AI isn't ready
Not everyone is convinced that the underlying technology is there yet, and the most articulate voice making that case is Avanti Studios, the live casino startup launched earlier this year by Gustaf Hagman and Jonas Delin.
Avanti's product is genuinely novel, but it resists the AI label. Rather than generating dealers computationally, the company brings professional land-based croupiers to a motion capture studio in Stockholm, records their movements and mannerisms in forensic detail, then uses Unreal Engine and the MetaHuman toolset – technology borrowed from the AAA video game industry – to construct photorealistic digital clones. The result is streamed live, with cards dealt in real time against a genuine RNG outcome.
Hagman is unambiguous about the reasoning. As he put it in a recent interview with Global Gaming Insider, "AI is not quite there yet" when the requirement is dealers that are "ultra realistic, completely human-looking."
The emphasis on lifelikeness is not aesthetic preference – it is a product philosophy. Avanti's argument is that the moment a player perceives something as artificial or uncanny, the trust relationship that makes live casino valuable in the first place begins to erode
Delin makes a related point about the structural integrity of the live format itself. Real-time dealing – where multiple players at a blackjack table each make independent decisions against a shared dealer hand – is not easily replicated by a pre-rendered or AI-generated character. The cards have to be dealt in the present tense, not simulated. That constraint puts a meaningful ceiling on what pure AI can deliver, at least today, for the games where player agency matters most.
The AI-dealer pitch to operators is essentially a software economics argument: infinite scalability, no studio rental, no staffing, no sick days and the ability to localise into any language at the flip of a switch
Is there a regulatory case for going digital?
There is also a regulatory dimension worth watching. Avanti's use of a cryptographically strong RNG as the basis for game outcomes – rather than physical cards handled by a human – may give it an unexpected advantage in markets where live casino regulation is evolving.
In Spain, for example, where roulette is currently the only live game permitted, Avanti believes its product may qualify regulators to expand that allowance, because the "dealer layer" is technically a presentation rather than a live game in the traditional sense.
So… Which model wins?
The honest answer is that it is too early to call – and quite possibly both can coexist. They are, to some extent, solving different problems.
The AI-dealer model addresses the economics of live casino supply with a brute elegance: eliminate the human, own the software, share the revenue. For operators in markets where localisation is mandatory but costly, or where table volumes are too low to justify traditional studio contracts, it offers a genuinely new commercial relationship.
The fact that BetHog is now pivoting this into a B2B product, and that Playgon and Sentient Gaming Group are following suit, suggests a real operator appetite exists.
Avanti, by contrast, is playing a longer game – betting that the ceiling on AI realism will continue to limit the format's reach, at least for the foreseeable future, and that operators serving high-value, discerning player segments will not want to compromise on visual fidelity.
It is a bet that live casino's premium positioning depends on the suspension of disbelief, and that disbelief is harder to suspend with today's AI than with a motion-captured clone rendered in Unreal Engine.
What the divergence reveals, ultimately, is that "live casino" is not a single product – it is a spectrum of player expectations, price points and regulatory environments. Let's not forget a third option still exists and may likely continue for eternity: the trust and luxury of the old-fashioned, purely human model.
The AI-native model may dominate in volume-driven, emerging markets. The digital-human model may hold sway where product quality and brand reputation are the competitive currency. Here, we can also consider the mass market versus VIP, high-yield players.
The interesting question, as AI technology continues to improve at pace, is how long that distinction holds.
Brazil’s regulated iGaming market introduces strict compliance and localisation requirements, meaning operators must tailor their offering to local players – a challenge that scalable digital or AI-driven dealer models may be increasingly positioned to solve
