G2E 2026 round-up: The industry’s new edge is player data. Most of its value is still undiscovered.

The Neuralift team just spent three days exhibiting at G2E 2026 in Las Vegas, where we had the opportunity to meet and talk with land-based, online and hybrid operators from across the US, Latin America and Europe.

Four themes came up again and again during the show: AI’s increasing prominence, the cost of fragmented player data, a renewed focus on responsible gaming, and the rise of prediction markets.

What links all four is player data. Operators already hold more of it than they use. Newer gaming systems and machines are built with data collection in mind. Those who learn from it fastest will be the ones to stay ahead of what’s next.

Fragmentation is a huge margin problem

Operators from around the globe came to the Neuralift stand over the three expo days, each with different versions of the same problem. Their player data is spread across systems that weren’t originally designed to talk to each other.

The root cause of this is history. Casino management and loyalty systems were built for the floor as it was decades ago. Hotel, dining and retail each often run on their own platforms. Online sportsbooks and iGaming arrived much later, bringing with them more modern stacks with their own player IDs or partnerships. Adoption and roll out of each was the right decision at the time, but as a whole they have led to one player looking like four or five different customers across the piece.

The State of AI in Gaming 2026, a recent report from the UNLV International Gaming Institute in collaboration with KPMG, surveyed 83 gambling companies. It found that land-based operators trail online operators on AI by a statistically significant margin, pointing to legacy systems and databases, and the sheer number of systems an integrated resort has to run, as primary drivers.

Operators were candid with us about what this is costing them:

• Bonus duplication. A player receives free play from the casino, a deposit match from the online team and a comp from the hotel, each funded from a different budget with no shared view of total reinvestment.
• Attribution that doesn’t hold up. When a player visits the property after an online campaign, neither team is able to say with confidence what drove the trip, so budgets are following assumptions as opposed to evidence.
• Personalization misfires. For example high-value table players that receive generic slots promotions online. Players notice this, and the risk is that your brand starts to feel more like several companies just sharing a logo.

Each of these leaks margin, and more importantly each chips away at the player experience that modern loyalty programs have been designed to protect.

But some operators are already ahead of the curve here — our own work with Delaware North is well documented. They have invested in using Databricks to build a unified view of players at a central level, on top of which they apply deep learning. This means that player segmentation lives across the whole estate instead of only inside individual properties or channels.

It shows how far the value can go, and the operational and revenue upside of tackling fragmentation: with one view of the player feeding deep learning, segments are able to resolve bonus cannibalisation, fuel greater personalization for key customer tranches like VIPs, and greatly improve measurement.

Operators don’t need to finish a multi-year data project to start on the journey. Discovery should begin with the data already available, whether that is single property, online brand or a loyalty extract delivered as a flat file. From there each data source that is added later makes the player model more valuable.

You would be hard pressed to find a vertical with more data than gaming. However it is the value not the volume that really distinguishes player data. Behaviors, transactions, locations, temporal and time-based patterns not to mention categories, games, bonuses and profitability all weigh heavily on operator growth. Bringing that data together at the ID level so intelligent decisions can be made from it is the first-step to better player relevance and operator growth.

AI takes over the show floor

2026 was the first year that G2E gave enterprise technology its own space at the show, with a new Business Solutions & Tech Zone sat alongside the usual buzz and lights. In this zone AI was prominent on almost every stand — ours included.

Recent data suggests that the gaming industry is earlier on its AI journey than the prevalence suggests. The aforementioned State of AI in Gaming report scored the industry at 45 out of 100 on AI maturity. Technology, security and product innovation account for nearly half of all AI activity. Customer-facing work sits further back, and marketing use cases such as acquisition, CRM and retention register at roughly the same level as fraud and player safety.

On stage, however, the ambition was clear. Boyd Gaming’s Eric Bowers described a “dream scenario” in which every player in the database gets their own AI host. Simo Dragicevic of IGI’s AI Research Hub talked about turning a static player database into a living one that models how players actually behave. We agree on both fronts.

Delivering on either won’t be turnkey while, per our previous point, an operator’s view of each player remains split across four or five systems. But operators already hold an unusual breadth and depth of player data, much of it thanks to years of investment in loyalty. Deep learning can work across all of those attributes at once and surface player segments that fixed rules and manual analysis would never think to define. Refreshed regularly, those segments start to look a lot like the living player database Dragicevic described. They give an AI host or personalization engine current behavioral evidence to work with. To act on that evidence, the AI also needs agreed business definitions and clear rules for which actions are appropriate. Together, these ground its decisions in the operator’s business context and player protection requirements.

AI was everywhere at G2E, but nearly every conversation came back to the same problem: player data that’s disjointed across multiple systems with no source of truth. That’s the gap that Neuralift and our product Alkemi are closing together, and why we’re so excited about our strategic partnership.

Responsible gaming took center stage

One theme defined September for this industry — player protection. The American Gaming Association marked its 28th Responsible Gaming Education Month, and the AGA’s David Forman estimates the industry has committed around half a billion dollars to responsible gaming programs. Operators backed it with action. Ahead of the show MGM Resorts and BetMGM contributed $320,000 to responsible gaming initiatives and took GameSense messaging into more stadiums. Hard Rock and Seminole Gaming expanded PlayersEdge with a new awareness campaign, a digital health app pilot and research partnerships. States from New York to Arizona ran their own campaigns.

The science came to Las Vegas too. The ICRG Conference on Gambling and Addiction ran at the Venetian Expo on the eve of G2E, with sessions on AI and data science for player protection and on detecting loss chasing from behavioral data. Researchers, regulators, treatment providers and operators all in the same room.

Regulators appear to be moving in the same direction. The UNLV and KPMG report found the most common regulatory action on AI in gambling is the use of automated tools for player harm detection, with jurisdictions including Belgium, Finland, Italy and several US states and Canadian provinces now mandating or encouraging automated monitoring. The same report is transparent about where the industry actually stands though. Tracked AI incidents include failures of harm detection systems and targeting that was too aggressive, and just 2% of companies say responsible AI is embedded throughout their organization.

On the show floor, the shift to responsible gaming was hard to miss. This year, no shortage of vendors led with use cases in this area, ahead of traditional player value and optimization stories that would normally headline.

We see the pent-up demand for this at Neuralift.

Operators want to use deep learning to find their best players, but in the same breath are asking how AI models can find the players they should leave alone, or perhaps even help. The data fragmentation problem comes back into play here just as much as it does for marketing and CX. A player chasing losses across a land-based property and an online account may look unrelated in each system on its own. Legacy segmentation rules are only catching the patterns that human teams have pre-defined, and risk allowing problem gaming through the cracks due to operational silos.

We believe the right path forward is clear:

• Operators need regularly refreshed, risk-indicative segments that are derived from their full estate of 1st party player behavioral data.
• These segments should go to player protection teams, with a mandate to review them and decide what happens next.
• The most obvious first action is suppression: pulling those players out of promotional activity.
• Support and outreach needs to be designed by people with responsible gaming expertise and focus, and be informed by explanations of why each segment formed, so nothing is a black box.

We encourage every operator working with Neuralift to put player protection at the top of its list of use cases. AI that finds your most valuable players should also help you look after the ones at risk.

Prediction markets: parity is the bare minimum

Nevada regulators have argued that prediction market operators are offering sports betting without complying with the state’s gaming framework, and a G2E main stage session brought tribal gaming leaders together to confront the challenge directly.

The concern from states, tribes and licensed operators is understandable. They have spent years building regulated markets, with licensing, age and identity checks, responsible gaming tools, self-exclusion and integrity monitoring. Watching a parallel market grow quickly under a different regulator, with a lighter set of obligations, feels unjust to many of them.

Washington is paying attention. In April, Senators Gillibrand and McCormick introduced the bipartisan Prediction Market Act of 2026, which would direct the CFTC to define insider trading standards specific to prediction markets. A separate bipartisan group led by Senators Curtis and Slotkin has introduced a bill to stop government officials trading on insider information. Others in Congress want to go further and keep sports contracts off these platforms altogether.

Consumers are adopting prediction markets for clear reasons. The products are simple to use, the range of markets is broad, and they offer a different way to engage with sport and events. We heard from operators that some players are now using prediction markets alongside traditional sportsbooks, including to hedge positions. The market is dividing, with some established operators exploring the space and others pushing hard against it.

Neuralift has no stake in picking a winner here, but we feel that whatever the outcome, regulatory parity should be the minimum. By this we mean that transparency, clear rules on insider information, and consumer protection should match what regulated operators already provide.

There is a potentially interesting angle here for regulated operators though. Their existing obligations produce a data asset that prediction markets currently lack: a long, verified, and consented history of how players behave across products. If this is used well, it can be a competitive advantage.

The stage is set

Gaming’s data complexity is also its opportunity.

Yes, when compared with other industries gaming data is challenging to unify. A single operator might run multiple casino floors, hotels, restaurants, retail sportsbooks, online casinos and loyalty programs, each producing its own data and each governed by individual regulatory requirements, often in several jurisdictions at once.

But that width of operator footprint also creates a breadth of player behavioral data that few other industries are able to match.

The operators tackling this challenge head-on and embracing AI capabilities like deep learning are finding revenue and margin upside, and outpacing their peers.

Neuralift will be back for G2E 2027, but you can catch us first at ICE Barcelona, 18 to 20 January 2027.