Giant X's Exclusive iTero Deal: AI Coaching and Esports' Unwritten Rule

The night Na'Vi lifted the Aegis of Champions at Gamescom 2026 is the reference point the Jack Williams interview chooses for itself. Fourteen years later, the memory of that trophy is intact, along with the 1,000,000 USD first prize claimed by the winners of the inaugural The International, when Dendi, Puppey and their teammates wrote the opening pages of professional Dota 2 history. But what is being argued about now does not sit on the stage. It sits in a software file beside the coach's bench: iTero, an artificial-intelligence coaching tool currently used exclusively by the organisation Giant X.

I read that interview three times over one night in Hai Phong. Each time I stopped at the two section headings it discloses: one on working with Giant X exclusively and the likelihood of being copied, one on AI-assisted cheating. Two topics, two names, and between them a gap nobody has named: fairness of access among teams inside a closed league.

Context: an exclusive arrangement in a league without relegation

The interview centres on Jack Williams, the figure attached to iTero, and Giant X, an organisation widely known within the EMEA ecosystem and operating inside a regional league run by Riot Games. The article's content is available only at heading level: exclusive partnership and the risk of copying; cheating with AI assistance. There is no bracket, no format, no patch version, no win rate of any kind.

In other words, this is a narrow B2B thought-leadership piece, not tournament reporting. Anyone looking for a group-stage prediction will come away empty-handed.

My trade is not prediction. I administer the transfer market, specialising in esports, and I am long accustomed to a quiet commercial agreement reshaping the competitive floor before anyone names it. Nights in Hai Phong taught me one thing: people watch the price board, I watch the movement board.

An AI coaching tool exclusive to one team is not the same as a sponsorship patch on a jersey. Sponsorship does not change results. An analytics tool can, if it genuinely works, and if it genuinely belongs to one side only.

Core: patch cadence decides the value of every model

The first thing worth separating is update cadence across major titles. Valve runs Dota 2 on large, infrequent, structure-breaking patches, with long stable stretches between shocks. Riot runs League of Legends on a two-week rhythm, small but constant.

That difference is not a technical detail. It is a first-order commercial variable.

In a slow-patch title, an AI model trained on historical data retains validity across long windows. The edge lies in depth of historical modelling. In a title patched every two weeks, the half-life of any learned pattern is far shorter. There, AI value shifts from solving the meta to detecting the meta delta faster than opponents, a speed advantage rather than a knowledge advantage.

A product promising equal effectiveness in both environments should raise questions. I have spent hours with datasets whose owners claimed universality. They were never universal.

Three things are missing before any performance claim can be assessed: the specific patch cadence, the tournament-server lock rules, and the permitted data window. Without those three, any accuracy figure is decoration.

I remember June 2026, presenting an analysis of Hai Phong striker Rimario Gordon: 14 matches, 0.32 xG per match, lowest among ten V.League foreign forwards. I predicted five goals that season. He scored exactly five. But what I took away was not that I was clever. It was that a model is only right when I know exactly what it measures, on which sample, over which window.

The Jack Williams interview supplies no sample. No evaluation methodology. No dataset size. That is the first gap.

The mechanics of exclusivity and the penalty of a closed structure

The second point matters more to me. A closed league with permanent member slots and no relegation creates a different kind of asymmetry from an open circuit. In an open system, structural advantages tend to be competed away: weak teams drop, strong teams rise, distortions flatten out. In a closed league, a structural advantage is not competed away. It persists season after season.

An exclusive analytics deal sits precisely in that box. If the tool creates a material difference, it is not a commercial gift. It is an institutionalised preparation advantage.

Esports has walked this road once before, at a different layer. In-game coach communication was once free, then progressively tightened, then strictly limited, because organisers realised an open information channel could change outcomes. AI tooling does not operate in-game, which has been clearly prohibited in every major title for years. The genuine grey zone is elsewhere: the window between games, inside a BO3 or BO5.

Thirty seconds between game two and game three is a window in which any model can act. Nothing in current rules was written to handle that window transparently.

From the shock of Germany leaving the 2026 World Cup, I learned this: respect the model, never trust it absolutely. I wrote that Germany would reach the semi-finals, based on 67 percent possession, 2.1 xG and 91 percent pass accuracy. Germany lost their opener to Mexico, then were eliminated by South Korea on 27 June. My model did not account for pitch temperature, Mexico's high press, or the psychology of the reigning champion.

Applying that lesson to iTero, I do not ask whether the tool is strong. I ask who verifies it, how, and for how long.

Contrarian angle: the missing frame is neither copying nor cheating

The two section headings frame the issue along two axes: commercial and integrity. Copying is the seller's story. Cheating is the player's story. Both skip the party in the middle, the tournament operator.

If a tool materially affects competitive outcomes, operators will soon have to choose: either mandate equal access for all teams, or restrict the tool. No third option holds for long. I have watched models collapse before, and I know structural penalties always arrive later than the argument.

I also have to be honest about my own limits. Correlation is not causation. A team using an AI tool and then winning proves nothing, unless the tool can be separated from squad quality, schedule, and luck. Without a control group, there is no conclusion.

In May 2026, when the Bundesliga returned to empty stadiums, I compared 26 rounds with crowds against nine without. Home advantage fell 15.3 percent, yellow cards rose 22 percent, and away-team PPDA dropped from 11.4 to 9.8. Empty stands taught me I had miscounted a variable: emotion is not in the spreadsheet.

An AI tool can compute PPDA. It cannot compute a shaking hand in the 89th minute.

Euro 2026 taught me the same from the opposite direction. I predicted Belgium would win because they had the highest total xG. Italy won, with a PPDA of 8.7, lowest of the 24 teams. I missed the defensive metric by focusing on attack. The chart does not lie, but it does not tell the whole story. I look for the missing part.

Takeaway: the signal to watch in the next cycle

What I am waiting for is not a statement from iTero or Giant X. I am waiting to see whether an operator of a closed league is forced to disclose exclusive tooling arrangements. If that happens, esports will have its first transparency standard for a competitive layer that has never been legislated. My numbers do not need applause. They need to be right. Time is the referee.