FT : AI drives a boom in new games but big developers dominate More instinctive

AI drives a boom in new games but big developers dominate
More instinctive technology is accelerating production amid concerns it risks losing gamers’ trust

When Stanislas Marchand was heading up game development for French mobile gaming unicorn Voodoo, his team tested 3,000 new titles per year. 

“It was an industrial process,” says Marchand, who was there until 2024. “Out of those, only one or two would go on to make $10mn-plus per year.” 

Those 3,000 ideas came from Voodoo’s internal teams as well as independent studios and, even after the advent of generative AI, productivity savings were relatively modest, he says. Pre-AI, it took about 14 days to design, test and get a mobile game ready for market. With AI, the team took 10 days. 

That appears at odds with predictions across many industries in 2026, as adoption of AI leads to grim forecasts about the future of work. Despite some lay-offs, the view of many in gaming is that AI will struggle to replace the central role of human creativity.

“What makes a good game is what we call ‘game feeling’,” he says. “So far, AI cannot do that,” Marchand says. 

Nonetheless, AI — and particularly the onset of “vibe-coding”, which allows people with little knowledge of software engineering to use AI to create concepts and games — has led to a massive increase in games coming to market. Global data from research company ATTN Economy shows that 181,000 games were released in the six months to May, 43,500 on iOS and 137,000 on Android — increases of 118 per cent and 73 per cent respectively year on year. 

This increase has come alongside some significant job losses in the gaming industry. Microsoft lost staff in its XBox division this month and cut 10 per cent in its Stockholm-based King division, famous for Candy Crush. A report by the GDC Festival of Gaming in May found that one in four employees in gaming had been laid off in the past two years. 

The report found that 52 per cent of gaming professionals thought generative AI was bad for the industry, up from 30 per cent in 2025 and 18 per cent in 2024. Workers in visual and technical arts (64 per cent) and game design and narrative (63 per cent) had the most unfavourable views. Only 7 per cent said that AI was having a positive impact. 

Some fans are also unhappy. The much-anticipated Crazy Taxi: World Tour was embroiled in online controversy over whether generative AI had been used in its development. This prompted Sega to acknowledge that it had utilised generative AI “as a support tool for developers [but] no AI was used in reference to performers in the game”. 

Those more bullish on generative AI see performers — or non-playing characters (NPCs) — as the most exciting use-cases. Where interactions with characters in role-playing and artificial world games such as Zelda or Call of Duty were heavily scripted, AI raises the prospect of true player-to-NPC dialogue. 

“Non-playable characters are already getting more powerful, but these things are about to become deeply personal. They’re going to remember you, and remember that you played this game two months ago,” says Josh Chapman, co-founder and managing partner at Konvoy, a venture capital firm that invests in early-stage gaming start-ups. 

Vlastimil Venclik, co-founder and chief executive of Czech gaming start-up Valka AI, agrees, and says there has been an exponential increase in the power of AI within games compared with two years ago. As models improve, so will storytelling and dialogue — and attitudes to AI may change. 

“I see some gamers are against it, but gaming studios are all using AI,” says Venclik. “I believe [their opposition is] just a trend which can fade away as fast as . . . it appears. In five, 10 years, my kids . . . probably won’t know the difference, right?”

Others are sceptical. ATTN Economy’s Lexi Sydow notes that data from Klaviyo, a digital marketing platform, shows that 13 per cent of consumers trust AI. When they see marketing content has been created with AI, they are four times less likely to trust that brand.

“There’s a cost to AI slop and it comes with a hefty price tag: trust,” Sydow says. “In gaming, it is like that on steroids. If it looks like it is AI-driven and not human-driven, it could fracture trust . . . if the storyline or the art drifts, people get really upset.”

The right approach will be key for the few large gaming conglomerates that dominate the industry. Sydow notes that the top 1 per cent of publishers controlled $75.6bn of revenue in 2025 with the other 99 per cent collectively making $6.1bn. That top 1 per cent saw 40.2bn downloads in 2025, accounting for nearly 80 per cent worldwide. 

Vibe-coding may have lowered the technical bars to entry in game development but Chapman predicts this will not break the big companies’ dominance. “The incumbents have such large balance sheets [and] are building up incredible talent and decades of data. That is really, really hard to compete with if you’re an up-and-coming start-up with a $4mn seed round.”  

From Slack workplace collaboration to Helsing military technology to Elon Musk’s SpaceX, gamers and gaming founders have played oversized roles in the wider economy. AI or not, that will continue to be the case. 

“The gaming market is the most competitive market in the entire planet . . . no other industry has 10,000 new [start-up] competitors a year,” says Chapman. “Gaming technologies proliferate into other markets and some of the world’s best technologies are being built by gaming founders.” 

FT : AI speeds the march of China’s factory robots into new sectors Artificial i

AI speeds the march of China’s factory robots into new sectors
Artificial intelligence is enabling the spread of automation to traditional industries

At a cavernous 100,000 sq m factory in the southern Chinese city of Changsha, dozens of automated guided vehicles whisk heavy loads between an array of giant robotic welding arms, machine tool beds and a small smattering of human workers, churning out a new concrete pump truck every 45 minutes.

In the factory’s space-age centre, manager Peng Yonghui can oversee each machine’s efficiency and state of repair from a giant screen flanked by blue-tiled fountains and palm trees. With a few clicks on a touchscreen, he can monitor machines and manufacturing sites across China.

Sany’s No.18 factory, opened in 2012, is one of the planet’s largest “lighthouse factories”, a World Economic Forum designation given to particularly smart manufacturing bases. With recent improvements in AI, the factory has continued to innovate, says Peng: “it makes us much more efficient”.

For instance, it previously took a number of humans to manually sort through the array of metal frames delivered to the factory’s eastern end. But with the advent of AI, the factory’s robotic arms can identify the shape and weight of the parts, decide which magnets are needed to lift them and feed them into machine toolbeds for processing, or drop them on automated vehicles to take them to the welding station.

“Originally, I had big parts, long parts, thick parts, small parts and short parts,” Peng says. Now, the system can not only discern size, weight and colour, but even the type of material. “Some materials are very similar, almost like twins, but they can still identify it.”

Policymakers in China are pushing automation and other smart factory technologies in the hope that they will enable the country to keep its manufacturing edge as its workforce ages.

But now, the rapid advance of AI technologies is improving the efficiency of production lines, boosting quality control and enhancing oversight. Factories, meanwhile, are deploying more intelligent robots on more difficult work. Others are preparing for the arrival of humanoid bots that could perform tasks previously reserved for humans.

Better AI models are enabling automation and smart systems to spread beyond the strongholds of large-scale auto manufacturing and other high-end fields into previously labour-intensive smaller factories in traditional sectors such as garments and footwear.


“First, the growth of AI supply-chain industries is generating direct demand for robots in manufacturing,” Samantha Mou, a senior analyst at robotics research consultancy Interact Analysis, wrote in a note this month.

“Second, AI-powered software is enhancing robot capabilities and ease of use, enabling robots to perform tasks they previously couldn’t, thereby creating demand in markets that were difficult to penetrate,” she added.

The introduction of AI has improved industrial robots’ visual guidance and made them easier to programme and better at co-ordinating with others. AI has also enabled remote monitoring and maintenance of robots, she wrote.

Robot vendors the world over are reacting. Japan’s Fanuc and Denmark’s Universal Robots have each announced collaborations with chipmaker Nvidia to work on AI-enabled programming tools. Switzerland’s ABB and Shanghai-based Step, meanwhile, have released new robot models designed for the electronics and semiconductor industries, to capitalise on new AI-driven demand.

Back in China, the huge manufacturing sector is driving demand.

“China is a mass market. [Its] industry reach and manufacturing scenarios [are] in the millions. So that generates . . . a real need for robotics,” says Jerry Liang, a partner at Beijing-based venture capital firm CCV.

Liang made early bets on a number of Chinese internet groups, including ecommerce retailer JD.com. But in recent years, his team has shifted its focus to AI-empowered robotics to capitalise on the pressing need for more robots in China’s economy.

Liang argues that deploying more robots in industrial settings will help to gather data for better physical AI models. Factories could also represent the best chance for China’s many humanoid robotmakers to monetise their products at scale, he says.

“Chinese wages, worker wages, are not cheap anymore, and new generations are reluctant to do the boring and dangerous jobs . . . the manufacturing sectors have great anticipation of either humanoid or non-humanoid [robots] to do more complex jobs out there.”

In the southern Chinese manufacturing hub of Guangzhou, GSK CNC Equipment was one of the country’s earliest industrial conglomerates to branch out into the relatively high-end field of industrial robots, developing its first models in the mid-2000s.

Like many Chinese industrial robot producers, the company’s biggest advantage over its Japanese and European rivals is its lower costs, says Huang Yongchao, a sales engineer at the company for 15 years. But he adds that workers in some of the lower-cost industries it sells to can struggle to adapt.

“The biggest challenge . . . is that our equipment is used in those kinds of factories where the work is dirty and tiring and the workers generally have lower levels of education. For tasks involving editing and manipulating text [to instruct robots], they might not be able to get to grips with them at first,” he says, standing in front of a phalanx of huge bronze-coloured robotic arms at a trade expo this month.

But Huang is hopeful that AI might have the answer. “They would prefer something simpler, or even something like cutting-edge AI technology, where people interact directly with the computer, which then uses language to tell them how to use it,” he says. “They want [the robots] to be able to understand their speech.”

FT : Quarter of UK water groups underspend investment allowances as network crea

Quarter of UK water groups underspend investment allowances as network creaks
Utilities failed to use maximum allowed despite arguing that higher bills were needed to fund improvements

A quarter of the UK’s privatised water companies underspent their investment allowances between 2010 and 2025 despite industry arguments that a failure to improve infrastructure was caused by customer bills being too low.

Water regulator Ofwat sets out every five years how much the utilities can raise prices, with the proceeds used to run their businesses and improve an ageing network. The companies have long said a focus on keeping bills low has constrained spending on infrastructure.

But Ofwat data showed that four out of the 16 water and sewage companies in England and Wales — Anglian Water, Northumbrian Water, South Staffordshire Water and Wessex Water — spent less cash than they were entitled to between 2010 and 2025.

Five others failed to spend all their cash in at least one five-year period during that window.

The findings, obtained by the FT under the UK Freedom of Information Act, come amid public anger and political scrutiny over the under-investment in infrastructure, which contributed to water outages, sewage spills and drought warnings.

Some customers are set to be hit with price rises of more than 50 per cent by 2030 compared with 2025 levels, the biggest increases since privatisation more than 30 years ago.

The average combined water and sewerage bill in England and Wales has risen by 373 per cent since 1990-91 including inflation, according to the Consumer Council for Water, which represents customers.

David Hall, visiting professor at Greenwich University, said water companies had “obvious incentives to underspend as it boosts profit margins, helping them to receive larger returns and the regulatory system makes it easy for them to do so”.

In some cases the groups underspent their allowances despite boosting dividends. Anglian Water paid out £2.5bn in dividends to investors between 2015 and 2020. In the period its spending was £383mn below its regulatory allowance, according to the FOI data, which was provided in 2013 prices.

The company said the £2.5bn included “payments made within our group structure to repay the debt raised to fund infrastructure investment” that did not leave the group. “Nonetheless dividends remain the best accounting term to use.”

Between 2015 and 2020 South West Water underspent its investment allowance by £255mn in 2013 prices. It paid out £781mn in dividends in that period. Its owner, Pennon, said: “The underspend in the regulatory framework reflects both capital and operational efficiencies, not just unspent investment.”


Since 2020 the water companies have increased their expenditure significantly as complaints over sewage pollution intensified. All but one used their entire allowance for 2020-25, with four companies overspending by more than £1bn each. The exception, South Staffs Water, underspent by £12mn.

Despite this, a National Audit Office report last year showed that critical performance metrics had not significantly improved and that some of the overspending was due to above-inflation cost increases, including for labour and energy.

Water mains were being replaced at an annual rate of 0.14 per cent between 2020 and 2024, which, if maintained, would mean the entire network would only be replaced once every 700 years, the NAO added.

Several water companies are financially stressed after the 16 utilities raised a total of £82bn in debt and paid out £85bn in dividends between 1991 and March 2025, according to research by the FT.

Industry body Water UK said water companies had overspent their allowances by 5.8 per cent since 2010. “No new reservoir has been built in more than 30 years because Ofwat and previous governments blocked them,” it said. 

Ofwat said: “We have approved a record £104bn investment programme which will strengthen infrastructure, improve resilience and support cleaner rivers and seas.”

FT : What would multilateral ‘AI arms control’ look like? Given the competition,

What would multilateral ‘AI arms control’ look like?
Given the competition, it’s debatable whether a US-China safety deal is even possible

With the strictly limited release of OpenAI’s GPT 5.6 model last week, it isn’t only Anthropic that’s bearing the brunt of the Trump administration’s newfound zeal for regulating AI. Sam Altman, OpenAI chief executive, announced his company’s latest model while noting delicately that “this isn’t quite the process that we think is optimal”. Yet America’s AI regulation debate isn’t limited to the White House and leading AI labs. It is being shaped by competition with Beijing. Could China and the US agree to rules that would address the risks in advanced AI systems while in the middle of the race? 

The frontier model contest changes nearly daily but Chinese companies continue to progress. The latest example is Z.ai’s newly released GLM 5.2 model, whose capabilities led David Sacks, co-chair of President Donald Trump’s Council of Advisors on Science and Technology, to declare: “We now have a Chinese open-weight model that is as good as the currently available models from OpenAI and Anthropic.”

In fact, model quality is growing harder to compare as the industry questions the efficacy of standard benchmarks such as mathematical problem completion. These benchmarks can be gamed if the model is trained on the specific problems in advance. What’s more, as OpenAI researcher Noam Brown recently noted, the right definition of capability is not only what problems a model can solve but how quickly and at what cost.

For complex, long-horizon tasks — the ones Anthropic and OpenAI are focused on — US companies may well have a longer lead than simple benchmarks suggest. But there’s no doubt Chinese models are improving. Before Trump, the Biden administration had hoped its AI chip restrictions on Beijing might give the US such a commanding lead in AI that Washington could dictate terms. This compute advantage has kept US companies ahead when measured by model capability or revenue, but it has not prevented Chinese competitors from releasing their own high-quality models. If Trump tightly regulates US companies he will kneecap American AI leaders without fully addressing global security concerns. 

There is one alternative to unilateral regulation: a deal between China and the US to address AI safety. The Trump administration appears keen for talks with Beijing. The two sides reportedly discussed the topic before the May Xi-Trump summit, though Beijing has not agreed to formal negotiations. So it is worth asking what a multilateral “AI arms control” regime could look like.

To start, we need to take cyber security — the reason that the White House is restricting foreign access to Anthropic’s Fable model — completely off the table. It’s nearly impossible to imagine the two countries agreeing not to use AI for cyber operations because it’s already happening so widely. Both Chinese hackers and the US National Security Agency use leading Anthropic models for this purpose, according to reports. This isn’t surprising: any model good at computer programming will be good at identifying cyber vulnerabilities.

A second regulatory priority is limiting the ability of AI to enable production of dangerous pathogens. It’s not impossible to imagine that China and the US could develop shared principles for assessing a model’s biosecurity risk. However, China’s record of biosecurity transparency is not reassuring. Nor are the cold war arms control parallels. When the US and Soviet Union signed the Biological Weapons Convention in 1972, the US had already begun dismantling its bioweapons programme. But the Soviets, convinced the US was lying, expanded work on anthrax and smallpox.

Could AI controls extend into the sphere of weaponry?

China’s pre-existing support for talks on banning autonomous weapons may sound reasonable. But will any major power renounce use of autonomous missiles or automatic missile-defence systems?

An even bigger issue is assessing compliance.

If a country promised not to use AI for a specific purpose, how could this be proved? It was hard enough to count missile silos during the cold war, even though these were visible from space. Neither China nor the US will give the other access to sensitive source code.

“Trust but verify” was Ronald Reagan’s wise approach to US-Soviet talks in the 1980s. So long as we distrust and cannot verify, we should not expect a significant deal. Both Washington and Beijing are likely to conclude that the priority is to keep racing instead.

FT : Elliott’s acolytes: how Paul Singer’s hedge fund became a spinout factory T

Elliott’s acolytes: how Paul Singer’s hedge fund became a spinout factory
The vaunted investor has spawned a small army of new firms mirroring Julian Robertson’s ‘Tiger cubs’

Elliott Management alumni are extending Paul Singer’s influence on Wall Street, as former staff use lessons learnt at the $80bn activist investor to launch firms of their own.

Members of the “Elliott diaspora”, as some former staffers call it, have founded at least seven hedge funds since 2020, mirroring the “Tiger cubs” that came out of Julian Robertson’s Tiger Management around the turn of the millennium.

They include Adam Katz’s Irenic Capital, Dan Gropper’s Carronade Capital, Quentin Koffey’s Politan Capital and James Smith’s Palliser Capital. Unlike Robertson, Elliott has not invested in any of the new firms, according to two people familiar with the matter. But even so, they have begun to find success in the same field.

“Elliott is a sophisticated investor, and people trained there tend to understand how to employ activism as a precise tool, not simply as a blunt instrument,” said one senior banker. “There’s a reason most defence advisers would rather have Elliott or one of its [alumni] across the table than a less disciplined activist.”


That Singer has started to amass his own lineage is a sign of his 49-year-old firm’s transition from a pugnacious hedge fund to a Wall Street institution with more than 600 employees.

It now manages $80bn compared with about $30bn a decade ago, and while it has struggled to keep up its momentum on returns in recent years, a new Elliott position has become so influential that its activism is often a self-fulfilling prophecy. When the hedge fund unveiled its stake in Honeywell two years ago, the company’s stock rose nearly 5 per cent.

But its targets also need to be increasingly large to move the needle, according to people familiar with the firm, creating opportunities for new funds that have the flexibility and interest in going after smaller companies.

Elliott’s scale means it has yielded many more spin-offs than rivals such as Starboard Value or Trian Partners, according to data from research firm Def 14. But it also has a distinctive approach, which firms such as Irenic and Carronade have tried to emulate.


Having waded into equity activism in the mid-2000s under the direction of Jesse Cohn, a wunderkind who has gone on to become a managing partner, Elliott now does everything from boardroom fights to takeover battles and distressed-debt brawls, all while maintaining an unusually intense eye on levels of risk.

The phrase Singer has used over the years to describe this phenomenon is “manual effort”, or the attempt to eke out better returns by sheer force of will and resources. Elliott will often have as many as 50 employees devoted to one investment.

Another reason the hedge fund has proved to be a finishing school for Wall Street’s next generation of activist investors is Singer himself, who has run Elliott since its founding in 1977.

“Paul was a great manager of people and he gave you rope,” said a former Elliott employee. “You either made it work or blew yourself up. But he was really good at enabling people.”

Singer allows portfolio managers to take big swings, a luxury enabled by Elliott’s uncompromising approach to risk.

Elliott tries to mitigate potential losses at every level of its investment through an elaborate web of hedges, according to two people familiar with the matter. Although this can be a costly strategy, it has given Elliott far greater precision in preventing losses.

“I like to joke that you can fit everyone who knows how to do that in my office, and my office isn’t that big,” said the former Elliott employee. “If you didn’t work at Elliott, you don’t know how to do that.”


While each of the newer hedge funds has its own style, one person close to Elliott described the firms as all running on the same “operating system”.

Carronade made 6.8 per cent in the first five months of this year, with a 10.5 per cent annualised return since its launch, according to people familiar with the figures. Irenic is up about 1 per cent, after making 16 per cent in 2025, according to people familiar with the numbers.

Elliott is known for going after intensely complicated companies and situations that other investors might shirk away from. Firms such as Trian Partners or Starboard are primarily associated with high-profile activist campaigns designed to lift a company’s share price.

Yet Elliott’s hallmark investments also often combine legal expertise, a savvy for credit documents and sometimes entire takeovers, as eventually took place with the $16.5bn deal for software company Citrix in 2022.

Another former employee said working at Elliott was particularly good for learning “how to prosecute a wide variety of weird and hairy and messy situations”.

Elliott offshoots have launched 50 activist campaigns since early 2023, according to Def 14, with targets as ambitious as Rio Tinto for Palliser and News Corp for Irenic.

Some of these campaigns have clinched early wins. Katz’s Irenic engineered a sale of Wagamama owner The Restaurant Group to Apollo Global Management; Smith pushed FTSE 100 systems testing company Intertek towards a sale to EQT; and Koffey took part in an almost $10bn sale of patient monitoring group Masimo to Danaher.

But starting a hedge fund in Elliott’s shadow brings its own challenges. The reputation of the brand can lead to investors feeling pigeonholed. “Once you work at Elliott, you’re labelled an ‘enemy to corporate management’ teams for life and thus more or less stuck as an activist,” said one executive.

Some of the Elliott offshoots have also struggled to emulate the hedge fund’s success.

Sparta Capital, which was founded by former Elliott trader Franck Tuil, made a series of ill-timed bets shortly after it launched, including on Spanish company Grifols. In 2024, multi-manager hedge fund Balyasny pulled its $250mn investment in the fund.

One person close to the fund highlighted that the positions were a small proportion of the overall bets the fund had made. “They have the same training but they don’t necessarily offer the same value proposition,” said one person familiar with Elliott.

One investor in Elliott, who has been pitched by a handful of the new hedge funds, said these founders had learnt the hard way that it was difficult to replicate what Singer had pulled off since the 1970s.

“You cannot recreate what you did at Elliott . . . because the costs are so enormous,” they said.

>>> US After Hours Summary: AVAV +18.2% sharply higher on earnings; ABVX +23.4%

After Hours Summary: AVAV +18.2% sharply higher on earnings; ABVX +23.4% on trial results; CNXC -22.2% under pressure after earnings
After Hours Gainers:
Companies trading higher in after hours in reaction to earnings/guidance: AVAV +18.2%
Companies trading higher in after hours in reaction to news: ABVX +23.4% (reports ABTECT maintenance Part 2 results for Obefazimod), CCOI +7% (closes on sale of 10 data center facilities), GMAB +2.2% (topline results from the Phase 3 EPCORE DLBCL-4 trial evaluating the combination of epcoritamab and lenalidomide), DHT +1.8% (will implement a proactive design upgrade for its two newbuildings delivering from Hyundai this year), REGN +1.3% (FDA selects seven particpants for PreCheck Pilot Program), MX +1.1% (names new CEO), CMCSA +0.7% (NBCUniversal weighing potential video-game entry post split; cable and connectivity business eyes technology investments, according to Reuters), AMRX +0.6% (FDA selects seven particpants for PreCheck Pilot Program), MEOH +0.4% (provides update on Trinidad and Tobago operations), FSM +0.4% (results from the feasibility study for its Diamba Sud Gold Project in Senegal), AZO +0.2% (Trump Administration issues memos relating to automotive repairs)
After Hours Losers:
Companies trading lower in after hours in reaction to earnings/guidance: CNXC -22.2%
Companies trading lower in after hours in reaction to news: NVCT -18.1% (stock offering), MAMA -12.8% (stock offering), VSH -10.4% (stock offering), FPS -4.9% (stock offering; also stock offering by holders), DLR -2.4% (to purchase BX interest in three data centers; also secondary offering of common stock by BX), TLRY -0.4% (announces acquisition of HelloMD), ORLY -0.2% (Trump Administration issues memos relating to automotive repairs), LLY -0.1% (FDA selects seven particpants for PreCheck Pilot Program)

>>> Abivax Data


Efficacy — refractory population held up well
Induction non-responders (the hardest cohort) on continued 50 mg at Week 44: clinical remission 37.2%, clinical response 61.5%, endoscopic improvement 48.0%, HEMI 44.6%, endoscopic remission 34.5% — clean dose-response vs 25 mg across every endpoint.
Part 1 relapsers recaptured: placebo→50 mg gave 45.0% remission / 69.7% response; 25 mg→50 mg escalation gave 45.5% / 66.7%. This is the practical takeaway — dose escalation rescues relapsers, supporting a flexible label.
Safety — the actual point of Part 2
Part 2 existed to fatten the safety database (1,704 patient-years integrated), since the malignancy/NMSC signal was the post-Part 1 overhang. Results:
Malignancy ex-NMSC (All Active) EAIR/100 PY UC background
Integrated UC program 0.35 0.30–0.70
Ph3 Maint (P1+P2) 0.56 0.30–0.70
Part 2 only 0.48 0.30–0.70
NMSC all-active sits at 0.59 (integrated) / 1.26 (P1+P2) / 0.95 (Part 2), against a 0.70–1.40 background. Note the 25 mg Part 2 NMSC point estimate is 1.52 — above range — but that's two events on small PYs, and all four Part 2 NMSCs had established risk factors (age, thiopurines, prior skin cancer). Two non-NMSC malignancies, both 50 mg, both deemed unrelated.
Bottom line: No new safety pattern, everything ex-25mg-NMSC within background, NDA reaffirmed for Q4 2026. The wide CIs on small event counts remain the bear's handhold, but management got what they needed here — the readout de-risks rather than re-rates the safety question ahead of filing.
Webcast at 10:30pm CEST tonight; detailed exposure-adjusted breakdowns by dose/PY/patient characteristics come there. Next catalysts: H1 results Sept 21, NDA Q4, ENHANCE-CD Crohn's mid-2027.
Want me to fold this into a Fiche Biotech update or draft the La Lettre item?