Codes of Culture | Issue 102
Europe’s largest seed round to build AI without humans
Welcome back to Codes of Culture. I’m Ashumi Sanghvi.
Excited to be in LA next week for the Milken Global Conference, where I will be hosting investors, family offices, and consumer tech brands in our network. There are a few times each year when we truly understand the signals and shifts across capital and culture, and being part of Milken is definitely one of them. If you are in LA/ SF next week, please reach out.
Something that caught my eye this week was Doji’s try-on, which features the latest Chanel Cruise show looks. You can try on yours by downloading the app, and for inspiration, see here.
In today’s issue, Gucci x Google partners to launch AI glasses, On’s lightspray robot for spray-on shoes under 3 minutes, Victoria Beckham and DressX AI twins, Europe’s largest seed round to build AI without humans, and finally, Google’s Pentagon deal, which Anthropic had refused, is now live.
If you are new here, or want to catch up on the best of Codes of Culture, we’d recommend you start here and remember to subscribe for full access to our news, insights, podcast and global events.
📖In this issue:
Kering is building the first luxury AI glasses.
On’s LightSpray robot is rewriting how performance footwear is manufactured.
Victoria Beckham and DressX are putting an AI twin in every online fitting room.
Ineffable Intelligence has raised the largest seed round in European history.
Google has signed the Pentagon’s AI deal.
1. LUXURY AND HARDWARE
Kering is building the first luxury AI glasses.
What’s happening: At its capital markets day in Florence, Kering CEO Luca de Meo confirmed plans to launch AI-powered smart glasses under the Gucci brand in partnership with Google, targeting a 2027 release. The move sits within de Meo’s broader agenda to grow Kering’s eyewear and jewellery divisions and diversify away from Gucci’s core fashion exposure. Kering will compete directly with EssilorLuxottica and Meta, which hold an estimated 82% of the smart glasses market through Ray-Ban Meta.
TLDR:
Gucci x Google for 2027, confirmed by the CEO. This is a strategic declaration, not a rumour.
Samsung entered with Gentle Monster and Warby Parker. Kering arrives with the most culturally weighted brand on the list.
Google’s Android XR platform already spans XREAL, Gentle Monster, and Warby Parker. Gucci raises the cultural register of the entire ecosystem.
De Meo’s broader thesis: a single brand identity imposed uniformly across every market is losing relevance. Smart glasses allow market-by-market differentiation.
Kering Eyewear already operates at scale across the group. The infrastructure exists. The missing element was a technology partner credible enough to carry the Gucci name.
Why it matters: The question of which brand holds cultural legitimacy at eye level is being answered in real time. Samsung sequenced its smart glasses strategy fashion-first; Google has now done the same, choosing Gucci. The houses that sit this cycle out will find the aesthetic language of wearable AI has already been written without them.
2. MANUFACTURING AND PERFORMANCE
On’s LightSpray robot is rewriting how performance footwear is manufactured.
What’s happening: Swiss sportswear brand On brought its LightSpray robot to London ahead of the city’s marathon, demonstrating the technology live at a pop-up in Hanover Square. LightSpray condenses approximately 200 traditional upper-assembly steps into a single three-minute robotic spray. On recently opened a second manufacturing facility near Busan, South Korea, with 32 robots, increasing production capacity thirtyfold.
TLDR:
200 assembly steps into three minutes. This is a manufacturing architecture shift, not an iteration.
The Busan facility moves LightSpray to consumer scale. The technology has moved beyond elite-only drops.
Hellen Obiri wore the Cloudboom Strike to a podium finish at the London Marathon. The performance credentials are live.
On framed the London activation around wellness programming alongside the robot demonstration. This is a cultural moment being narrated as much as a product one.
LightSpray reduces material waste and lowers carbon footprint relative to conventional assembly. The sustainability case compounds the design story.
Why it matters: The manufacturing process is becoming part of the brand narrative, and On brought the robot onto the streets of London to make that argument directly to consumers. The Busan expansion is the more consequential signal: this is no longer a proof of concept. The brands that can demonstrate genuine material innovation at consumer scale and simultaneously build the cultural frame around it will define what premium performance means in the next cycle.
3. FASHION E-COMMERCE AND AI IDENTITY
Victoria Beckham and DressX are putting an AI twin in every online fitting room.
What’s happening: Victoria Beckham has launched an AI-powered virtual try-on tool on its e-commerce site in partnership with DressX. Shoppers can try selected garments on a full-body photograph or via a personalised AI twin generated from a selfie, with video try-on showing garments from multiple angles and in motion. The platform generates size recommendations from either a full-body image or a selfie, along with user measurements. DressX reports that try-on users convert to purchase at ten times the rate of those who do not, with up to seven times higher long-term retention.
TLDR:
An AI twin from a selfie, showing garments in motion and from multiple angles. A different proposition from image overlay.
Conversion ten times higher, retention up to seven times higher. The commercial case is unambiguous.
Live on Victoria Beckham’s own site, not a marketplace test. Applied to the brand’s direct customer relationship.
Sizing uncertainty drives the majority of returns in luxury e-commerce. A credible solution restructures channel unit economics.
DressX frames this as a precision problem: fashion demands visual accuracy that standard AI tools miss.
Why it matters: The conversion and retention numbers are not a feature story; they are a margin argument. Every fit interaction builds a proprietary data asset: how garments behave on real bodies, at scale, across a direct channel. The brands that integrate this now, rather than waiting for platform adoption, will hold a compounding advantage. The AI twin is not a novelty layer. It is becoming the infrastructure of the fitting room.
4. CAPITAL AND AI ARCHITECTURE
Ineffable Intelligence has raised the largest seed round in European history.
What’s happening: David Silver, who led the reinforcement learning team at Google DeepMind and was the principal researcher behind AlphaGo and AlphaZero, emerged from stealth with Ineffable Intelligence, a London-based lab founded in November 2025. The company raised $1.1 billion at a $5.1 billion valuation, the largest seed round in European history. Sequoia and Lightspeed co-led, with Nvidia, Google, Index Ventures, DST Global, and the UK Sovereign AI Fund participating. NVIDIA’s venture arm contributed at least $250 million. No product, no roadmap. The mission: a superlearner that discovers knowledge through experience rather than human-generated data.
TLDR:
$1.1 billion, $5.1 billion valuation, no product. The capital is pricing a paradigm thesis.
Silver’s argument: LLMs are trained on human data, a finite resource. Reinforcement learning learns from experience and carries no such ceiling.
AlphaGo Zero proved the logic at the game level: trained on no human data, it beat the world’s best players. Silver is betting the same principle scales beyond games.
Sequoia, Lightspeed, Nvidia, Google, DST Global, and the UK Sovereign AI Fund in the round. Compute, distribution, sovereign backing, and long-duration capital simultaneously.
The UK government’s involvement in the Sovereign AI Fund marks a shift from policy statements to direct capital positioning in frontier research.
Why it matters: What this raise tells you is where sophisticated capital believes the LLM architecture cycle ends. The thesis is not that language models will become irrelevant; rather, the most significant value will be captured by systems that generate genuinely new knowledge, not by those that recombine existing human output. The industries most exposed are those where the relevant knowledge has never been written down: manufacturing, physical medicine, and materials science. That is a different disruption map from the one most operators are currently planning against.
5. POLICY AND THE GEOPOLITICS OF AI
Google has signed the Pentagon’s AI deal.
What’s happening: Google has reached a classified agreement with the US Department of Defence allowing its AI systems to be used for "any lawful government purpose," according to Bloomberg and The Information. The deal was finalised despite a letter from more than 700 Google employees, including senior DeepMind figures, urging CEO Sundar Pichai to refuse. It places Google alongside OpenAI and xAI in the Pentagon’s supplier base. Anthropic remains in an active lawsuit against the DoD after refusing equivalent terms and being designated a supply chain risk.
TLDR:
Google’s agreement covers "any lawful government purpose,” the same language Anthropic rejected.
More than 700 employees, including DeepMind directors and vice presidents, signed an open letter opposing the deal before it was signed.
In 2018, Google withdrew from Project Maven following employee protests. Proceeding now is a deliberate reversal.
Anthropic’s lawsuit remains active. A judge has granted an injunction against the supply chain risk designation while the case proceeds.
Pentagon AI chief Cameron Stanley confirmed the expanded relationship with Google and noted that overreliance on a single vendor is not strategically sound.
Why it matters: Two AI companies faced the same terms and made opposite decisions. The cost of each is now visible: a supply chain designation and a lawsuit on one side; significant internal dissent and a public reversal of stated principle on the other. Every AI company with government adjacency is watching. The frame for what responsible AI commitments mean in practice, when they conflict with a client’s requirements, is being set inside this dispute.