Continental Breakfast

Continental Breakfast

What a time to be live. The main thing happening today is Meta saying its new Muse Wonder model has caught up with the frontier. Here's your Order of Play, plus Luke on how Bending Spoons buys its growth.

12:45 PM - Alex Emeshev, Co-Founder at Vivid

1:00 PM - Will Wang, Founder & CEO at Even Realities

1:15 PM - Kanishka Narayan, Minister for Artificial Intelligence at Cabinet Office

1:30 PM - Bela Wiertz, Co-Founder & General Partner at Calibration

What’s Happening Today

Meta says its new Muse Wonder model is now at the AI frontier, with state-of-the-art performance on agent tasks.

The new model comes with better memory, stronger reasoning across complicated tasks and more expertise in areas like health, finance and law. Meta's own benchmarks have it ahead of OpenAI's Astra and Anthropic's Fable on several tests, and it'll soon be powering Muse, its personal AI agent.

Muse has attracted nearly nine million downloads since launching last month, but they are are slowing. Fewer users are sticking around compared with ChatGPT after two weeks according to the WSJ. Meta has spent the past year rebuilding its AI efforts after the disappointing launch of Llama 4, bringing in Alexandr Wang to lead its new Superintelligence Labs.

But better models only solve part of the problem. Amazon has blocked Muse from completing purchases on its website, and other sites have restricted its access to their services. For AI agents to reach their full potential, they need websites to let them in.

We’re getting into all of it in today’s show at 12 BST.

Daily Op-Ed — Bending Spoons

Jason Lemkin wrote a good tweet about what’s happened in the last 102 days since Bending Spoons went public after their Head of Product, Federico Simionato said they were laying off a large part of the Airtable team This morning Monty made a graph (below) showing how Bending Spoons has spent roughly $8.8B buying companies and is now worth about $26.1B EV. That’s about 3x what it paid. Their strategy is a masterclass in how buying growth can create more value than growing organically. Nearly all of the added revenue came from acquisitions. Organic revenue growth of the businesses it already owned was only 3% and as Jason points out almost all “the $393M of added revenue in Q2 came from six acquisitions”. Their strategy is wholly based on the multiple arbitrage of buying revenue at low multiples (2-3x) and trading at a higher public market value (around 8x). When they buy a company their CEO Luca Ferrari says the goal isn’t to optimise growth but optimise EPS. They cut costs quickly - fire a bunch of the workforce, centralise the operations and use AI extensively. So the same revenue is therefore worth significantly more once it is inside Bending Spoons than it was when inside a separate company. In the graph you can see the value creation. Estimated total spent on acquisitions = $8.8B (major deals like AOL, Vimeo, Miro, Airtable, and Tractive) compared to current EV of $26.1B. The gap is both the higher public-market multiple and the operational improvements (cost cuts, higher margins, price increases etc) made after the deals close. The market assigns a higher multiple for a few reasons. Investors are valuing the acquisition engine and the ability to keep buying companies cheaply, rather than simply the sum of the individual businesses. The model assumes Bending Spoons can continue finding deals at attractive prices and successfully improving margins after each acquisition. Debt is also used to fund many of the purchases, which is why enterprise value exceeds market cap.

— Luke

etn. Insight

Airwallex CEO Jack Zhang joined the boys in the studio and told the story of turning down Stripe's $1.2 billion offer when Airwallex was making just $2 million in revenue!

It came down to a five-year lockup. "I was 34," he said. "What am I going to do when I'm 39, when I finish the lockup? Do I have the energy to build another startup?" He wasn't even sure he'd survive the five years, and felt Airwallex was "more likely to be succeeding on our own than sell ourselves".

Watch the full clip below.

Continental Breakfast

AI Researchers Start Smoking Expecting AI To Cure Cancer

City of London Wants Non-Competes Back

Anthropic Hunts for Australia Data Center Deals