Hi there,
Welcome to the 132nd edition of Heartcore Insights, curated with 🖤 by the Heartcore Team.
In this edition, instead of a thematic Prepared Minds essay, we are sharing our perspective on the ongoing AI-bubble conversation.
If you missed the past newsletters, you can catch up here. Now, let’s dive in!
Over the past year, the idea of an “AI bubble” has dominated headlines. Valuations have grown faster than fundamentals, and each week brings new commentary questioning whether expectations can possibly be justified. NVIDIA’s latest earnings call showed the same tension: record results and rising demand, yet the stock fell. Fundamentals and sentiment are moving on different clocks.
That same anxiety had been amplified earlier by the MIT article claiming that AI had yet to produce “measurable economic value”, a line that briefly made every early-stage investor look misguided for pouring billions into the space. The moment passed quickly, but it revealed something important: the debate around AI is happening at the extremes. And when conversations are set at the edges, the narrative becomes binary and short-term, ready to flip overnight the moment markets decide things have gone too far, too fast.
The dot-com boom and bust is the masterclass example of this dynamic. The internet arrived, and with it a belief that everything would be remade instantly. When the crash came, it was easy to claim that expectations had been unrealistic. But history tells a different story: the underlying thesis for the internet was correct. It was the timeline that was wrong.
The internet didn’t disappear after 2000; it simply grew into its valuation over time. Companies that failed in 1999 later became mainstream a decade or two after:
Grocery delivery (Webvan → Instacart)
Digital music (MP3.com → Spotify)
Social networks (Friendster → Facebook)
Specialist e-commerce (Pets.com → Chewy)
Ideas that “didn’t work” in the first cycle failed not because they were bad ideas, but because the world around them wasn’t ready.
The dot-com era taught us that the hype, and ultimately the bubble, played a decisive role in building the infrastructure that became the backbone of modern digital life: fibre networks, broadband and the early generation of data centres. In that sense, it is worth considering whether it takes a bubble to fund what only a bubble would dare to build.
The cycle of optimism → excess → collapse → reconstruction may simply be how societies finance the future.
If that pattern holds, today’s AI spending on models, chips, compute clusters, data centres, agent frameworks and new workflows might look reckless in the short term but inevitable in hindsight. Most companies won’t survive, and that is not a defect. It is the mechanism through which foundational technologies are tested, sorted and strengthened.
Which leads to the real question: Not: “Are we in an AI bubble?” But: “What will remain after it?”
For founders, the key question is not whether a downturn will come, but whether their company deserves to exist after it does. Can the product stand without narrative tailwinds? Will customers care if the hype disappears?
For early-stage investors, the calculation is equally simple: the opportunity cost of not participating in a generational platform shift is far greater than the risk of participating. Early-stage investing is, at its core, the business of backing the future which is why the “AI bubble” question has always been somewhat beside the point.
In the end, the cycle will do what cycles do best: it will sort out the true winners and clear out the rest. That selective pressure isn’t a flaw of the system. It is the system. And everything else is the cost of progress.
~ Lærke Hansen, Principal, Heartcore Capital
Europeans, You Need Silicon Valley, But Silicon Valley Doesn’t Need You, Judith Dada
AI Will Not Make You Rich, Colossus
The Anti Mimetic Index, The Capitalist
dotega, a Germany-based prop-tech for homeowners’ association, raises €1.3M with HTGF - link
Motley, a Switzerland-based AI-powered business reporting platform, raises $1.5M with Seedcamp - link
Spinnable AI 🖤, a Portugal-based provider of scalable AI agents to SMEs, raises $2M with Heartcore - link
Tsuga, a France-based observability platform, raises $10M with General Catalyst - link
Faktus, a France-based neo-bank for construction SMEs, raises €9M with Lakestar - link
GetVocal, a France-based developer of hybrid human-AI agents, raises €22M with Creandum - link
FOMO, a social crypto trading app, raises $17M with Benchmark - link
Agency, a customer success startup, raises $20M with Menlo/Sequoia - link
Sphere, an AI-native tax compliance software, raises $21M with A16Z - link
Kaizen, a digital services platform for governments, raises $21M with NEA/Accel/A16Z - link
Method Security, a startup building autonomous cyber systems for security teams, raises $26M with A16Z/General Catalyst - link
ElectronX, an energy exchange created to address volatile short-term price exposure to electricity, raises $30M with DCVC - link
Harvey, a US-based domain-specific AI for legal and professional services, raises $150M with A16Z - link
Legora, a Sweden-based collaborative AI platform for lawyers, raises $150M with Bessemer/Iconiq/General Catalyst - link
Mercor, a US-based data-labelling firm raises $350M with Felicis/Benchmark/General Catalyst - link
Ramp, a US-based corporate expense management platform, raises $300M with Lightspeed - link
Kalshi, a US-based predictions market platform, raises $1B with Sequoia - link
Anysphere, the US-based maker of coding assistant Cursor, raises $2.3B with Accel/Coatue - link
Spinnable AI 🖤 raises $2M in a pre-seed round led by us! 🚀
TravelPerk becomes Perk 💚
Corti 🖤 launches clinical-grade Danish speech recognition model 🩺
Skynopy 🖤 awarded with the ESA Rising Star Award 2025! 🎉
AAVE 🖤 launches a consumer facing app! 📲
We welcomed Michael Baczyk to the investment team to back the next generation of quantum founders! 🥳
Thank you for being a loyal subscriber of our monthly Insights. Please feel free to share this newsletter with anyone you’d think would appreciate it!
Comments
Nothing yet. Say the first thing.
Sign in to join the conversation.