Supabase at $10B, Vercel at $9B.
The “build layer” is becoming one of the most valuable parts of the AI stack.
As AI makes software easier to create, the platforms that handle database, deployment, APIs, app infrastructure, etc..become even more important.
Supabase has raised $500M at a $10B valuation
In this round we are giving @supabase employees the opportunity to cash out 25% of their vested options. We have done this in every round since inception.
We do it as a “cashless transaction” so that employees don’t need to front any cash to exercise their options. This is the friendliest way we could design it until we can offer RSUs.
On top of that, we give employees a 10 year exercise window: whether they stay or leave the company. The typical/default window is 3 months. IMO, equity is earned and employees shouldn't be penalized because they don't have the cash to exercise within 3 months of leaving a job (often that's the time they need the cash/certainty the most).
Suno is the prime example that AI is not just productivity software.
It is entertainment + creation
Great consumer AI companies will not only save time. They will create new formats, new habits, and new culture.
AI music startup Suno raised $400M led by Bond at a $5.4B valuation, up from $2.45B after raising $250M in November 2025, and passed 2M subscribers in February (@lucas_shaw / Bloomberg)
(Visit Techmeme dot com for the link and full context!)
Come watch Game 1 of the NBA Finals at The Jeffrey on Wednesday. If the Knicks win, we’ll pick up your tab!!
Offer valid during the game only for guests who arrive before the tip. Up to $100 per guest. Tax and gratuity not included. Drink responsibly. No purchase necessary.
I’ll be in NYC for ETHConf and around for a few days after.
Availability may be tighter when the Knicks are playing, but if you’re in town, DM me, would be great to meet up!
Heading to @ethconf NYC.
Ethereum continues to be a proving ground for the next generation of applications, digital ownership, on-chain coordination, and AI-native systems.
@justpat and @paulhsu will be there representing Decasonic and looking forward to meeting the founders and builders turning these ideas into products.
See you in NYC!
Politics gets discussed in binarys, but power is not a binary variable.
Just like politics it is a spectrum shaped by probabilities,
What makes kpow interesting is that it tries to turn that high-dimensional political state space into a single market-priced signal.
Not perfect, but directionally important.
Instead of arguing only about who is winning, we can start measuring the magnitude and direction of how power is shifting. imo a great metric to start using in debates as well.
Someone asked me a few days ago what the real use cases for prediction markets are.
Something like this was basically my response.
Discoverability here has been simplified.
identifying the risks and translating them into something that can actually be hedged.
Happy to ship the first version of Hedgebook, powered by @Kalshi : an app showing how every SPX companies can hedge their core business risks using Kalshi event contract markets.
Event contracts and perpetual contracts consolidated into a single, onshore, regulated exchange is how we get to perfect Arrow-Debreu markets.
Thanks to @j0hnwang and @0x_ultra for inviting me to build this, and to Nicole from @KalshiResearch for inviting me to present the early frameworks at the Kalshi Research Conference in March.
Late night thought as I build product around this. (View's not bad)
Everyone's anchored on what a company earns in year ten. We can't know that, never could. The instinct is to chase a cleaner number.....wrong instinct.
The answer keeps coming back to the opposite: aggregate dozens of methods, map the whole landscape, and let the truth live in the range instead of the point. Then underwrite to what's actually in front of you and structure for the upside you can't price. The single number on a ten-year horizon was always narrative. The discipline is holding the range without flinching.
Culture is coming back to crypto...in a different and better way.
The last cycles proved people want to trade culture. NFTs, memecoins, etc.. (these still have traction not on scale) had the attention layer, but often lacked real backing + utility
The next wave is cultural assets with actual underlyings: collectibles, creators, athletes, moments, and markets that people already care about.
Not culture as a meme but as an asset class.
the AI market is obviously scaling fast, but i believe we are seeing the bottleneck change with the scale.
model -> coordination
Security, maintenance, verification, and incentives for builders are all becoming increasingly relevant as AI gets embedded deeper into workflows and outcomes.
frontier labs will no question capture a lot of value at the model and app layer.
But decentralized AI becomes more compelling as the bottleneck moves beyond just model quality. Once AI is embedded across software, security,and workflows, the bigger question is how participation, economics, and trust rails help builders coordinate around the systems being deployed.
This feels like a large asymmetric opportunity.
I don’t think the key AI x prediction market insight is just AI markets getting more volume. it is infrastructure. Traditional finance was built for traditional human institutions Prediction markets are increasingly built around APIs, wallets, stablecoins,settlement. A much more natural environment for agents.
Prediction markets can become the first place where an agent can actually express belief, price uncertainty, and coordinate.
AI progression feels natural in micro dramas.
This is less about replacing production teams and more about changing the cost curve of experimentation. There has been a high barrier to entry for creating quality, typically full production infra.
Micro dramas already depend on fast payback: test the hook, retention, conversion, then scale. The issue is that traditional production still has real fixed costs even when the content is only max a few minutes long.
AI fits because it makes more creative bets economically viable: more experimentation at a much lower marginal cost.
China’s short drama market is such an important signal. AI can change the actual workflow AND ECONOMICS.
China is KILLING it right now in AI short dramas.
They did $14B in revenue last year and are on track for $16.5B this year, which is bigger than their entire box office.
One of the AI shorts drama, "Huo Qubing" did 500M views on a $450 budget.
They launch a new AI drama every 90 seconds, 470 a day, 14,600 in January alone.
AI is now 38% of China's top 100. last year it was 7%.
it's one of the biggest real-world examples of AI changing entertainment
2026 TGEs are clearing when the market can already underwrite real traction, disciplined supply, quality liquidity, and credible value accrual.
Wrote more on what is actually working in token launches right now.