Raghav Agarwal-1

IPOs and acquisitions are shaking up Web3, but let’s not overlook another crucial path: collaborating with TradFi institutions for scaling adoption. @DeutscheBank’s recent DAMA 2 whitepaper highlights how Web3-native teams like @zksync, @axelar and @Memento_Bc are playing a vital role in enabling tokenization on a global, compliant scale.

Key takeaway for Web3 teams: don’t just stick to your crowd—grow the pie! Partnering with established institutions isn’t just clout; it’s your path to scale, revenue, and real-world credibility. Pumped to see this unfold and hopefully many more similar announcements will come soon.

The @solana expansion of @1inch in April is starting to show results already. From 32% marketshare in March to 60% in May.

May 2025 DEX aggregator volume hit an ATH of $55B, and @1inch processed $33B.
The Solana aggregation wars are just getting started.

The recent Ethereum Protocol Research Call 2 was packed with many exciting updates on how @ethereum is innovating in the short to medium term. Here is what you should not miss 👇 🧵

The one chart in the @ethereum Protocol Research Call 2 that excited me the most, was this one 👇

A powerful L1 with rollups clustered around it is preferred over a weak L1 with dominant rollups, as the latter risks rollups becoming independent or moving elsewhere. Even in a rollup centric future, we need the L1 to scale. It is imperative that Ethereum continues to innovate, improve and inspire!

The story of perp DEXs is best told in 7 distinct phases — a timeline of dominance, disruption, and the rise of a new king:

1. Early–Mid 2021: @dYdX Dominance
Perp DEX = dYdX. No real challengers. Deep liquidity, strong PMF, and the only game in town.

2. Late 2021: @GMX_IO Appears
GMX v1 entered with an oracle-based model and a community-first ethos. Still tiny vs dYdX, but it proved new models could work.

3. 2022: Market Starts Fragmenting
@GainsNetwork_io, @synthetix_io, and others began attracting niche users. dYdX’s monopoly was intact, but the moat showed cracks.

4. Early 2023: @HyperliquidX Arrives
A quiet entrant with serious momentum. Product velocity + incentives helped it quickly carve out share from both dYdX and GMX.

5. June 2023: A Brief Pullback
Hyperliquid’s share dipped to less than 2%. A blip, not a bust. The retreat was temporary.

6. June 2023–Dec 2024: The Climb
Hyperliquid steadily rebuilt, volume growing month after month. By late 2024, it had leapfrogged everyone. @JupiterExchange also managed to gain some marketshare against the others.

7. Jan 2025 to current: Hyperliquid Dominates
Hyperliquid now commands the perp DEX market with more than 80% marketshare in May 2025. dYdX v4, ApeX, and others still trade—but the race isn’t close anymore.

Wir erleben derzeit den Aufbau des einflussreichsten und vertikal integrierten KI-Imperiums, das von @sama aufgebaut wird.

Es ist kühn. Es ist erschreckend. Und während es anfängt, Gespräche zu entfachen, werden die Auswirkungen immer noch massiv unterschätzt.

Hier ist, was er Stück für Stück zusammenstellt:

1) Kern-KI-Modelle
@OpenAI unter Altman hat Grenzmodelle wie GPT-4.5 und GPT-4o veröffentlicht - mit GPT-5 in Arbeit. ChatGPT ist bereits die am häufigsten heruntergeladene App auf der Erde, während diese Modelle leise die KI-Engines der größten Unternehmen der Welt antreiben.

2) KI-Hardware
6,5 Milliarden Dollar zur Übernahme von Jony Ives Design-Startup, io. Ive - der Kopf hinter dem iPhone - leitet jetzt die Hardware-Vision von OpenAI. Erwarten Sie KI-native Geräte, die definieren, wie Menschen mit Maschinen interagieren.

3) Recheninfrastruktur
Ein 1-Gigawatt-KI-Super-Datencenter - Stargate VAE - wird mit G42 gebaut. Das ist Rechenleistung im Maßstab einer Nation in einem einzigen Projekt. In der Zwischenzeit richtet OpenAI ähnliche Ventures in den USA ein.

4) Identität & menschliche Verifikation
@worldcoin, Altman’s kryptobiometrisches Projekt, scannt Iriden, um die Menschlichkeit zu verifizieren - bereits bei 26 Millionen Nutzern. In einer KI-saturierten Zukunft könnte es zur Standardebene werden, um nachzuweisen, dass man online menschlich ist.

5) Soziale & Verbraucherplattformen
OpenAI baut Berichten zufolge seine eigene X/Twitter-Alternative und hat Interesse an der Übernahme von Chrome gezeigt, falls Regulierungsbehörden es jemals von Google trennen. Das ist ein Browser, ein sozialer Feed und eine KI-Ebene, alles potenziell unter einem Dach.

6) Entwicklerwerkzeuge & Infrastruktur
OpenAI hat Windsurf (KI-Coding) für 3 Milliarden Dollar und Rockset (Echtzeitanalytik) übernommen. Diese bilden das Software-Rückgrat für den nächsten KI-Stack.

Das ist nicht nur ein Unternehmen - es ist ein Stack. Eine Full-Stack-Wette, KI vom Silizium bis zur Gesellschaft zu besitzen.

Es ist wie das Zuschauen, wie eine Black Mirror-Episode live geschrieben wird... und wir sind im Cast!

In ernsthaften Angelegenheiten, kann mir jemand helfen zu verstehen, wie Move sicherer sein sollte und warum das im Fall des Sui-Hacks nicht funktioniert hat?

While @ethereum and @solana were battling it out for dominance, @HyperliquidX emerged as the market leader in terms of daily fees.
HL went from ~17% in mid-Jan to ~64% in May

Data source: @DefiLlama
Area chart of dailyFees of Hyperliquid, Solana, Ethereum.

I can finally show my face at family gatherings—everyone who bought $BTC on my advice is in the green.

One of the more subtle but troubling takeaways from Token2049 Dubai: too many teams are spending heavily at conferences with little regard for accountability or actual value creation. And this isn’t just limited to side events—what’s more concerning is the kind of personal expenses being claimed, often without much scrutiny or justification.

Yes, dinners, networking, and side events are an essential part of these gatherings. But what stood out to me was the almost complete absence of ROI thinking around these expenses. In some cases, it felt like spending was done for the sake of spending.

Now, defining ROI for conferences is tricky—it’s not always about direct returns. But that shouldn’t be an excuse for unchecked or outlandish costs. At the very least, teams should be asking: is this aligned with our broader goals?

Leadership has a key role here—not through micromanagement, but by establishing high level frameworks that guide how budgets are used and what success looks like. Without that, it's easy for good intentions to turn into expensive inefficiencies.

Personally, every time I attend a conference, I stay mindful of the cost my company is bearing. I always ask: what value am I creating in return? That mindset shouldn't be rare. It should be the baseline.

This is especially necessary for teams which are not generating revenue at the moment and require runway to find PMF and revenue in the long run.

At Token2049, every friend from RaaS and DA teams echoed the same thing: demand for launching new rollups is cooling. The OP Stack hasn’t had a major announcement since Unichain. EF is focused on scaling L1. Yet Caldera keeps popping up all over my timeline.