Twitxy Saved

Alex

Verified account This account is verified. Learn more

@obchakevich_

Research for private companies and VCs. Trusted by Visa, Polygon, Avalanche. Featured in a16z, ARK Invest, Flagship Advisory Partners and others.
Not financial advice! Joined October 2018
2.1K Following 67K Followers 30K Posts

Latest posts 5

Pinned post

Alex Verified account @obchakevich_
The payments table tells a story nobody planned.

@Visa is up 5% for the year, @Mastercard barely 1%, and @circle - down 42% over twelve months - just gained 30% in a single month. Read quickly, it looks like noise. Read slowly, it's the market changing its mind about who actually earns money on stablecoins.

The old thesis was simple: pay a premium for distribution, avoid the issuers. Visa ( $V ) and Mastercard ( $MA ) own the rails, the merchants, the habits of 8b cards.

Circle ( $CRCL ) owns a token and an interest rate. When rates fall, circle bleeds; the giants just keep clipping their 2-3%.

That thesis is quietly breaking from both ends. Circle stopped being a pure rates trade - a federal trust bank charter, a swing to profit in Q2, transaction revenue doubling. @USDC did $849b in july volume, 62% of the entire market.

The rally isn't a discovery, it's a correction of an overdone funeral. The stock still sits 58% below last year, insiders are selling, and wall street can't agree on a target within a factor of two. But the business underneath has changed shape.

Meanwhile the "safe" side got less safe. Mastercard paid $1.8b for @BVNKFinance - the same firm that powered Visa's stablecoin payouts, which is the kind of irony m&a lawyers dream about.

Both networks joined the open standard consortium behind $OUSD. They are no longer toll collectors watching the disruption from a bridge; they are issuers now, competing in the market they were supposedly immune to.

And the reason is visible on-chain: $33t in stablecoin transfers last year, growing 72%, with AI agents learning to route around anything that charges interchange.

So the divergence in the table isn't old rails versus new rails. It's a market splitting one bet into three - distribution, issuance, and the consortium coin that could squeeze both.

The premium for standing still just stopped being free.

Data: @artemis
17 replies 17 reposts 129 likes 243K views View on X
Alex Verified account @obchakevich_
In its October 2026 report, the @AtlantaFed says US companies spend about $1,024 per employee per year on AI. But if we look at their May post, the very same survey shows $2,068.

I think we can be fairly confident that both numbers are right. The only difference is how many of the biggest spenders you count. Cut the top 5%, and the average falls by half. A small number of companies account for almost all the spending, while the rest simply buy a few subscriptions.

$1,024 a year is roughly one ChatGPT license + one Copilot, and that’s it. No agents, no API, no premium tools. If we look at @tryramp card data, we see the same picture: the median company spends $190 a year, the top 1% spend $115k.

So while headlines like “AI spending has doubled” are just words, in reality for most companies it doubled to the price of two basic subscriptions.

The ones actually paying for the data centres sit 600x to the right of the median firm.
23 replies 6 reposts 40 likes 2.3K views View on X
Alex Verified account @obchakevich_
Idk why this hasn't become popular yet. It's the best thing I've read in the last six months. Everything is on point and gets the message across.

Fundraising needs to be engineered, not improvised.

If you're preparing to raise funds for your project, this material is a must-read 👇🏻
Jamie Gull Verified account @jamiegull 4d
How to engineer your fundraise, Part 1: building the pipeline before you ever pitch. x.com/i/article/2118…
View quoted post on X
18 replies 3 reposts 35 likes 2.9K views View on X