A-Portfolio

The company in one breath

Circle Internet Group, ticker CRCL on the New York Stock Exchange, issues USDC, a digital dollar backed by cash and short-term U.S. government assets, and sells the plumbing that lets businesses move it across blockchains. Circle earns most of its money from interest on the assets behind USDC, then pays partners such as Coinbase (COIN) to distribute the coin.

That makes Circle less mysterious than crypto people prefer. It is a regulated dollar reserve business wearing a software jacket. The bigger dream is Arc, Circle's own blockchain for payments, trading, and financial assets. If Arc works, Circle can own the dollar, the road, and some of the tollbooths.

Why now

Arc's public network launches September 16. A public mainnet is the live blockchain where real transactions and real money move, rather than a test environment. Good would look like applications, active users, and transaction fees arriving quickly after launch, not twelve months of partnership logos doing jazz hands.

Circle enters launch week with serious names. BlackRock (BLK), Intercontinental Exchange (ICE), Mastercard (MA), Visa (V), and DTCC (private) sit among Arc's 11 founding validators, the institutions that check transactions and secure the network. More than 100 builders already use the private network. That means Arc has credible launch partners. Good would look like those partners running valuable transactions, because a famous validator without activity is a nightclub bouncer guarding an empty room.

The stock closed Friday at $102.05, then traded at $101.60 after hours. The regular close values Circle near $25.9 billion, which means investors already pay for more than today's interest business. Good entry quality would give us a price that assumes Arc may work, not a price that assumes it already does.

The numbers, translated

• USDC circulation ended the quarter at $73.3 billion, up 19%. Circulation is the amount of digital dollars customers hold. More USDC gives Circle more reserve assets that can earn interest. Good for a network chasing category leadership would be sustained growth above 25%, so 19% is healthy but below management's long-term 40% ambition.
• Onchain transaction volume reached $14.8 trillion, up 151%. This counts the value moved through supported blockchain transactions. It means usage grew much faster than the coin supply. Good would look like that activity producing transaction and service revenue, because moving oceans of money for free wins applause, not owners.
• Quarterly revenue and reserve income was $701.3 million, up 6.6%. This is all money Circle brought in before paying distribution partners and operating bills. Growth below 10% is modest for a company valued like a category leader. Good would be growth above 20%, driven by products rather than higher interest rates.
• Reserve income was $667.7 million, or 95.2% of total revenue. Reserve income is interest earned on the safe assets backing USDC. The number means almost the entire business still depends on coin balances and short-term rates. Good would look like non-reserve products reaching at least 20% of revenue, so Arc and the payments network can carry weight before rates fall.
• Revenue after distribution and related costs was $288.8 million, a 41.2% margin. Circle calls this revenue less distribution costs. It is the money left after paying Coinbase and other partners that help put USDC in customer hands. Good would look like a margin above 45% without slowing circulation, so the latest 41.2% shows improvement but not dominance.
• Standard-accounting operating income was $34.4 million, a 4.9% margin. Operating margin is what remains after distribution and normal business costs, before taxes and certain other items. Circle kept about 5 cents of operating profit from each revenue dollar. Good for a scaled financial network would be at least 15 cents, so today's profit is real and still thin.
• Net income was $48.2 million, a 6.9% margin. Net income is what common owners get after the full accounting bill. Good would be a margin above 15% that repeats without token-sale help. Circle is profitable, but nobody should confuse one clean quarter with a cash cannon.
• First-half operating cash flow was $538.5 million, up from $303.7 million. Operating cash flow tracks cash generated by daily business activity. The increase looks strong, but $222 million came from an Arc token presale. Good would look like recurring customers producing the cash, not buyers prepaying for a token before the network opens.
• Stock compensation was $53.6 million, or 7.6% of quarterly revenue. This is employee pay delivered in shares, which spreads each owner's claim across more pieces. Good would be below 5% of revenue and annual share growth below 3%. Circle is close on the expense and not yet clean on dilution.
• At $102.05, Circle trades near 8.9 times trailing revenue and 57.8 times trailing earnings. The first number means investors pay almost $9 for every $1 of annual sales, while the second means they pay nearly $58 for every $1 of annual profit. Good for a company growing total revenue only 6.6% would be closer to 6 times sales, unless Arc quickly creates recurring, high-margin growth.

What I like

Circle owns a scarce position. USDC combines scale, liquidity, regulatory work, and connections into exchanges, banks, wallets, and payment systems. A rival can mint another stablecoin by lunch. Building trust with institutions takes years and a lot more than a logo generator.

Arc gives the company a credible route beyond reserve interest. The network uses USDC for transaction fees, promises final settlement in under a second, and adds privacy controls that banks can understand. Recent GitHub work from Trust Wallet (private) and ShapeShift (private) shows outside developers preparing Arc support before launch. That is a small signal, but it beats another executive saying ecosystem seventeen times.

The institutional roster has teeth. BlackRock expects to deploy its BUIDL tokenized Treasury fund on Arc, and DTCC plans work on tokenized assets in the second half of 2027. Good would be measurable balances and settlement activity from those integrations. If that happens, Circle stops looking like a crypto yield trade and starts looking like financial infrastructure.

Founder Jeremy Allaire still runs the company. He spent more than a decade dragging regulated stablecoins from fringe idea to public company. Management also secured a federal trust-bank charter and renewed the Coinbase distribution agreement on existing terms. Execution gets respect here.

What worries me

The current earnings machine hates falling rates. Circle estimates that a 1 percentage point rate decline, using June balances, would cut annual reserve income by about $737 million while lowering distribution costs by about $360 million. The net hit is roughly $377 million before operating expenses and taxes. Good would look like circulation growth and product revenue replacing that lost economics. They are not there yet.

The Arc headline flatters this year's guide. Circle raised expected 2026 other revenue to $310 million to $330 million from $150 million to $170 million, but about $180 million comes from recognizing token-presale revenue as milestones arrive. That is real revenue under the accounting rules. It is not proof of recurring network fees. Selling tickets before opening night tells me people are interested, not that the show can run for five seasons.

Distribution remains expensive. Circle spent $412.5 million on distribution, transaction, and related costs during the quarter, equal to about 59 cents of every revenue dollar. Good would be that burden falling below 50 cents as Circle holds more USDC directly and earns more software revenue.

Competition can attack from both ends. Tether (private) dominates stablecoin supply, Coinbase controls major distribution, banks can issue regulated digital dollars, and PayPal (PYPL) already has PYUSD. Arc also fights established blockchains with larger developer crowds. Circle has a strong seat. It does not own the table.

The stock gained roughly 63% in one month and now sits 229% above its $31 IPO price. That means the easy skepticism has left the building. Good price action would digest the run while fundamentals catch up. Buying after a sprint because the chart looks exciting is how people volunteer to become exit liquidity.

What would make this a buy

I want one of two setups.

The first is price. At $70 to $80, Circle would carry a market value near $17.8 billion to $20.3 billion, or roughly 6.1 to 7.0 times trailing revenue. That means the bill would still recognize USDC's quality while giving Arc room to stumble. Good would look like the stock building a stable floor in that zone, not falling through it because USDC circulation or the launch broke.

The second is proof. I would pay more if Arc reports at least $25 billion of annualized transaction volume within two quarters, non-reserve recurring revenue clears $50 million per quarter, and revenue after distribution costs grows above 20% without another token presale doing the lifting. I also want the retained margin above 45% and standard-accounting operating margin above 10%. Those marks would show the platform, not Treasury rates, starting to earn the valuation.

This is research and commentary, never a buy signal. Right now the company is compelling and the transaction is stingy.

What makes me walk away

I walk if Arc misses the September 16 launch or suffers a material security failure. Financial infrastructure gets one first impression. Good networks launch on time, keep money safe, and explain incidents without hiding behind jargon.

I walk if USDC circulation growth falls below 10% for two quarters while the reserve return rate keeps dropping. That would squeeze both variables feeding the core engine. Good would be circulation accelerating enough to offset lower yields.

I walk if non-reserve recurring revenue stays below 10% of total revenue through the end of 2027. That would mean Arc and the payments network became expensive demonstrations rather than a second business.

I walk if distribution and related costs stay above 60% of revenue or annual dilution stays above 5%. A network that pays away most of its economics and prints shares for the rest is not a tollbooth. It is charity with servers.

IPO Watch

Siyata PTT (private, proposed ticker PTT), expected September 8, no underwritten price range. It sells rugged push-to-talk devices and vehicle communications gear. This is a direct listing of up to 6.1 million shares, so existing holders can sell and the company receives no sale proceeds. I would write it up only if Nasdaq approves trading, liquidity proves real, and management shows enough cash to survive at least a year without another financing scramble.

Earnings On Deck

Oracle (ORCL), September 10 after the close. Watch cloud-infrastructure growth and the cash cost of building AI capacity. A good print means signed demand turns into revenue while free cash flow holds up. A bad print means the backlog is huge because the spending bill is huger.

Adobe (ADBE), September 10 after the close. Watch whether generative AI adds paying users and recurring revenue without hurting margins. A good print means AI strengthens pricing and retention. A bad print means usage grows while customers trade down or leave.

The verdict

Score: 73 out of 100. Bucket: watch. Circle earned points for USDC's scale, 151% transaction-volume growth, regulatory credibility, founder execution, real profit, strong cash resources, and an Arc launch backed by serious institutions. It lost points because total revenue grew only 6.6%, reserve interest still supplies 95.2% of revenue, distributors take a heavy cut, token-presale revenue flatters the near-term outlook, competition is brutal, and the stock already trades near 8.9 times sales after a 63% monthly run.

A candidate starts at 80. Circle could become the operating layer for regulated digital dollars. At today's price, investors already tipped the bartender for drinks Arc has not poured.

The DBTech read

Interest-rate sensitivity tells you how much a company's earnings change when rates move. Circle estimates a 1 percentage point decline could reduce annual reserve income by $737 million, partly offset by $360 million of lower partner costs. That leaves a rough $377 million hit before normal expenses and taxes, assuming USDC balances stay flat. Good sensitivity means another revenue stream grows fast enough to absorb the shock. When one outside number can swing profit this hard, track that number like it owes you money.

Sources and as-of date

• Price: $102.05 at the regular close on September 4, 2026, and $101.60 after hours on September 4, 2026, CNBC market data retrieved September 7, 2026 at 7:06 a.m. Eastern. U.S. markets are closed September 7 for Labor Day, so Friday's session is the latest regular trade.
• Circle second-quarter 2026 results, August 5, 2026.
• Circle Form 10-Q for the quarter ended June 30, 2026, filed August 5, 2026.
• Circle's official Arc validator and integration announcement, August 5, 2026.
• CNBC Circle quote page, market data through September 4, 2026.
• Oracle earnings announcement, September 2, 2026.
• Adobe earnings announcement, checked September 7, 2026.
• Siyata PTT Form F-1, filed August 6, 2026.
• Community and developer pulse covered 1 Reddit thread, 1 YouTube video, 4 Hacker News stories, and 5 GitHub items from August 8 through September 7, 2026. The useful signal came from Arc integration work in GitHub. The social sample was thin, so I did not use it to prove financial claims.
• Financial figures are as of June 30, 2026 unless stated otherwise. Market-derived calculations use the $102.05 close. Research and commentary only, never a buy signal.