An AI agent that wants to buy something, move money or open an account faces a problem that has nothing to do with model quality: nobody can tell whether it should be trusted. A startup has just raised $35 million to solve that problem.1 At the same time, exchanges are extending the hours during which an agent could act. This piece separates what the record verifies from what is still only announced. It also says plainly where the record is thin.
The trust layer: identity for software that spins up and disappears
Baselayer, a San Francisco AI startup that helps financial institutions verify businesses and assess fraud risk, has raised $35 million to extend its identity technology to AI agents. M13 led the Series A, with Picus Capital also participating.1 The founder-side framing comes from Jonathan Awad, who describes the core difficulty this way: "Agents spin up and they spin down. How can you trust this random one-task agent?"1
His commercial conclusion is blunt. Unless agents can establish that they act on behalf of legitimate people or businesses, he says, "agents will just get blocked everywhere."1 For practitioners building agents, this is the practical point. Capability is not what limits deployment in regulated finance. Provable authority is.
Awad also describes how slow that trust is to build. Relationships with financial institutions typically take 12 to 18 months to establish, and large merchant partnerships can take up to 24 months.1 He says Baselayer has "essentially streamlined 10 years' worth of selling into two years."1 That is a founder's claim about his own company, not an independently measured result. Treat it as such.
The clock is changing: 23-hour trading and event contracts
Nasdaq says the industry's transition to a 23-hour trading day is expected on Sunday, December 6, 2026. That is pending SIP readiness and SEC rule changes, and subject to SEC approval Nasdaq would add a 9 p.m. to 4 a.m. ET session.2 It is an expectation, not a done deal.
Cboe announced on September 30, 2026, at Robinhood's third annual HOOD Summit in Houston, that it plans to launch binary contracts tied to company-specific key performance indicators. The launch is expected in October 2026 and is subject to regulatory approval, with products tied to 23 companies. Robinhood would be the first retail broker to offer them.3 Our dossier holds no confirmation that the contracts have launched, so we do not report them as live.
Why this matters for AI builders: a market that runs nearly all day, with contracts on specific company metrics, is the environment where automated agents are most useful and where the identity problem above bites hardest. That link is our reading. None of the sources we hold says that agents will trade these products.
The retail broker underneath: what Robinhood's filings show
Robinhood is the retail channel in both announcements, so its SEC-filed numbers are the checked backbone here. These are canonical, fidelity-checked figures.
The company swung from a net loss of $541 million in fiscal 2023 to a net income of $1.411 billion in fiscal 2024 and $1.883 billion in fiscal 2025.4 Earnings per share moved from a loss of $0.61 in 2023 to $1.56 in 2024 and $2.05 in 2025.4
Cash tells a less steady story. It stood at $4.835 billion at the end of fiscal 2023, $4.332 billion at the end of 2024 and $4.261 billion at the end of 2025.4 It then rose to $5.012 billion in Q1 2026 and $5.362 billion in Q2 2026.4 Quarterly EPS in 2026 was $0.38 in Q1, $0.62 in Q2 and $0.51 in Q3.4 That Q3 figure is below both the Q2 2026 figure and the $0.61 reported for Q3 2025.4 The dossier does not explain why, so we do not guess.
Two seams are worth showing. First, the dossier carries two values for Robinhood's cash on 2025-09-30: $4.331 billion and $4.3 billion, a gap of roughly $31 million that looks like rounding rather than a real disagreement. The canonical filing figure is $4.331 billion.4 Second, the Q3 2026 cash figure of $5.829 billion is identical to the Q2 2023 figure of $5.829 billion.4 Both are marked canonical. A match to the dollar across three years is unusual enough that we flag it for a re-check rather than build any argument on it.
None of these filings measures AI agent activity. They show a broker that is profitable and holds cash, which is context for why it can push into new products. They are not evidence that agents are driving its growth.
The platform side: LY Corp raises guidance
LY Corp reported Q2 fiscal 2027 results on November 5, 2026, with revenue up 11.5% and adjusted EBITDA up 20.9% year on year. The company raised guidance across the board.5 Full-year revenue guidance rose from JPY2.24 trillion to JPY2.33 trillion, up 14.8%.5 Adjusted EBITDA guidance rose from JPY585 billion to JPY640 billion, up 28.9%.5 Adjusted EPS guidance rose from JPY30 to JPY35, and the annual dividend from JPY11 to JPY12 per share.5
These are very large numbers, and the sources give no outside yardstick for them, so the useful comparison is internal: guidance moved up by roughly JPY90 billion on revenue and JPY55 billion on adjusted EBITDA within one quarter's report.5 Commerce revenue is forecast at JPY1.06 trillion with adjusted EBITDA up 33.6%, and Strategic Business adjusted EBITDA at JPY160 billion, up 69.1%.5 The company expects sustainable double-digit profit growth next fiscal year, supported by structural cost reductions and revenue growth from cumulative businesses like LYP Premium and LINE Official Account.5
The headline of the source article ties the raise to AI, but the excerpt we hold does not break out AI revenue or say how much of the raise AI explains. So the AI link is the publisher's framing, not a number we can verify.
Regulation and the stablecoin rail
Agents that pay need money that settles. On the stablecoin side, Coinbase set a three-day deadline tied to Europe's MiCA rules. It converted non-compliant stablecoins (USDT, PayPal, Trump-linked) to USDC automatically and told users not to deposit them after October 30.6 The pattern is that regulators are deciding which payment instruments are acceptable. That shapes what an agent can legally hold and spend.
How much to trust the sources
Reliability varies. The NewsEOD feed that carried the Cboe announcement and the LY Corp summary has a measured record of 54% of 6,692 checked claims holding up.3,5 The Crunchbase News pieces sit at 49% of 41 checked claims, a much smaller sample.1 Neither figure means any specific claim here is wrong. It means the SEC-filed numbers above deserve more weight than company-sourced projections, and the quotes are one founder's view.
What to watch
- Whether the SEC approves the Nasdaq overnight session, and whether the December 6, 2026 expectation holds.2
- Whether Cboe's KPI contracts launch with Robinhood as planned, and for which of the 23 companies.3
- Whether Baselayer's 12 to 24 month sales cycles shorten as its agent-identity product reaches customers.1
- Whether LY Corp's upgraded guidance converts into reported results in coming quarters.5

