Markets are closed, but Pro Rata is open for business.

Also publishing on Monday (aka Patriots' Day here in Mass.), although I may push early to catch some of the marathon (usually stand near the top of Heartbreak Hill, if you're running by). OK, here we go...

Early-stage venture deals keep getting bigger, even once stripping out AI juggernauts. And that raises a multi-billion dollar question:

• If AI is supposed to supercharge productivity, why are startups raising more money? Shouldn't they need less?

By the numbers: Median early-stage round sizes are up year-over-year for most industry sectors, easily outpacing inflation, according to Q1 data compiled by PitchBook and provided to Axios.

• Not just in software, which includes many AI developers, but also in pharma/biotech (+29%), media (41%), IT hardware (71%), health-care systems (30.5%), and energy (79%).
• For PitchBook, "early-stage" means the company must be less than five years old and, if a series is specified, it should be an A or B.

The best explanation for this disconnect may be that AI hype hasn't yet translated into a ton of actionable use cases.

• The top ones so far are coding assistance and customer service automation, both of which can drive down startup costs but aren't necessarily game-changers.
• This is far different than when cloud computing took hold, dramatically cutting startup costs for both equipment and real estate (and, arguably, leading to the NYC tech boom).
• There also are some real-world cost increases, like Bay Area housing, but VC round sizes are climbing at a much faster clip.

The more cynical explanation may be that round size doesn't always match a startup's capital requirements.

• Venture capitalists often have their own calculus, tied to fund dynamics and ownership stake thresholds. Or the desire to use money as a competitive moat, particularly if they view a startup as operating in a winner-take-all market.
• Founders often lament taking more funding than they want, but either don't have better options or let future fundraising fears win out.
• Sometimes both sides are aligned on the gluttony, since larger checks can correspond to larger valuations.

Finally, there's the academic explanation of Jevons paradox, formulated by a British economist 160 years ago, whereby increased efficiency can lead to increased consumption.

• It's mostly been debated around areas like energy and agriculture, but soon may come to AI.

The bottom line: Venture capital talks a ton about the AI revolution, but so far is putting too much money where its mouth is.

Google lost a major antitrust case yesterday, with a federal judge ruling that it illegally dominates both the online advertising and ad-tech markets.

Why it's the BFD: This could lead to forced divestitures and sudden opportunity for runner-up rivals, were Google to also lose its appeal.

• Plus, Google may need to recalibrate some of its AI and cloud investments, which are largely funded by its ad business.

The bottom line: "The court drew a distinction between the markets for advertising exchanges and ad servers, where it found Google has an illegal monopoly, and the general market for display ads online (i.e., the old DoubleClick acquisition), where it found Google does not." — Sara Fischer & Scott Rosenberg, Axios

• Goodfire, an AI interpretability research company, raised $50m. Menlo Ventures led, joined by Lightspeed, B Capital, and Anthrolpic. axios.link/42m988M

• Meadow, a financial services provider for college students, raised $14m in Series A funding. Matrix Partners led, joined by insiders Susa Ventures, Giant Ventures, Treble Capital, and GoGlobal Ventures. axios.link/4jzbFCJ