Feds Investigate $5B in Suspicious…

1. Key Themes

Regulatory scrutiny of prediction markets is intensifying

Kalshi's suspicious trading pattern is drawing federal attention, signaling that the fast-growing prediction market industry may face a regulatory reckoning.

"Kalshi is drawing regulatory scrutiny again after nearly one million trades in a single ether market were placed in nearly identical $5,500 amounts, accounting for more than $5 billion in volume. The CFTC is examining the activity, while Kalshi says it found no evidence of wash trading."

The AI capital pouring continues across every layer of the stack

Massive fundings this cycle span infrastructure (cloud, hardware), data/training tooling, vertical AI agents, and even wildly speculative "self-improving" models — suggesting capital is still chasing AI broadly rather than concentrating on proven use cases.

"Mirendil, a San Francisco startup founded this year that develops self-improving AI models designed to advance with little or no human input, is in talks to raise up to $1 billion at a $5 billion post-money valuation." "Snorkel AI... raised a $350 million round at a $3.5 billion post-money valuation." "Cyera... raised a $400 million extension of its Series G from Goldman Sachs Alternatives. The company has raised more than $2 billion in total funding."

Government and courts are grappling with AI/tech power asymmetries

Multiple items show public institutions pushing back on tech's information and market advantages — from lawsuits over data feeds to a president literally rebranding AI terminology.

"San Francisco just sued Trump Media over Truth Social's $100,000-a-month data feed, arguing that giving paying subscribers fractional-second early access to President Trump's potentially market-moving posts violates California's Unfair Competition Law." "President Donald Trump said the U.S. government will begin referring to AI as 'super intelligence' in official documents after polling the public on alternative names."

Alternative education and non-traditional talent pipelines are gaining institutional backing

Big tech and VC are directly funding new pathways outside the traditional university system, betting on AI-native skill-building.

"Andreessen Horowitz is investing $35 million in an unaccredited, two-year San Francisco college alternative for 16- to 22-year-olds, with Anthropic, Google, Meta, Nvidia, and OpenAI providing software, compute, office access, and internships."

2. Contrarian Perspectives

A sitting head of state publicly admitting uncertainty about AI policy

Most political leaders project confidence on tech policy; Mitsotakis broke from that norm by openly admitting he lacks answers — a candid, non-obvious stance for a leader on an investment-courting trip.

"He said openly that he doesn't have answers to many of the AI questions leaders around the world are more privately debating." "'We're Already Fighting Yesterday's Battle': Greece's Prime Minster Gets Candid About AI"

Regulators, not just markets, may become the check on prediction-market growth

Rather than assuming prediction markets like Kalshi will simply scale unchecked (the prevailing fintech narrative), the CFTC probe suggests structural exploitation (potential wash trading at massive scale) could stall the category's momentum.

"nearly one million trades in a single ether market were placed in nearly identical $5,500 amounts, accounting for more than $5 billion in volume."

3. Companies Identified

Kalshi — Prediction market platform. Mentioned as subject of CFTC investigation into suspicious trading patterns. "The CFTC is examining the activity, while Kalshi says it found no evidence of wash trading."
Trump Media — Owner of Truth Social. Sued by San Francisco over a paid data feed giving early access to Trump's posts. "San Francisco just sued Trump Media over Truth Social's $100,000-a-month data feed."
Mirendil — Startup building self-improving AI models. Notable for enormous early valuation ambition. "...in talks to raise up to $1 billion at a $5 billion post-money valuation."
Snorkel AI — AI training data/RL environment company. Notable for reaching $3.5B valuation with broad investor syndicate. "raised a $350 million round at a $3.5 billion post-money valuation."
Cyera — Israeli data security startup. Notable for scale of funding and total raised. "has raised more than $2 billion in total funding."
Heidi Health — AI clinical documentation startup. Notable for blended equity/nondilutive financing structure. "also secured $240 million in nondilutive growth financing from General Catalyst's Customer Value Fund."
RealSense — Intel spinout robotics/computer vision company. Notable rapid value creation post-spinout. "raised $50 million after its spinout, tripled revenue, and has been profitable for the past two quarters," being acquired by Cognex for "roughly $600 million."
Iambic Therapeutics — AI drug discovery startup filing for IPO. Notable as a case of AI-bio convergence reaching public markets. "filed to go public on Nasdaq under the ticker IAM."
Waymo — Robotaxi company expanding teen access. Notable for growth-strategy shift beyond geography. "opening its robotaxi service to teenagers as young as 13 in Nashville... as the Alphabet-owned company looks to drive growth beyond geographic expansion."
Amazon — Drone delivery expansion facing community backlash. Notable tension between scaling ambition and local externalities. "drawing backlash in Richardson, Texas, where residents complain of dozens - sometimes more than 100 - noisy flights a day."
Microsoft / Ninja Theory — Gaming layoffs and possible studio closure. Notable as sign of continued gaming industry contraction. "Ninja Theory... may close after two attempts to sell it fell through."

4. People Identified

Kyriakos Mitsotakis — Prime Minister of Greece. Notable for candidly discussing AI uncertainty and pitching Greece as a tech destination. "I think it's another indication that the economy is doing well and that Greece is no longer treated as a special case."
Donald Trump — U.S. President. Notable for directing a rebrand of AI terminology and considering an "AI czar." "the U.S. government will begin referring to AI as 'super intelligence' in official documents."
Scott Bessent — Treasury Secretary. Notable as candidate for new White House AI czar role. "is being considered for President Donald Trump's new 'AI czar' role, alongside... Michael Kratsios, OPM Director Scott Kupor, and National Cyber Director Sean Cairncross."

5. Operating Insights

Blended capital stacks are becoming standard for capital-intensive AI health/infra startups. Heidi Health paired traditional equity ($100M Series C) with $240M in nondilutive growth financing — a model founders in capital-intensive verticals should study to reduce dilution while scaling.
Speed to profitability post-spinout can command premium exits. RealSense went from Intel spinout to a $600M acquisition in just 14 months by tripling revenue and reaching profitability quickly — a signal that lean, fast-execution spinouts can outperform slower-moving independent startups in exit speed.
Regulatory risk is now a first-order consideration for fintech/market-structure startups, not an afterthought — Kalshi's situation shows how quickly unusual trading volume can trigger existential scrutiny even absent proven wrongdoing.

6. Overlooked Insights

Big Tech is directly subsidizing alternative education/talent pipelines, not just funding startups — a16z's $35M investment in a non-accredited college alternative, backed by compute/internship access from Anthropic, Google, Meta, Nvidia, and OpenAI, hints at an emerging talent pipeline strategy that bypasses traditional universities entirely, potentially reshaping how AI talent is sourced long-term.
Detection-evasion arms races are escalating in unexpected sectors — the mention of AI cheating detection tools like Pangram colliding with counter-tools designed to evade them foreshadows a broader "authenticity verification" market opportunity beyond just academia (e.g., hiring, content, journalism).