TBPN订阅

Chip Wars and Google Earnings Recap

Happy Thursday.

We’re live on YouTube and 𝕏.

The current thing in tech and business is Jensen’s close relationship with a Chinese business exec that was actually a spy.

Today’s Lineup

Kino AI Co-Founder Lucas Igel at 11:50 AM
BIG3 Co-Founder & CEO Ice Cube at 12:00 PM
UNA Sports Group Co-Founder, President & COO David Rader at 12:25 PM
NFL Hall of Famer Jared Allen at 12:25 PM
Serval Founder & CEO Jake Stauch at 12:35 PM
Eve Legal Founder & CEO Jay Madheswaran at 12:45 PM
Vibe.co Co-Founder & CEO Arthur Querou at 12:55 PM

Run of Show

Chip Wars and Google Earnings Recap

The New York Times has an exclusive that Jensen Huang formed a close relationship with a Chinese business exec who U.S. intelligence officials believe was gathering intelligence for Beijing. Here’s the full story.

We actuallydiscussed this a year ago on the show.

Google Earnings recap:

Google stock is cratering (down 12.5%) after an earnings miss. Here are the results:

Here are 5 key takeaways:

Cloud is a monster business.

There was some reporting by The Information that Larry Page considered pulling the plug on Google Cloud in 2018 because it wasn’t meeting expectations. Completely different world now. Revenue doubled to $30.7B, operating profit tripled to $11.1B, and operating margin reached 36.2%. Cloud backlog hit $561B, with just over 55% expected to be recognized as revenue within two years.

Search is growing, but monetization is the concern.

Search advertising growth slowed from 17% last quarter to 12%. AI is accelerating query volume, but AI Mode searches currently monetize less effectively than traditional searches. Will be tricky to get back to a world where the first full page of pixels you see when you type a search is fully paid for by advertisers.

Google is spending roughly $200B on AI infrastructure this year.

2026 CapEx guidance of $195B–$205B with another significant increase expected in 2027. Q3 spending alone was $51.4B, let’s go.

The earnings miss was partly a paper loss on SpaceX.

An approximately $8.7B loss on equity investments reduced EPS by roughly $0.55. Excluding that impact, EPS would have been approximately $2.67, still below the roughly $3.01 Wall Street consensus.

The Gemini app crossed 1B MAU.

Google’s first-party model APIs process nearly 32 billion tokens per minute, up 40% quarter over quarter. The company also announced Gemini 4 Carbon, its newest frontier model.

RABBIT HOLE

The rise of retail proprietary trading firms is one of the stranger developments in modern finance. Traditionally, a prop trading firm meant a sophisticated institution like Jane Street or Citadel Securities, which employs professional traders to invest the firm’s own capital. But over the past decade, an entirely different industry has emerged, promising ordinary people the opportunity to become professional traders without risking much of their own money. Companies like Topstep, FTMO, FundingPips, and Alpha Capital allow customers to pay an entry fee, often less than $100, to participate in trading competitions using simulated capital. Pass the evaluation, and you can graduate to a supposedly “funded” account, sometimes advertised as controlling $100,000 or $200,000, with the opportunity to receive as much as 90% of your trading profits. According to a December 2025 Bloomberg investigation, the industry has exploded from approximately 10 firms five years earlier to more than 400, while three of the largest companies claim a combined six million customers.

The model was pioneered in part by Topstep, founded by former Chicago futures trader Michael Patak, whose predecessor business dates to 2009. Patak had experienced significant personal trading losses and developed a system to evaluate aspiring traders through simulations before allowing them to manage real capital. In 2012, Topstep introduced its Trading Combine, borrowing the concept from professional sports, where prospective athletes compete to demonstrate their skills. The idea was to invert the traditional recruitment model: instead of a trading firm paying to identify and train talent, aspiring traders would pay the firm for the opportunity to prove themselves. This turned out to be an extraordinarily scalable business. By 2020, Topstep was reportedly generating more than $25 million in annual revenue and over $3.7 million in EBITDA. The company eventually expanded into trading software, educational media, and brokerage services, while hundreds of competitors copied its evaluation model, particularly in foreign exchange and futures trading.

The catch is that most customers aren’t actually trading the firm’s money, even after they become “funded.” They’re trading simulated accounts whose profits and losses exist only on a computer screen, although successful participants can receive genuine cash payouts based on those fictional returns. An advertised $100,000 account might have a maximum permitted drawdown of just $3,000, meaning the trader is working with a much smaller effective risk budget than the headline number suggests. To qualify for payouts, traders must navigate strict rules governing daily losses, position sizing, consistency, and sometimes restrictions on trading during major news events or holding positions over weekends. Violate a rule and the account can be terminated, requiring another purchase to start over. According to Bloomberg, industry software provider FPFX Technologies estimates that only 4% of customers ever receive a payout. Topstep’s own published statistics indicate that just 0.71% of its Express Funded Account participants graduated to live trading in 2025. The central economic insight is that these companies can generate substantial revenue from evaluation fees without ever deploying much real trading capital.

Bloomberg’s investigation illustrates how this structure can become psychologically addictive. One London accountant, Samim Sharif, estimated that he had spent approximately $10,000 on trading challenges with a firm called PipFarm, repeatedly purchasing new attempts after failing. A former PipFarm employee described customers trying to recover their losses by buying yet another challenge, creating a cycle reminiscent of gambling. Another trader interviewed by Bloomberg acknowledged taking maximum allowable risks precisely because the account was simulated: if he failed, he could simply purchase another one. The industry nevertheless has real success stories, including customers who have collected substantial payouts and even quit their jobs to trade. Firms argue that their products offer valuable education, enforce risk management, and democratize access to an otherwise exclusive profession. But there is an obvious tension in the business model: while customers aspire to become consistently profitable traders, the companies can benefit financially from repeated failures. The UK’s Financial Conduct Authority has characterized the sector as high-risk and largely unregulated, while Belgian regulators have warned that gamified trading can encourage overconfidence and reckless behavior.

The regulatory history is equally revealing. In 2023, the US Commodity Futures Trading Commission brought a major fraud case against MyForexFunds, alleging that the company had misled customers about whether they were trading real capital and profited from rules designed to make traders fail. But the case collapsed spectacularly in May 2025 when a federal judge dismissed it with prejudice and sanctioned the CFTC for misconduct, leaving the underlying fraud allegations unproven. Meanwhile, FTMO stopped accepting new American customers in 2024, strengthened its disclosures that trading was simulated, and subsequently acquired regulated broker OANDA. Topstep has likewise moved toward conventional brokerage services. These developments suggest an industry attempting to mature from a lightly regulated, simulation-based business into something resembling mainstream financial services. Yet the fundamental question remains unresolved: are retail prop firms primarily discovering talented traders, or are they selling millions of aspiring traders an expensive, recurring opportunity to pretend they’re managing Wall Street money? The evidence suggests that while genuine trading opportunities and real payouts exist, the industry’s most reliable commercial product may be the dream of becoming a professional trader rather than professional trading itself.

Clip Spotlight: Mark Gurman believes famously M&A-shy Apple could be more trigger-happy with acquisitions under John Ternus.

The biggest signal is new M&A leadership at the company.

Mark broke the story a few weeks ago that Adrian Perica, the former head of M&A, was moved to a new position under Eddy Cue on the services side of the business.

Steve Smith, Perica's former deputy, is now Apple's head of M&A.

Mark believes this shakeup happened for two reasons:

1. Eddy Cue will eventually need a successor, and Perica is being positioned as a candidate.

2. John Ternus and the hardware team have had a decades-long gripe that Apple buys no one, and now they're finally in charge.

"The joke with Apple M&A is they do nothing. It's the easiest job in the world. This is how the hardware guys think of it."

"Ternus wants someone who will get cracking."