🌐 Trump-Xi Summit, Space Data Centers & The Rise of Vertical Video

🇨🇳 Trump Lands in Beijing with Tech Titans in Tow

The global business world turned its attention to Beijing this week as President Trump arrived for a high-stakes summit with Chinese President Xi Jinping. The Wall Street Journal's front-page coverage captured the moment: Trump and Xi dueling for wins at a summit that brought together an unprecedented delegation of American business leaders.

The roster read like a who's who of American industry: Elon Musk (Tesla), Tim Cook (Apple), Jensen Huang (Nvidia), Kelly Ortberg (Boeing), David Solomon (Goldman Sachs), Stephen Schwarzman (Blackstone), Larry Fink (BlackRock), Jane Fraser (Citi), and Dina Patel McCormack (Meta).

"President Trump visits Chinese President Xi Jinping in Beijing on Thursday and the preempting US spin is a search for stability. It's a nice idea as long as Mr. Trump doesn't think personal rapport can overcome Mr. Xi's anti-American purposes." — Wall Street Journal Editorial Board

✈️ The Mystery of Jensen's Late Arrival

Speculation swirled around Jensen Huang's participation in the summit. Early reporting suggested he wasn't invited, prompting a Truth Social correction from Trump clarifying that Jensen was indeed welcome—though perhaps not personally invited by the President himself. Jensen reportedly boarded Air Force One in Alaska, leading to questions about whether he was a late addition or simply joining from a convenient location.

Given that Nvidia sits at the center of US-China semiconductor tensions, Jensen's presence carried particular weight. His company faces complex export restrictions while remaining critical to China's AI ambitions—making him perhaps the most strategically important executive on the trip.

🎯 What's Actually at Stake?

According to the Wall Street Journal's editorial analysis, the summit agenda spans trade, technology, rare earths, and Iran. On the US side, the goal is stability—ratifying the status quo on tariffs with promises from Beijing not to weaponize rare earth supplies. For Xi, the weak Chinese economy creates incentives to cooperate, likely through promises to purchase more American farm goods and aircraft.

The catch? Xi has made similar promises before, and US farmers never regained their lost market share in China.

More concerning is what the Journal calls "the Venus fly trap"—Xi's push for the US to formally oppose Taiwanese independence (versus the current posture of simply not supporting it). The editorial warns this seemingly minor diplomatic tweak could disrupt decades of policy that has maintained peace in the Taiwan Strait.

"Taiwan is not the aggressor in the Taiwan Strait. Xi Jinping's fiction that opposing independence would indulge. Mr. Trump might not care about Taiwan's freedom or its example that a prosperous Chinese democracy is possible. But the president doesn't want a crisis on his watch, which would be an economic and geopolitical catastrophe."

🤖 The AI Policy Contradiction

A striking inconsistency emerged in US AI policy this week: while Japan's three megabanks gained access to Anthropic's Mythos model—marking the first expansion to an East Asian nation—the US continues restricting the most advanced models due to national security concerns. Secretary of the Treasury Scott Bessent personally informed Japanese officials of the decision.

Yet simultaneously, Jensen Huang flew to Beijing on Air Force One, presumably to discuss selling advanced AI chips to China. As one observer noted: "Tonight, the secretary of the Treasury is personally vetting each company that gets access to the most advanced US AI model because the risks are so high. Also tonight, Jensen Wong is flying on Air Force One to Beijing to sell China the AI chips it will use to develop its own Mythos-level AI model as soon as possible."

The administration's AI policy, in short, remains inconsistent and incoherent.

🛰️ Space Data Centers: The Next Frontier

SpaceX and Google are in talks to launch data centers in orbit, with the deal potentially boosting SpaceX's business ahead of its historic public listing. Google is exploring rocket launch partnerships as it expands efforts to put orbital data centers in space—an unproven technology that Elon Musk has called "the next frontier."

Google announced plans last year for Project Suncatcher, aiming to launch prototype satellites by 2027. CEO Sundar Pichai offered a cautious timeline: "There's no doubt to me that a decade or so away we'll be viewing it as a more normal way to build data centers."

SpaceX has filed applications to launch up to one million satellites for orbital data center ambitions—a massive leap from the roughly 10,000 Starlink satellites currently in orbit. The value proposition is clear: data centers powered by solar panels, eliminating the major power constraints faced by earthbound facilities.

However, significant engineering challenges remain, including thermal management through louvers, deployable radiators, and phase-change materials to handle heat in the vacuum of space.

🚀 Varda: The First Commercial Space Factory

In a groundbreaking development, Varda Space Industries announced that United Therapeutics—a $25 billion publicly traded pharmaceutical company—will use its own balance sheet to produce physical products in low Earth orbit.

This marks the first time in human history a large public company is manufacturing physical products in space for commercial purposes. Co-founder and President Delian Asparouhov explained the vision:

"We're the only company that makes physical products in space. Everybody else—it's all radio waves, internet, inference, photos of Earth. This is literally the first time there are physical products being moved in space as a way to generate revenue for a real business."

The application? Pharmaceutical crystallization in microgravity. In 2019, Merck demonstrated that crystallizing their blockbuster cancer drug Keytruda in space allowed them to shift administration from IV drip to subcutaneous injection—dramatically improving patient access.

Varda's technology enables similar transformations across therapeutic categories. The company expects to fly seven missions in the next year—higher cadence than everything the ISS and Space Shuttle have accomplished in all of human history combined.

📈 Inflation Concerns Resurface

Wall Street's anxiety about inflation intensified after Tuesday's consumer price index release. Producer prices rose 6% year-over-year in April, significantly above the 4.8% estimate. The closure of the Strait of Hormuz is having ripple effects across supply chains, with even unexpected products like Diet Coke facing shortages due to aluminum smelter disruptions in the Middle East.

Inflation expectations, measured by the gap between ordinary Treasury yields and Treasury Inflation-Protected Securities (TIPS), recently hit their highest level since October 2022. Investors now expect annual inflation to average around 2.7% over the next five years.

The concerning scenario: while AI-driven GDP growth concentrates in capex and data centers, the real economy is growing at just 0.1%. If the broader economy suffers from 2.7% inflation while remaining stagnant, the result is stagflation—a particularly challenging environment for policymakers and markets alike.

🎬 The Vertical Video Revolution

True Short emerged from stealth with $22 million in funding from Coastal Ventures, Wonder, and a roster of Hollywood investors to build AI-generated vertical films for mobile. The company, led by founder and CEO Nate Ter, is positioning itself as the bridge between Netflix and TikTok.

The inspiration? Chinese micro-dramas—vertical soap operas that generate more revenue than the box office in China. Some apps in the space reportedly generate $50 million per month in revenue from minute-long episodes consumed vertically on mobile devices.

True Short's initial wedge: a true crime app that became the number one true crime app within six months, reaching $3 million in annualized revenue. Investor Jeffrey Katzenberg saw the potential for the "next Hollywood" being built from the ground up, with creators leveraging AI to produce content at unprecedented speed and scale.

🤖 Humanoid Robots Reach New Milestones

Two major developments in robotics captured attention this week:

1. Unitree's GD1 Mecha
Chinese robotics company Unitree unveiled a transformer-style mech that can switch between bipedal and quadrupedal modes. The demonstration video showed the robot breaking through a brick wall—though the company's launch statement urged users to "please everyone be sure to use the robot in a friendly and safe manner."

2. Figure's Live Production Stream
Brett Adcock's Figure launched a live stream of humanoid robots running full eight-hour shifts in production, sorting packages at remarkable speed. The movement appeared significantly smoother than previous demonstrations, leading some observers to question whether the robots were being teleoperated—a claim the company would need to address with transparent demonstration or independent verification.

💰 Affirm Guides Higher, Doubles Down on Credit Excellence

Max Levchin returned to discuss Affirm's 2026 investor forum, where the company announced plans to reach $100 billion in GMV—doubling from current levels—while compounding at 25% annually and improving profitability margins from the current 3-4% range to 3.75% floor.

The striking detail: three years ago, Affirm projected 20% compound growth. Instead, the company delivered 30%+ growth. Now, as a larger company, Affirm is accelerating rather than decelerating—a testament to network effects and the flywheel dynamics of the business.

On the question of how rising inflation might affect demand, Levchin noted the company saw increased demand during the last inflation spike as consumers preferred to preserve cash flow. The key constraint? Credit performance, which Levchin called "job number zero."

"We grow as fast as credit performance will allow us to grow. Full stop. We have to print consistent credit returns, otherwise we lose credibility with our debt investors—and that's the most important thing."

🧬 Recursive Emerges with $650M to Build Self-Improving AI

Richard Socher announced Recursive, a new AI lab focused on building "recursive self-improving superintelligence to automate knowledge discovery." The company raised north of $650 million at a $4 billion pre-money valuation, led by Google Ventures with participation from Nvidia and AMD.

The team includes seven co-founders with exceptional pedigrees: Tim Rocktäschel (inventor of retrieval-augmented generation and Rainbow Teaming), Jeff Clune, Tim Shi (built Cresta to unicorn status), Josh Tobin (early OpenAI, led work on ChatGPT agents and Codex), Sam Yang (co-inventor of prompt engineering), Alexey Dosovitskiy (co-inventor of Vision Transformer), and Yandong Chen (led RL at Meta).

The focus? Enabling AI to automate the scientific method on itself—the ideation, implementation, and validation of ideas. Once successful at self-improvement, the technology can expand to other domains including software, digital work, and eventually physical applications like preclinical biology.

🌐 Vory Raises $22M to Modernize Grocery Operations

Third-generation grocer Brandon Hill is building the "self-driving operating system for supermarkets" with Vory, which raised $22 million to automate everything from inventory management to pricing to marketing for independent grocery stores.

The problem Vory solves is striking in its inefficiency: today, a pricing coordinator manually reviews invoices from hundreds of suppliers, uses a ruler and highlighter to identify price changes, then types updates into the ERP system—a 12-step process that Vory reduces to one click.

The platform includes electronic shelf labels—the same technology Walmart just spent $1 billion deploying across 4,600 stores—bringing enterprise-grade infrastructure to the 75% of the $1.5 trillion food and beverage retail market that isn't dominated by giants like Walmart and Amazon.

☁️ Nebius Posts 841% YoY Revenue Growth

Roman Churn, co-founder and Chief Business Officer of Nebius, joined to discuss the company's explosive earnings. The AI cloud division posted 841% year-over-year revenue growth, with earnings per share significantly beating expectations.

Nebius describes itself as a "vertically integrated, full-stack, AI-specialized cloud"—not simply bare metal compute or a marketplace, but a true cloud experience with flexibility, fast time-to-value, and multiple layers of developer tools.

The company's recent acquisitions of AEN AI and Clarify AI focus on inference optimization—one team specializing in model optimization (the engine of inference), the other in system optimization (routing, KV caching, orchestration across large compute clusters). The goal: maximize the efficiency of converting GPUs into tokens and value for customers.

As Churn explained: "Every three months you can lose a big piece of the pie if you don't move fast enough. It's not just about margins and earnings—unoptimized chips mean you won't get the yields you want, and the tokenomics will flip."

🎯 Looking Ahead

As this week's developments make clear, we're witnessing the convergence of geopolitical tensions, technological breakthroughs, and economic uncertainty. The Trump-Xi summit set the stage for how the US and China will navigate AI competition. Space-based infrastructure moved from science fiction to commercial reality. And inflation concerns are tempering enthusiasm about rate cuts—potentially setting up a challenging environment for the real economy even as AI investment soars.

The question facing policymakers, investors, and technologists alike: how do we manage the transition to an AI-driven economy without leaving the broader population behind in a stagflationary squeeze?