Bloomberg Tech
2026-05-18 · Hosted by Caroline Hyde, Ed Ludlow · Bloomberg / iHeartMedia
Executive Summary
Bloomberg Tech covers a tech-heavy Friday marked by a ~4% chip stock selloff as markets digested the Trump-Xi summit’s lack of chip deal progress, while the Nasdaq 100 fell 1.3% on the day. Bright spots included Figma’s blockbuster Q1 earnings with revenue growth accelerating to 46% year-over-year, defying fears of AI-driven disruption to the design stack. Other major stories: Cerebrus giving back gains after its historic 68% IPO debut, the OpenAI vs. Apple relationship fracturing with potential legal action under consideration, the OpenAI/Elon Musk trial closing arguments, and Figure’s humanoid robots completing 50+ hours of autonomous operation at human speed.
Key Stories & Changes
1. Trump-Xi Summit: No Chip Deal, Taiwan Tensions Flare
President Trump confirmed he discussed Nvidia H200 chips with Xi Jinping but no purchase agreements materialized — Trump said China “wants to develop their own”
Taiwan emerged as the “number one issue”: Xi Jinping asked Trump directly whether the US would defend Taiwan militarily; Trump was evasive, saying he was the only one who knew the answer
$14 billion US arms deal with Taiwan remains unconfirmed
Boeing officials were present in Beijing through the visit, viewed as a positive sign for potential aircraft purchases, but no Chinese confirmation
USTR’s Jameson Greer indicated chip sales were a “sovereign decision” by China
2. Chip Stocks Sell Off Despite Strong Fundamentals
Nvidia: -4% — Ended 7-day win streak; biggest decline since February; earnings next Wednesday
SOX Index: -4% — Broad chip selloff post-Trump-Xi summit
Micron: Down — Profit-taking after supply-concern-driven rally
Cerebrus: -4.8% — Profit-taking after 68% debut; still trading well above $185 IPO price
Cerebrus IPO backstory: Three VC backers (Benchmark, Eclipse, Foundation Capital) invested ~$25 million in 2016; now poised to make billions each
ARM and SoftBank made an acquisition attempt on Cerebrus in the weeks before its listing — rebuffed
3. Figma Defies AI Disruption Narrative — CEO Dylan Field
Q1 revenue growth accelerated to 46% year-over-year
Net dollar retention for >$10K ARR customers: 139%
Non-GAAP operating margin: 16%; Free cash flow: 27%
Company raised guidance after the quarter
Figma launched AI credit monetization in March — paying users can purchase additional credits beyond included allowance; some users dropped off but many were willing to pay
Figma Weave: New workflow tool connecting model outputs (images, video, 3D) for creative industries — architecture firm NBBJ cited as customer replacing expensive on-site photoshoots
Field’s thesis: As AI commoditizes code, “the layer above code” (design) becomes more valuable
On margins: Short-term margin pressure may come from aggressive growth investment, but company is targeting a “massive TAM” in design and advertising/marketing
4. Hyperscaler AI Capex — Goldman Sachs Analysis
Eric Sheridan (Goldman Sachs TMT research) noted hyperscaler revenue backlogs now exceed $900 billion combined across Alphabet and Amazon cloud divisions
CapEx continues with upward bias but growth was more muted this quarter — received positively by investors
Cloud margins surprised to the upside as non-AI workloads accelerate
Custom silicon (TPUs, Amazon chips) viewed as the “most underappreciated narrative” — competitive on price-to-performance and captures more cloud margin
Interdependence between foundation model companies and hyperscalers seen as a structural feature, not a bug — only a handful of companies can build at this scale
5. OpenAI vs. Apple: Relationship Fracturing
OpenAI is weighing possible legal action against Apple, arguing Apple hasn’t delivered the expected distribution benefits from their partnership
OpenAI’s technology remains “limited and hard for users to find” within Apple Intelligence according to the complaint
Apple is now testing Anthropic and Google Gemini for Siri alongside OpenAI
OpenAI CFO Sarah Fryer said the company needs more compute and already has raised $120+ billion; confirmed strong working relationship with Sam Altman
OpenAI vs. Elon Musk trial: Closing arguments wrapped Thursday; Musk is seeking $134 billion in damages, removal of Altman and Brockman, and reversion to nonprofit status; jury’s verdict is advisory
6. Figure Humanoid Robots: 50+ Hours Autonomous Operation
Figure completed 50+ hours of continuous autonomous package sorting with 3 robots working in shifts
Robots operate at roughly human speed (~3 seconds/package) with 90%+ success rate on barcode scans
CEO Brett Adcock confirmed absolutely no teleoperation despite social media skepticism
Manufacturing facility BACU producing 60-70 humanoid robots per week; on track to thousands annually
Company has $1+ billion cash on balance sheet
OpenAI reportedly exploring re-entry into robotics; Figure took the full-stack approach in-house after ending prior OpenAI software partnership
7. Additional Stories
Samsung management in last-minute talks with union leaders to avert 18-day strike that could cost $700 million/day and disrupt AI chip production
Bill Ackman’s Pershing Square disclosed new stake in Microsoft, arguing market underestimates its AI resilience; Microsoft down ~13% YTD
XAI (Elon Musk) launched Grok Build coding agent to compete with Anthropic’s Claude Code
Andy Jassy profile: Amazon CEO is “all over the details,” steering Amazon aggressively into AI and data centers
Trends Identified
1. Post-AI-Threat Vindication for Design/Software Tools
Figma’s 46% revenue acceleration represents a direct refutation of the “AI will destroy SaaS” narrative that has weighed on software stocks since 2024. Field’s thesis — that AI commoditizing code actually elevates the value of the design layer above it — is gaining market validation, and the success of credit monetization shows that AI features can become new revenue streams rather than margin destroyers.
2. The Interdependence of the AI Ecosystem
The Goldman Sachs analysis of hyperscalers underscores a key structural trend: the AI buildout is creating deep interdependence among a small number of companies that have the scale, capital, and custom silicon to compete. The blurring of lines between foundation model providers, cloud infrastructure, and enterprise customers (the Amazon-Anthropic-Goldman relationship cited on air) is becoming a defining feature of how AI economics will concentrate returns.
3. Chip Geopolitics: US-China Tech Decoupling Deepens
The Trump-Xi summit produced no chip deal, and China’s stated preference for domestic semiconductor development (protecting Huawei) alongside Taiwan’s elevated position as a flashpoint confirms that the bifurcation of the global chip supply chain is structural, not tactical. This creates sustained uncertainty for companies like Nvidia with large potential China market exposure.
4. Humanoid Robotics Crossing the Reliability Threshold
Figure’s 50-hour autonomous demonstration represents a qualitative shift from lab-to-live readiness for humanoid robots. The company’s claim of operating at human parity in speed and reliability for a specific industrial task — while manufacturing at scale — marks an inflection point that could accelerate enterprise adoption timelines.
5. AI IPO Cycle Reaching Public Markets
The Cerebrus IPO’s 68% debut and Apple/OpenAI’s fracturing partnership both reflect the same underlying dynamic: the AI buildout is now generating enough commercial momentum that the private → public transition is underway. The valuation discussions around SpaceX ($1.75 trillion), Anthropic ($900 billion), and OpenAI ($850 billion) suggest retail investors are finally getting access — though at significantly higher price levels than early institutional backers. —-
Sentiment Analysis
Overall Market Sentiment: Cautious / Consolidating
The day’s tone was one of digesting parabolic gains after summit disappointment, with AI fundamentals broadly intact but near-term catalysts clouded by geopolitics and rising yields.
Risk Factors Highlighted
US-China chip deal failure: No H200 chip purchases from China and Taiwan tensions mean Nvidia and other chipmakers face sustained China revenue uncertainty
Post-summit market reaction: Markets had priced in some positive chip deal outcome; the miss could trigger further chip stock consolidation heading into Nvidia earnings
Cerebrus concentration risk: $20 billion of its $25 billion backlog is a single OpenAI contract; diversification is the key execution risk post-IPO
OpenAI-Apple legal dispute: Could disrupt ChatGPT’s consumer distribution strategy at a critical time when rival models (Gemini, Anthropic) are catching up
Samsung strike risk: An 18-day work stoppage could cost $700 million/day and add to the already 25-26 supply chain shortages tracked by Seaport analysts
Musk trial outcome risk: If advisory jury sides with Musk, judge could pursue structural remedies for OpenAI, creating existential corporate uncertainty
Figure full-stack execution risk: Humanoid robots require simultaneous excellence in hardware, software, manufacturing, and data — vertical integration is a moat but also a single point of failure
Valuation extension in AI IPO cycle: Cerebrus trading at 134x trailing revenue and 5x+ Nvidia’s multiple; the public market is pricing years of flawless execution
This episode was covered in today’s The Market Signal — 2026-05-18, a cross-source synthesis of multiple podcast reports.