🎧 Listen to this Newsletter on the Sun app
Here's what's been trending:
Musk testified he was a "fool" for backing OpenAI. Then admitted xAI "partly" distilled OpenAI's technology.
NSA is testing Mythos to find vulnerabilities in Microsoft apps. The White House is drafting a memo to avoid agencies relying on a single AI vendor.
China blocked Meta's $2B Manus acquisition in a 54-character decree. Meta is preparing to unwind the deal.
Combined 2026 capex for Google, Meta, Amazon, and Microsoft is on track to hit $725B. Samsung warns the memory shortage will worsen in 2027.
KKR secured $10B+ to launch Helix Digital Infrastructure. The AI data center company will be led by ex-AWS CEO Adam Selipsky.
Startup Intros has a simple mission: to help tech founders navigate everything in the early-stage, from compliance to capital.
🎧 Listen to this Newsletter on the Sun app!
🔥 Startup Intros Events Coming Up
We’re excited to have you at the a16z Speedrun AMA with SUN 👋
Before the session, we put together a short audio walkthrough on SUN covering how Speedrun works, what the application process looks like, and what founders should keep in mind going in:
Give it a quick listen: https://sunapp.ai/course/a16z-speedrun
Bring your questions. We’ll see you in the room!
Other Events Coming Up:
May 12: Agentic AI Founders’ Night-Out: We are going beyond copilots and talking about the rise of Agent Orchestration! Register on Luma!
May 15: Caffeine & Capital: Our morning coffee meetup is back at Corgi Cafe in FiDi! Grab coffee, meet founders and investors & register on Luma!
⚖️ Musk v Altman: "I Was a Fool"
🗣️ The Testimony
Day One: Musk's lawyer told the jury that Altman "made a mockery" of OpenAI's public mission. Musk testified that he sued because "it is not okay to steal a charity" and that OpenAI's pivot sets a "concerning precedent for philanthropy."
Day Two: Musk called himself a "fool" for backing OpenAI, accusing Altman and Brockman of manipulating him into donating tens of millions. Internal 2017 emails show Musk sought majority voting control over OpenAI co-founders, withheld promised funding, and tried to poach researchers. The emails undercut his narrative of a passive donor betrayed by insiders.
Day Three: When asked whether xAI has ever distilled technology from OpenAI, Musk said the claim is "partly" true. That admission hands OpenAI's legal team exactly the kind of cross-contamination evidence they need to reframe the case from Musk-as-victim to Musk-as-beneficiary.
The Judge: Both Musk and Altman were told to "control your propensity to use social media to make things worse outside this courtroom." Musk had boosted a Ronan Farrow article about Altman's alleged deceptions on X the night before the trial. Some seated jurors expressed negative feelings about Musk and AI, but assured the court they could be impartial.
📉 The Missed Targets
The Numbers: OpenAI missed an internal goal of reaching 1B weekly active ChatGPT users by the end of 2025 and missed multiple monthly revenue targets earlier in 2026. OpenAI called the report "prime clickbait" and says its business is "firing on all cylinders."
The Pricing Shift: Internal projections show OpenAI expects $8/month ChatGPT Go subscribers to grow ~36x to 112M in 2026, while $20/month Plus subscribers will fall 80% to ~9M. The bet: volume at a lower price replaces premium subscribers who are downgrading.
⚡ The 10GW Pivot
The Compute Grab: OpenAI has signed contracts for 10GW of US AI compute capacity, with 3GW+ added in the past 90 days, hitting a goal it once aimed to reach by 2029. Sources say OpenAI has abandoned the Stargate JV in favor of large bilateral deals. Its guiding principle remains: "build more compute."
The AWS Expansion: Amazon and OpenAI expanded their partnership, making OpenAI's models and Codex available on Amazon Bedrock, a day after OpenAI ended Azure exclusivity with Microsoft.
Translation: Musk's own emails show he wanted control of OpenAI and withheld funding when he didn't get it. Then he admitted xAI "partly" distilled OpenAI's tech. When the plaintiff admits to the very behavior he's accusing the defendant of, what exactly is this trial deciding?
🏛️ Anthropic's Pentagon Problem
💰 The $900B Question
Anthropic has begun weighing a new funding round at $900B+, after previously resisting investor proposals at $800B+. This comes days after Google committed up to $40B and Amazon committed $33B total. But the valuation is running into a national security wall.
The Mythos Standoff: The White House opposes Anthropic's plan to expand access to Mythos for 70 additional companies and organizations due to security concerns. At the same time, the White House is developing guidance for government agencies to get around Anthropic's supply chain risk designation and onboard models like Mythos anyway. The administration is simultaneously blocking and seeking workarounds for the same model.
🔄 The Pentagon Shuffle
The DoD Diversification: Pentagon Chief Digital and AI Officer Cameron Stanley confirmed the DOD is expanding its use of Gemini, saying "overreliance on one vendor is never a good thing." Sources say the White House is preparing a wide-ranging AI policy memo outlining rules for national security agencies, including a ban on reliance on a single vendor. Goldman Sachs stopped its Hong Kong bankers from using Anthropic's models; Anthropic says its models were never officially "supported" in Hong Kong.
The NSA Connection: Sources say the NSA has been testing Anthropic's Mythos to identify vulnerabilities in Microsoft products and other widely used software. Meanwhile, AISI's cybersecurity analysis found that GPT-5.5 reaches a performance level similar to Mythos Preview and is the second model to solve a multi-step cyberattack simulation. Anthropic's competitive moat in national security AI just got thinner.
The Google Contradiction: Google signed a classified deal allowing the DOD to use Google's AI for "any lawful government purpose." Google told employees in a memo that it "proudly" works with the US military. Then, separately, Google withdrew from a $100M Pentagon drone swarm contest following an internal ethics review.
🔍 The Ethics Collision
The Chinese AI Probe: House Homeland Security and the House China Select Committee are probing Airbnb and Cursor maker Anysphere over their use of Chinese AI models. The investigations reflect growing concern that American companies are building on Chinese-developed technology, while the State Department warns of reverse distillation.
The Employee Revolt: More than 600 Google employees demanded Pichai bar the DOD from using Google's AI for classified work. Google's response: sign the deal anyway and tell employees it's a point of pride. The gap between what Google tells its employees and what it tells the Pentagon has never been wider.
Translation: The NSA is using Mythos to find vulnerabilities in Microsoft products while the White House drafts a memo telling agencies not to rely on a single AI vendor. It seems that the government is simultaneously deploying Anthropic's model for offensive security, while telling agencies to diversify away from it.
🇨🇳 China's AI Crackdown Cascade
❌ The Manus Kill
Beijing blocked Meta's $2B acquisition of Manus in a 54-character decree, reviewing whether the deal violated investment rules and ordering both sides to cancel the deal. Meta is preparing to unwind the deal, and Manus investors have already received their returns. Bloomberg says the move killed the "Manus template" for Chinese AI startups hoping to build in China and sell to American buyers. Manus had moved to Singapore in 2025, but that wasn't enough.
🔗 The Supply Chain Leverage
China Still Has Cards: Sources tell FT that 10%+ of Meta's revenue comes from Chinese advertisers, and Goertek, a key manufacturer of Meta AI glasses, sits in China's supply chain. Chinese billionaire Chen Tianqiao says his startup MiroMind is adding strict "firewalls" between its Chinese and US businesses in response to the Manus precedent. At least two China-based funds are using parallel fund structures to continue fundraising from US investors.
Meanwhile, China suspended new Level 4 autonomous vehicle licenses after 100+ Baidu Apollo Go robotaxis stalled, disrupting traffic in Wuhan in March. Even China's own AI deployments aren't immune to crackdowns.
🤫 The Quiet Dependency
The Other Direction: While Beijing blocks American acquisitions, Chinese companies are still building on American AI. Tencent employees used Claude Code to evaluate and fine-tune the company's Hy3 model. The US Commerce Department ordered chip equipment companies to halt some shipments to Hua Hong, China's second-largest chipmaker. But Beijing's GPU maker Moore Threads posted a $4.3M profit on $107.9M in Q1 revenue, up 155% YoY. And Lightelligence, the first mainland Chinese photonics chipmaker to IPO in Hong Kong, jumped nearly 400% on its debut.
Translation: China killed the Manus template and started building firewalls between Chinese and US AI businesses. But Tencent is still fine-tuning models with Claude Code. When Beijing is simultaneously blocking US acquisitions and relying on American AI tools to train its own models, the decoupling is more performance than reality.
💰 The $725B AI Capex Arms Race
📊 The Numbers
Earnings week delivered the clearest picture yet of Big Tech's AI commitment:
Alphabet: Q1 revenue up 22% to $109.9B. Google Cloud up 63% to $20B. Capex raised to $180-190B for 2026, expects capex to "significantly increase" in 2027. GOOG up 6%+ after hours.
Microsoft: Q3 revenue up 18% to $82.9B. AI business hit a $37B annual run rate, up 123%. Azure up 40%. Copilot has 20M+ seats.
Amazon: Q1 revenue $181.5B, up 17%. Net income up 77% to $30.3B. AWS up 28% to $37.6B, its fastest growth in 15 quarters. Capex: $44.2B in Q1 alone.
Meta: Q1 revenue up 33% to $56.3B. Net income up 61% to $26.8B. But capex raised to $125-145B, exceeding estimates. META dropped 6%+ after hours. Reality Labs lost $4B+, total losses now exceed $80B since late 2020.
Samsung: Q1 revenue up 69% to ~$90.2B. Operating profit up 756% to ~$38.5B, driven by AI memory demand. Samsung's founding family hit ~$45.5B net worth, up from ~$20.1B in 2025. Samsung projects the memory shortage will worsen in 2027 as customers are already placing orders for next year.
The Full Picture: FT analysis puts combined Q1 capex for Meta, Alphabet, Microsoft, and Amazon at a record $130B, with 2026 capex set to reach $725B, up 77% from $410B in 2025. Meanwhile, Nvidia's B300 server prices have nearly doubled to ~$1M each in China as a crackdown on chip smuggling dries up black market supply.
☁️ The Cloud Boom
The AI bet is paying off in cloud revenue. Google Cloud grew 63%, the fastest among hyperscalers. AWS grew 28%, its fastest in 15 quarters. Microsoft's AI business surpassed a $37B run rate, more than doubling in a year. Google's paid subscriptions hit 350M, up 25M QoQ. Gemini Enterprise paid MAUs grew 40% QoQ.
The Ad Boom: Digital ads are booming for Google and Meta, helped by AI tools for translation, real-time ad copy, and targeting. Google's ad revenue hit $77.25B, up from $66.89B in Q1 2025. Amazon's ad revenue rose 24% to $17.24B.
📉 The Market Split
The market is rewarding AI revenue and punishing AI spending without clear returns. Alphabet jumped 6% on cloud growth. Meta dropped 6% after raising capex again and reporting that family DAP fell 20M QoQ to 3.56B. Qualcomm jumped 12%+ despite weak smartphone revenue because a top hyperscaler will start using its chips. Seagate jumped 15%+ on AI storage demand. South Korea's stock market surpassed the UK's at $4.04T, fueled by Samsung and SK Hynix.
The Chip Tax: TSMC sold its remaining Arm shares for $231M, having invested ~$100M in Arm at its 2023 IPO. NXP jumped 20%+ on a stronger automotive chip forecast. The EU is drafting a revised Chips Act II, set for late May, that would let the Commission invest directly in large, cross-border fab projects.
Translation: Combined 2026 capex for Google, Meta, Amazon, and Microsoft is on track to hit $725B, up 77% from 2025. When the infrastructure bill is $725B and rising, how long before the market demands proof that the AI revenue justifies it?
🚀 Product Launch Quick Hits
OpenAI Symphony: Open-source spec for agent orchestration that turns a project board like Linear into a control plane for coding agents.
Anthropic Creative Connectors: Claude integrations for Blender, Autodesk, Adobe, and Ableton, embedding AI directly into professional creative workflows.
Gemini File Generation: Gemini app can now generate PDF, Word, Excel, Google Docs, Sheets, plain text, and Markdown files.
Poolside Laguna: US startup debuts Laguna XS.2 (33B open-weight) and Laguna M.1 (225B proprietary) for agentic coding.
Nvidia Nemotron 3 Nano Omni: Open multimodal model with 30B-A3B hybrid MoE architecture; Nemotron 3 family saw 50M+ downloads in the past year.
Snap AI Sponsored Snaps: Conversational ads in Snapchat's Chat tab where users talk to brand-specific AI agents for product recommendations.
Bloomberg ASKB: Bloomberg Terminal getting a chatbot-style interface overhaul, currently in beta for ~125K users.
Lovable Mobile: AI coding app launches on iOS and Android, letting users code via voice or text prompts and switch between PC and mobile.
Google Ask YouTube: Conversational AI search that generates pages with videos and text summaries, for Premium users in the US aged 18+.
AWS Quick AI Desktop: Desktop app for Amazon's Quick AI assistant, connecting tools and local files to build custom apps and dashboards.
Anthropic Claude Security: Opus 4.7-powered code vulnerability scanner, formerly Claude Code Security, now in public beta for Enterprise users.
Cloudflare Agent Accounts: AI agents can now create Cloudflare accounts, start paid subscriptions, register domains, and deploy apps on behalf of users.
Spotify Verified Badges: "Verified by Spotify" badge to identify human artists, based on artist presence on and off platform and consistent listener activity.
Meta Ads AI Connectors: Open beta letting advertisers use third-party AI tools to create, manage, and analyze campaigns through Meta's ad ecosystem.
👔 Personnel Quick Hits
45,800 tech layoffs in March: The worst month for reported tech job cuts in at least two years, per Layoffs.fyi.
eFishery founder sentenced to 9 years: Gibran Huzaifah convicted in $300M scandal that destroyed one of Southeast Asia's top startups.
Senate AI child safety bill advances: Senate Judiciary Committee unanimously endorses bipartisan bill requiring AI companies to implement age verification.
🌟 Editor's Note
At Startup Intros, our mission is to bring the latest founder-investor news straight to your inbox, keeping you ahead in the fast-paced world of Silicon Valley.
🎧 Listen to this Newsletter on the Sun app!
💭 Parting Thoughts
The thread connecting everything this week is the gap between what's being spent and what's being decided:
Big Tech's combined 2026 capex is on track to hit $725B, up 77% from 2025. Google Cloud grew 63%. Samsung memory profit surged 756%. But Meta dropped 6% for spending without clear consumer AI returns.
Musk admitted on the stand that xAI "partly" distilled OpenAI's technology while suing OpenAI for betraying its mission. OpenAI missed its growth targets but signed 10GW of compute and projects 112M subscribers at $8/month.
The NSA is using Mythos to hack into Microsoft products while the White House drafts a memo telling agencies not to rely on a single AI vendor. GPT-5.5 just matched Mythos on cyberattack simulations.
$725B in capex. Zero consensus on who controls the models it's building.
Till next time!
 | Dev Chandra |
Startup Intros Trending Thursday: Your trusted source for founder-investor insights, delivered with clarity and focus.