The Frontier Wire - Weekly Roundup: April 18 - April 25, 2026
The Frontier Wire - Weekly Roundup: April 18 - April 25, 2026
Top stories from the past week. Full edition at thefrontier.fm.
AI & Tech
Musk buys Cursor to win AI coding war
TL;DR
- SpaceX secures $60B option to buy Cursor, pairing its AI models with a top coding platform.
- The deal gives Musk access to developer data needed to train self-improving AI.
- Google and Amazon respond with founder-led teams and massive compute deals.
The Story
SpaceX is no longer just a rocket company. It’s now a central player in the AI arms race after securing a $60 billion option to acquire Cursor, the AI-powered coding environment favored by elite developers. If the acquisition closes, it will pair Cursor’s real-time developer workflows with xAI’s Colossus supercomputer - a cluster of millions of H100-equivalent GPUs. This isn’t a vanity purchase. It’s a bid to close the gap between Musk’s AI ambitions and his underwhelming track record.
According to Nathaniel Whittemore on The AI Daily Brief, xAI has struggled to release models competitive with OpenAI’s GPT-4.7 or Anthropic’s Opus. It lacks both mindshare and production data. Cursor changes that. Its Composer 2 model already ranks among the best in coding benchmarks, and its IDE captures granular logs of how developers use AI - the kind of behavioral data that fuels recursive self-improvement. By folding Cursor into the xAI stack, Musk gains a live pipeline of code traces to train smarter models.
The $10 billion collaboration fee SpaceX is paying even if the deal fails underscores the urgency. David Sacks on All-In argued it functions like a breakup fee, locking in access during a critical window. While Anthropic and OpenAI throttle usage due to compute constraints, a SpaceX-backed Cursor would have near-unlimited training capacity. “This is vertical integration at scale,” Sacks said. “Hardware, model, and interface - all under one roof.”
“The real value isn’t the IDE. It’s the data pipeline - the logs, the keystrokes, the corrections. That’s the training set for the next leap.”
- Jason Calacanis, All-In with Chamath, Jason, Sacks & Friedberg
Google sees the threat. Sergey Brin has returned to lead a strike team at DeepMind focused on turning Gemini into a primary developer tool. Internal data shows Anthropic’s models now write nearly 100% of their own code, while Google lags at around 50%. Brin’s team is training models on Google’s private codebase to accelerate internal agentic development - a direct response to Musk’s move.
Amazon isn’t waiting either. It’s committing $25 billion to Anthropic in a compute-for-equity deal. AWS will provide 5 gigawatts of power and Trainium chips; Anthropic delivers models. The deal shores up Anthropic’s capacity while giving Amazon a hedge against OpenAI. It also follows Amazon’s shuttering of its own AGI lab - a tacit admission that outsourced intelligence is now the only viable path.
Meanwhile, Apple’s leadership shift adds another layer. Tim Cook exits with a $4 trillion valuation but a legacy of AI stagnation. John Ternus, his successor, inherits a company strong in hardware but weak in AI integration. While Apple bet on privacy and on-device processing, that strategy left Siri irrelevant and the Mac Mini an accidental hub for open-source agents like OpenClaw.
Ternus must now decide whether Apple becomes a platform for others’ models - toggling between Grok, Claude, and ChatGPT - or finally builds its own AI soul. The SpaceX-Cursor deal raises the stakes. Musk isn’t just buying software. He’s buying time, data, and leverage in a race where the winner doesn’t just sell tools - they define how software gets made.
Sources: All-In with Chamath, Jason, Sacks & Friedberg, This Week in Startups, The AI Daily Brief: Artificial Intelligence News and Analysis, The AI Daily Brief: Artificial Intelligence News and Analysis
Mythos leak forces AI safety reckoning
TL;DR
- Anthropic’s ultra-powerful Mythos AI leaked to hackers, exposing power grids and banks to automated attacks.
- Governments compare the threat to a Strait of Hormuz blockade - yet no agency regulates such models.
- Sam Altman mocks Anthropic’s ‘fear-based marketing’ as the industry’s self-policing facade cracks.
The Story
Anthropic’s most advanced AI, Mythos, was never meant to escape. Designed to find hidden software vulnerabilities - like a 27-year-old OpenBSD bug - it was shared with just 11 US institutions and the UK government. That control failed. A hacker group on Discord accessed the model, weaponizing a tool built for defense into a blueprint for cyberwarfare.
The breach reveals a dangerous illusion: that AI safety can be maintained through secrecy and elite access. The Bank of England warns Mythos could “crack the whole cyber risk world open.” Canada’s finance minister equates the threat to a blockade of the Strait of Hormuz. These are not hypotheticals. The model’s ability to automate attacks on critical infrastructure means the damage is already in motion.
On The AI Daily Brief, Nathaniel Whittemore noted the leak undercuts Anthropic’s “safety-first” branding. If the model is as dangerous as claimed, its exposure via a third-party vendor is a catastrophic failure. Sam Altman seized the moment, calling Anthropic’s strategy “fear-based marketing” - selling $100 million “bomb shelters” while building the bomb.
“We are betting the stability of the global financial system on the server security of a single company.”
- Krystal Ball, Breaking Points
The regulatory void is glaring. No body reviews or licenses models like Mythos. A medical drug requires years of trials for a small patient group; a tool that can collapse a bank’s firewall is governed by a startup CEO’s discretion. Krystal Ball argues for a presidential advisory body to set standards - “transparent review,” not self-policing.
Meanwhile, the Pentagon labels Anthropic a supply chain risk, while the NSA uses Mythos in secret. This contradiction, reported by Axios, shows inter-agency chaos. President Trump signaled détente after a White House meeting, calling the team “very smart,” even as his administration previously blacklisted Claude.
“If Mythos can map a roadmap for hackers to attack power grids, the time for voluntary safety pledges has passed.”
- Saagar Enjeti, Breaking Points
The myth of containment is over. AI this powerful cannot be gated by promises. The leak proves the need for binding oversight - before the next exploit hits a live grid.
Sources: Breaking Points with Krystal and Saagar, The AI Daily Brief: Artificial Intelligence News and Analysis, The Intelligence from The Economist, The AI Daily Brief: Artificial Intelligence News and Analysis
Politics
SPLC indicted in hate scheme
TL;DR
- Federal indictment alleges SPLC paid extremists to fabricate hate incidents and drive donations.
- The DOJ is auditing SPLC’s $500M endowment amid money laundering charges.
- Media leaks aim to discredit investigators probing hidden FBI files.
The Story
A federal grand jury in Alabama has returned an 11-count indictment against the Southern Poverty Law Center. The charges allege the organization didn’t just track extremism - it funded and manufactured it. According to Acting Attorney General Todd Blanche, the SPLC channeled money through shell entities like ‘Rare Books Warehouse’ to pay members of the KKK and National Alliance. These were not informants. They were paid provocateurs.
The indictment ties SPLC funding directly to the 2017 Charlottesville ‘Unite the Right’ rally. That event became a national narrative, later used to propel Joe Biden’s 2020 campaign. Hosts Adam Curry and John C. Dvorak on the No Agenda Show argue the SPLC operates as a high-margin protection racket: donate or stay on the hate list. With $500 million in assets, the group avoided scrutiny for years.
“If the source of truth is a paid provocateur, the entire media infrastructure collapses.”
- Adam Curry, No Agenda Show
Now, the DOJ is conducting a forensic audit of the SPLC’s finances. The case expands beyond fundraising. FBI Director Kash Patel has filed a $250 million defamation lawsuit against The Atlantic. He claims reporters fabricated stories about his sobriety to derail his investigation into sealed FBI files. Those files, he says, contain evidence of election interference and FISA abuses dating back a decade.
Former officials like John Brennan have dismissed the probe as ‘retribution.’ But Dvorak notes the irony: Brennan, who oversaw expansive surveillance programs, now defends the privacy of bureaucrats. The conflict is clear - an entrenched bureaucracy versus a new administration intent on exposure.
“The smear machine is a desperate attempt to prevent these documents from reaching the public.”
- Kash Patel, All-In with Chamath, Jason, Sacks & Friedberg
The implications go beyond one organization. If the SPLC - a cornerstone of global fact-checking networks - is charged with inventing the threats it claims to fight, the credibility of the entire ecosystem is at risk. The case is no longer just about fraud. It’s about who gets to define hate - and why.
Sources: All-In with Chamath, Jason, Sacks & Friedberg, No Agenda Show
Business
Thoma Bravo exit signals SaaS debt crack
TL;DR
- Thoma Bravo’s Medallia fire sale reveals the collapse of debt-leveraged SaaS models.
- AI agents are replacing per-seat software, eroding legacy valuations.
- Shadow banks now bet against the private credit boom they fueled.
The Story
Thoma Bravo is handing Medallia back to creditors. The move wipes out billions in equity and signals the end of an era: the debt-fueled SaaS growth model is cracking under AI-driven deflation.
On All-In, David Friedberg argued that the predictable cash flows once seen as bulletproof for leveraged buyouts are evaporating. AI agents now perform tasks - like customer surveys and HR functions - for a fraction of the cost of licensed software seats. Chamath Palihapitiya added that tokens are cheaper than licenses, and enterprises are switching fast. Salesforce and ServiceNow have already seen sharp valuation compression as investors abandon the old ‘net dollar retention’ gospel.
“The per-seat model is dead. If you can spin up an AI agent for pennies, why pay $50 a month per seat?”
- David Friedberg, All-In with Chamath, Jason, Sacks & Friedberg
Meanwhile, private credit firms - like those behind Tricolor’s subprime auto loans - are repeating the 2008 playbook. As Mia Wong detailed on Behind the Bastards, these non-bank lenders bundle risky loans into tranches, using them as collateral for more debt. JP Morgan already lost $170 million on such instruments. Redemption gates - capping withdrawals at 5% per quarter - mean investor money is frozen when panic hits.
Wall Street isn’t waiting for the collapse. JP Morgan and Barclays are now trading credit default swaps on Blackstone and Apollo funds. They’re not just funding the shadow banks - they’re betting on their failure. As Molly Conger noted, the most profitable move in today’s economy isn’t building, but betting on implosion.
“It’s a casino. The house is now selling insurance on its own collapse.”
- Molly Conger, Behind the Bastards
The common thread? Overleveraged systems assuming perpetual growth. Whether in SaaS or auto loans, the model relied on stable cash flows. AI and economic reality have broken that assumption. The result isn’t just a market correction - it’s a structural unwind of the post-2008 financial imagination.
Sources: Behind the Bastards, All-In with Chamath, Jason, Sacks & Friedberg
Iran’s strait blockade triggers global energy crisis
TL;DR
- Lloyd’s insurance hikes, not just Iranian missiles, are blocking the Strait of Hormuz.
- $300 oil and global food shortages loom as fertilizer production stalls.
- US missile defenses are half-empty, leaving no room for further escalation.
The Story
The Strait of Hormuz is closed, but the real barrier isn’t Iranian warboats - it’s the cost of insurance. Adam Curry on No Agenda Show argues that Lloyd’s of London raised war risk premiums five-fold, forcing shippers to choose between ruinous premiums or idle fleets. The Joint War Committee’s move effectively shuttered the strait, even as Iran demands tolls and the US enforces a naval blockade.
“The real story is not the Iranian fast boats - it’s the five-fold increase in war risk premiums by Lloyd’s. That’s what’s stopping trade.”
- Adam Curry, No Agenda Show
Meanwhile, the energy shock is rippling through global supply chains. Ole Hansen on Macro Voices warns that natural gas shortages are already slashing fertilizer output. With nitrogen fertilizer tied to gas prices, farmers from Kansas to Punjab are under-fertilizing crops. Hansen predicts a 2026 yield collapse in wheat and corn - setting the stage for a global food inflation surge by 2027.
The US response has been erratic. Trump’s blockade, meant to pressure Iran, backfired when Pakistani mediators arranged a ceasefire deal - only for Trump to extend the blockade at the last minute. Iran responded by seizing three container ships. Jeremy Scahill on Breaking Points calls this “whiplash diplomacy” - a term that captures how Trump’s unpredictability is undermining his own leverage.
“The Pentagon burned through 50% of its high-end missile interceptors in 38 days. We’re not winning - we’re depleting.”
- Ryan Grim, Breaking Points
Behind the scenes, the US military is exposed. Internal DIA reports confirm 60% of Iran’s navy and two-thirds of its air force remain functional. Despite Trump’s claims of total victory, the US has exhausted half its THAAD and Patriot interceptors. A Pentagon assessment warns mine-clearing in the strait could take six months - pushing oil toward $300 a barrel.
Jack Mallers on his namesake show argues Iran isn’t fighting a military war - it’s waging a financial one. By choking oil flows, Iran forces inflation into a US economy already drowning in debt. “They don’t need to win on the battlefield,” Mallers says. “They just need to break the Treasury market.”
The fallout is global. China’s exports slowed in March as supply chains seize up. TSMC and nickel processors are cutting output due to sulfur and helium shortages. Even rural China is feeling it - Western fast food chains are retreating from new store builds as costs spiral.
Jeffrey Sachs on The Tucker Carlson Show calls the conflict a continuation of the 1996 Clean Break doctrine - Israel’s roadmap to dismantle seven nations. Iran is the last target. “Netanyahu called this war his dream come true,” Sachs says. The US, he argues, is not acting in its own interest but as Israel’s enforcement arm.
Sources: The Tucker Carlson Show, Human Action Podcast, Macro Voices, Breaking Points with Krystal and Saagar, Breaking Points with Krystal and Saagar, Breaking Points with Krystal and Saagar, The Jack Mallers Show, The Intelligence from The Economist, No Agenda Show
AI beats Bitcoin in energy hierarchy
TL;DR
- Governments will shield AI data centers during blackouts, leaving Bitcoin miners first to lose power.
- AI’s thirst for water and electricity sparks rural revolts in Ohio, Wisconsin, and Virginia.
- As big miners pivot to AI, Bitcoin’s hash rate could fragment - helping decentralization.
The Story
AI is no longer just a tech story. It’s an energy war now - and Bitcoin is losing. Michael Dunworth on What Bitcoin Did argues that governments will soon treat AI infrastructure like water or power: essential. When the grid strains, regulators will cut power to Bitcoin miners before touching AI data centers running critical services.
This isn’t hypothetical. Forty of the top 200 ASX-listed firms have already shifted infrastructure toward AI. And in Franklin Furnace, Ohio, Google is building an 800-acre data center that guzzles up to a billion gallons of water and emits a constant, grating hum that drives neighbors indoors. Locals say they’re getting 50 permanent jobs in return - hardly a fair trade.
“AI is being classified as essential infrastructure. Bitcoin is not.”
- Michael Dunworth, What Bitcoin Did
The backlash is spreading. Voters in Wisconsin and Virginia have blocked similar projects. Senator Jon Ossoff is investigating how these facilities inflate household power bills. Ryan Grim notes this local resistance is the only real check on Big Tech’s land and water grabs.
Meanwhile, the AI surge is reshaping Bitcoin’s mining landscape. Firms like IREN are pivoting hard to AI, pulling capital from mining operations. Dunworth sees a silver lining: as industrial-scale miners retreat, the hash rate could fragment, weakening central control over the network.
But Bitcoin’s institutional adoption brings new risks. Companies like MicroStrategy are stacking sats through regulated custodians like Coinbase. That concentrates supply. A government in crisis wouldn’t need to hunt millions of wallets - just seize a few custodial vaults holding 30% of the coin.
“When machines need to transact, they can’t call a bank manager. They need Bitcoin.”
- Michael Dunworth, What Bitcoin Did
The irony is sharp. AI may starve Bitcoin of energy today - but tomorrow’s AI agents might run on Bitcoin anyway. Dunworth calls it a ‘harmonic loop’: energy becomes computation, computation becomes Bitcoin, Bitcoin buys more energy. The machines won’t trust humans. They’ll trust math.
Sources: What Bitcoin Did, Breaking Points with Krystal and Saagar
Central banks flee Treasuries for gold
TL;DR
- Central banks are buying gold, not Treasuries, to dodge dollar weaponization.
- The U.S. is now a net exporter of gold, signaling a reserve shift.
- A blocked Strait of Hormuz could trigger cascading supply collapses.
The Story
Central banks are dumping U.S. debt and buying physical gold in bulk, a quiet but decisive break from the dollar system. The move isn’t theoretical - it’s visible in trade flows. For four of the last five months, gold has been the top U.S. export, not planes or cars. Most of it is heading to China and the Middle East.
Luke Gromen on Human Action Podcast laid out the stakes: by sanctioning Russia and Iran, the U.S. proved it would weaponize the dollar. That pushed other nations to seek neutral assets. Gold, not Treasuries, is now the go-to reserve. “The dollar is a token used at a play center,” Gromen said. “The world wants real commodities.”
“The single biggest export from the U.S. has been gold, not planes or cars.”
- Luke Gromen, Human Action Podcast
This isn’t just financial reshuffling. The Strait of Hormuz remains a flashpoint, and any prolonged closure would halt oil, sulfuric acid, and helium flows - all critical to refining and chips. Restarting idle plants takes months. One failure now risks a chain reaction.
Meanwhile, the U.S. military faces a paradox: it relies on Chinese factories to build the very missiles meant to counter China. While Washington spent on wars, Beijing bought copper mines and built supply dominance. The U.S. can’t scale up without Chinese parts.
“Beijing controls the American military by supplying the parts for its missiles.”
- Luke Gromen, Human Action Podcast
The financial and physical supply systems are now intertwined. De-dollarization isn’t a future risk - it’s underway. Gold is no longer a relic. It’s the anchor in a breaking system.
Sources: Human Action Podcast
The Frontier — 50+ shows, one read.
Write a comment