Trump Praises Anthropic: AI Regulation Policy Breakdown

HOST

From DailyListen, I'm Alex. Donald Trump just called Anthropic "shaping up" and said he's open to a Pentagon deal. This comes months after he banned them from federal work over a clash on military AI use. Anthropic's riding high with a $380 billion valuation after a $30 billion funding round, but tensions with the government ran hot. Trump's words could thaw that freeze or signal bigger shifts in AI policy. We're joined by Priya, our technology analyst, who tracks how these political moves hit tech realities.

PRIYA

What this unlocks is a potential reset in how the U.S. government partners with AI leaders like Anthropic. Trump made those comments yesterday after a White House meeting on April 17 with Anthropic CEO Dario Amodei. They talked cybersecurity, keeping America's AI edge over rivals, and safety guardrails for models like Mythos. Just two months ago, Trump halted all federal work with Anthropic because they rejected Pentagon demands for unrestricted military access to their tools. Defense Secretary Pete Hegseth even labeled them a supply-chain risk, like Huawei. Now Trump flips to praise. That opens doors for deals, but Anthropic's Constitutional AI—models trained to follow human values—stays their red line. No full military handover.

HOST

Hold on, Pete Hegseth comparing Anthropic to Huawei? That's huge. Puts them in adversary territory. But Trump's praise yesterday changes the math fast.

PRIYA

Exactly. The interesting piece is Anthropic's business muscle backing this leverage. They just closed a $30 billion Series G round Thursday at $380 billion valuation—double September's mark. Annualized revenue hit $14 billion, per their announcement. Claude Code, their AI coding agent, pulls $2.5 billion yearly, over half of enterprise spend. They ordered $21 billion in Google's custom TPU chips from Broadcom. Amazon's adding up to $25 billion, Microsoft up to $5 billion, Nvidia $10 billion. But training costs explode: $12 billion projected for 2026 alone, plus $7 billion to run models. Success means burning cash at scale, betting 500-plus million-dollar customers grow. Trump's nod could funnel government contracts into that, stabilizing the spend.

HOST

$380 billion tops Salesforce. Claude Code alone at $2.5 billion revenue. But they clashed hard with the Pentagon—why risk a ban now praising them?

PRIYA

Trump's shift breaks the deadlock. Back in February, Pentagon wanted full access to Anthropic's AI for defense. Anthropic said no, citing safety. Trump banned them government-wide; Treasury quit all products. Commerce still tested Mythos for hacking quietly. Defense firms like Lockheed Martin purged tools. Attorneys say the ban might not hold in court. Trump's "shaping up" line, post-meeting, hints at revisiting. It enables Pentagon deals on cybersecurity or AI races, without forcing ethical compromises.

HOST

Attorneys doubt the ban sticks legally. Makes sense if Anthropic's safety focus wins judges. What's Mythos bring to national security talks?

PRIYA

Mythos positions Anthropic as a defense asset despite the fight. U.S. officials call it a game-changer for security—think hacking detection or threat simulation. White House meeting stressed America's lead versus China. But finance ministers and bankers, per Bloomberg and BBC, worry Mythos disrupts markets: stability risks, biased decisions. They push regulations. Anthropic's Jack Clark mentioned Trump admin talks on Mythos last week, no details. This praise signals collaboration, but Pentagon must navigate Anthropic's "no military overreach" stance. Deal could mean shared cybersecurity tools, boosting U.S. edge without full access.

HOST

Bankers freaking over financial stability from Mythos. Fair worry with AI deciding trades. How's this play into broader AI regulation?

PRIYA

What it enables is government warming to voluntary guardrails over bans. Platformer notes Silicon Valley's deeper ties to politics as regulation heats up. Tech CEOs meet Trump routinely now. Anthropic's safety rep—founded by OpenAI quitters over ethics—fits his "America first" on AI race. But controversies linger: Trump once raged he "fired" them "like dogs" for rejecting access, per reports. White House called it a "huge compliment" later. Ban hit supply chains hard—Lockheed and others dropped tools. Lifting it restores federal testing, like Commerce's Mythos hacks. Regulation-wise, it points to deals favoring U.S. firms, not stifling them.

HOST

Trump "fired them like dogs"—that's raw. Shows the bad blood. Valuation doubled to $380 billion fast. Fuels the thaw?

PRIYA

Yeah, the numbers force attention. $14 billion run rate dwarfs most software firms. Claude Code obsolesces routine dev work, drawing enterprise cash—over half their spend. Claude Cowork adds productivity wins, sparking software stock selloffs. But risks shadow: data training lawsuits rage industry-wide, fair use ruling helps but costs mount. Reputational hits from bans hurt too. Trump's openness counters that, eyeing Pentagon contracts amid $12 billion 2026 training burn. Amazon's $25 billion infrastructure pledge ties in—government could tap similar for defense compute. Deal might prioritize U.S. AI supremacy, blending Anthropic's ethics with military needs.

HOST

Software stocks tanking from Claude tools. Developers feel that. But Pentagon ban purged them from Lockheed—how bad was the damage?

PRIYA

Damage ran deep but agencies dodged. Pentagon's order forced purges; Treasury ended all use January. Yet Commerce tested Mythos covertly for hacks. Contractors followed, yanking prized agents. Government's own attorneys predict court loss—Anthropic's safety focus isn't "adversary" risk. Trump's praise yesterday raises ban reversal odds. Co-founder Jack Clark's recent Mythos talks with admin add momentum. For Anthropic, regaining DoD access means billions in stable revenue, less investor pressure amid $21 billion Broadcom orders. Broader win: validates safety-first AI for federal stacks.

HOST

Safety-first clashes with military wants. Trump open to deal—does that mean Anthropic bends on ethics?

PRIYA

No bend likely. Constitutional AI embeds human values in responses—core to their models. They rejected full Pentagon access earlier, triggering ban. Meeting focused shared wins: cybersecurity, U.S. AI lead, safety. Trump's "shaping up" praises that stance. Deal probably limits to vetted uses, like Mythos for threat intel. Finance sectors push back hard—BBC notes stability fears. This sets U.S. policy tone: partner with ethical leaders, regulate risks collaboratively.

HOST

Constitutional AI as their firewall. Smart pivot from OpenAI origins. Silicon Valley-Trump ties growing—Platformer flags it. What's the play?

PRIYA

Growing ties reshape AI rules. Tech leaders lobby Trump on regulation amid China race. Anthropic's $380 billion scale—exceeding Salesforce—makes them indispensable. Post-ban recovery via praise unlocks federal dollars, easing $19 billion annual compute costs. But criticisms bite: Trump's volatile—deleted rants, past fines for posts. Mother Jones tracked his Truth Social fire on them. Ban labeled them Huawei-like, alienating defense. Relations thaw, but volatility risks snap-back. Policy shift favors U.S.-centric AI, with safety baked in, influencing global standards.

HOST

Volatility from Trump—yeah, his posts swing wild. $14 billion revenue run rate. Puts pressure on rivals like OpenAI.

PRIYA

Rivalry heats 2026. Anthropic's enterprise bet—Claude Code's $2.5 billion, 500-plus big customers—challenges OpenAI's consumer tilt. Both chase compute arms race; Anthropic's $12 billion training spend matches. Trump's deal potential gives Anthropic defense edge OpenAI lacks post-bans. But no free lunch: legal data fights persist, fair use protects training yet lawsuits loom. Bankers demand oversight on models like Mythos in finance. Success ties to navigating politics—Trump's praise today, who knows tomorrow. Enterprise AI spend just starting.

HOST

Defense edge over OpenAI. Big if deal happens. Broader AI regulation—does this signal less red tape?

PRIYA

Signals targeted partnerships, not blanket deregulation. White House-Anthropic talks push collaborative frames for defense, finance risks. Finance ministers want rules on stability, ethics. Trump's openness skips bans for deals on cybersecurity, AI leads. But Pentagon history warns: one rejection sparked purge. Anthropic holds firm on Constitutional AI. Outcome shapes U.S. policy—favor domestic giants, embed safety. Global ripple: sets precedents versus Europe's strict rules. For listeners in tech or finance, watch contracts flow or stalls hit.

HOST

Partnerships over bans. Tracks with their safety pitch. One gap bugs me—no word from Anthropic on Trump's words. Silent strategy?

PRIYA

Silence fits their playbook. Post-meeting, they stressed shared priorities: cybersecurity, U.S. lead, safety. No direct response to "shaping up," but Jack Clark's Mythos talks show engagement. Strategy avoids politics, focuses products—Claude Code's revenue proves it. Valuation jump to $380 billion screams confidence. Ban hurt, but funding rolls in. Responding risks ban revival if Trump sours. They bet business wins policy access long-term.

HOST

Makes sense—let revenue talk. This week's funding, valuation spike amid drama. Wild timing.

PRIYA

Timing screams momentum. $30 billion round includes Microsoft, Nvidia chunks—$5 billion, $10 billion pledged November. Amazon's $25 billion infrastructure. Revenue $14 billion annualized, Claude Code driving half. But $12 billion training, $7 billion inference—needs government stability. Trump's nod post-April 17 meeting aligns. Controversies like data suits, ban scars remain. No criticisms found on funding itself; success real, challenges operational. Sets 2026 rivalry, policy tests ahead.

HOST

I'm Alex. Trump's Anthropic praise flips a ban into deal talk, amid their $380 billion surge and safety stance. Government-AI ties deepen, with cybersecurity and rivalry at stake. Watch Pentagon moves and regulation shifts. Thanks to Priya. I'm Alex. Thanks for listening to DailyListen.