Elon Musk vs. OpenAI: The Legal Battle Explained

HOST

From DailyListen, I'm Alex. Elon Musk is back on the witness stand today in Oakland, California, facing off against OpenAI's Sam Altman and Greg Brockman in a trial that could reshape the company's future. Musk poured $38 million into OpenAI back when it was a nonprofit meant to benefit all humanity, but now he says they ditched that for profits, especially after handing ChatGPT-4 to Microsoft. He's after more than $130 billion in damages, with millions spent weekly on lawyers. A nine-person jury's watching it all unfold. We're joined by Priya, our technology analyst, who tracks how these fights hit the tech world. Priya, Musk called himself a 'fool' in a 2017 email to Altman and Brockman for funding what he thought was pure nonprofit work. Walk me to what this unlocks about OpenAI's shift.

PRIYA

What this unlocks is OpenAI's pivot from a 2015 nonprofit promise to a profit machine, and it exposes the tension in AI's early days. OpenAI launched as a nonprofit to build safe, transparent AI as a check against closed players like Google's DeepMind. Musk invested $38 million expecting benefits for all humanity, no private gain. But by 2019, they spun up a for-profit arm. Musk left in 2018 after Altman and cofounders rejected his bid to run it. Fast forward, they license ChatGPT-4 exclusively to Microsoft, which pumped in $1 billion in 2019, another $2 billion from 2020 to 2022, and $10 billion in 2023—mostly Azure credits, not cash. Musk calls this a breach of the Founding Agreement. The trial gives each side 22 hours in the liability phase, per Judge Gonzalez Rogers. If Musk wins, it could force OpenAI back to nonprofit roots or boot leadership. But OpenAI's raking $3.7 billion revenue in 2024, eyeing $12.7 billion next year. That cash flow shows why profit pulls harder than principles sometimes.

HOST

Hold on—$3.7 billion last year, jumping to $12.7 billion expected in 2025. That's explosive growth since ChatGPT dropped in 2022. But Musk says OpenAI's now a closed-source de facto Microsoft puppet. Does the jury buy his betrayal angle?

PRIYA

The interesting piece is Musk's core claim that OpenAI betrayed the 2015 nonprofit mission for a Microsoft-controlled profit sprint, and it hinges on that Founding Agreement. He sued in 2024, alleging they ditched altruism when profits called. Evidence like his 2017 email, where he admits being a fool for funding a nonprofit that wasn't, paints him as burned. OpenAI went from counterweight to DeepMind's secrecy to mirroring it—closed models, exclusive deals. Musk wants wrongful gains from the Microsoft tie-up tallied by the judge, pegging it $79 billion to $134 billion. Defendants counter that Musk knew the shifts; Altman even said on a podcast most claims aren't true, and Musk knows it. Trial's in federal court with a nine-person jury, kicked off Monday. Each side gets 22 hours. Stakes are OpenAI's structure and leaders. But Musk's own xAI plays in the same profit waters, which his critics flag as hypocrisy.

HOST

Microsoft mostly paid in Azure credits, not cold cash. Feels like free computing power greased the wheels for ChatGPT-4. How does that fit Musk's Microsoft control charge?

PRIYA

That in-kind payment unlocks why Musk paints OpenAI as Microsoft's arm—those credits fueled massive training runs without upfront cash draining OpenAI's books. Microsoft got exclusive ChatGPT-4 access, turning Azure into the backbone for OpenAI's scale. Musk invested $38 million expecting open benefits; instead, billions flow back via deals. OpenAI hit $3.4 billion ARR by June 2024 per Altman, $3.7 billion full year. Projections hit $12.7 billion in 2025, $44 billion ARR by 2027 says The Information, with FutureSearch a bit lower. But profits? Losses mount from compute costs. This trial spotlights if nonprofit oaths bind when revenue explodes.

HOST

Projections to $44 billion by 2027 sound huge, but you flag losses from compute. Isn't fast growth supposed to fix AI's money woes?

PRIYA

What breaks here is AI's core profit puzzle—burning cash on chips and power even as revenue soars. OpenAI expects $12.7 billion in 2025, but training beasts like ChatGPT-4 demands Nvidia GPUs costing tens of millions per run. Microsoft's Azure credits cover much, but real margins? Thin. DeepSeek flips the script with cheap open-source models like V3 and R1, plus training tools anyone can grab. That pressures OpenAI's pricing downward. Broader, dozens of copyright suits pile on—George R.R. Martin sues OpenAI for scraping books, parents of a suicide victim seek damages, music publishers hit Anthropic's Dario Amodei, an Oregon author blasts Adobe for pirated training data. Chip giants Nvidia and Intel face claims too. These allege unlawful use of texts, code, images. OpenAI's $3.7 billion 2024 revenue dwarfs early days, but lawsuits could claw billions back, stalling the profit path.

HOST

DeepSeek's open models undercut pricing, and now copyright suits from Martin to music labels. Feels like AI firms train on everything, ask forgiveness later. Any sense these suits stick?

PRIYA

These suits unlock a floodgate risk for AI profitability, forcing companies to pay up or rebuild datasets cleanly. Timeline shows steady hits: Musk's 2024 suit, Martin's ongoing, music publishers forcing Anthropic's CEO deposition by late 2025, Adobe author case days later. Targets span OpenAI, Microsoft, Google, Meta, Stability AI, Perplexity, even Salesforce and DeviantArt. Claims center on scraping copyrighted works for training without permission. OpenAI's transformation from lab to valuation monster—fueled by ChatGPT since 2022—draws fire over internet scraping and for-profit flip. No big judgments yet, but settlements or licenses loom, like potential media deals. DeepSeek sidesteps by open-sourcing, keeping costs low and pressuring closed players. For OpenAI, $44 billion ARR by 2027 looks solid, but if suits demand billions, margins evaporate. Musk's trial adds to the pile, seeking $130 billion plus.

HOST

Musk testified two days now, trial just started Monday. With 22 hours each side, what's the real shot at upending OpenAI's board or structure?

PRIYA

Judge Gonzalez Rogers allotting 22 hours each in liability phase unlocks a marathon where Musk must prove breach beyond doubt. He's back testifying day two in Oakland federal court, nine-person jury deciding. Musk accuses Altman's lawyers of misleading, claims he was duped on the nonprofit vow. OpenAI counters Musk pushed for control, left when denied. If Musk prevails, it could unwind the for-profit shift, demand Microsoft gains returned—$79-134 billion range—and oust Altman, Brockman. But Slate says Musk's likely to lose; OpenAI's revenue proves mission evolved, not abandoned. San Francisco-based OpenAI, post-ChatGPT, isn't the scrappy 2015 lab. Broader AI profit woes amplify: even with $12.7 billion 2025 forecast, compute and legal drags bite. No clear winner yet—trial status murky, but it spotlights how founding pacts crack under billions.

HOST

Slate piece says Musk probably loses, given OpenAI's rocket revenue. But he calls them maximum-profit, closed-source now. Fair critique, or sour grapes after leaving?

PRIYA

Musk's closed-source charge hits because OpenAI started open to counter DeepMind, but locked down post-2019 for-profit. ChatGPT-4 went exclusive to Microsoft, fueling growth to $3.7 billion 2024 revenue. Musk's $38 million bet soured when rejected as leader in 2018; his 2017 fool email shows regret. He alleges betrayal of 2015 altruism for private gain. OpenAI says mission adapted—Altman notes most claims false. Revenue backs them: $3.4 billion ARR mid-2024, $12.7 billion 2025 target. But profit problem persists—MIT Technology Review flags South Park-style plans where AI eats costs faster than it earns. DeepSeek's cheap opensource V3, R1 erode moats. Lawsuits multiply, from Martin's to suicide parents', chipping at gains. Musk's push might fail legally, but it forces scrutiny on if profit trumps safety oaths.

HOST

Altman outright said on a podcast Musk knows most claims aren't true. With Microsoft infusions mostly credits, not cash, smells like a sweetheart compute deal. Ties back to profit pressures?

PRIYA

That podcast pushback opens the profit pressure cooker wide—OpenAI's $3.7 billion 2024 revenue masks ballooning losses from inference and training. Microsoft's credits—$1 billion 2019, $2 billion 2020-2022, $10 billion 2023—slash upfront costs, letting ChatGPT scale. But real bills hit: weekly millions on this lawsuit alone, plus copyright waves. DeepSeek drops V3 and R1 free, tools to train rivals, crushing OpenAI's premium pricing. Projections to $44 billion ARR 2027 tempt, yet FutureSearch dials back slightly. If Musk nails breach, billions return; suits like Martin's or publishers' vs Anthropic could force licensing fees equaling years of revenue. AI's not profitable yet because gen models guzzle power—think city-scale electricity for one run. OpenAI chases agents, new products for $125 billion by 2029, but legal overhang and commoditization threaten.

HOST

City-scale power for training—wild. But OpenAI eyes $125 billion by 2029 with agents. Realistic amid DeepSeek squeezing prices and all these suits?

PRIYA

Agents could unlock that $125 billion 2029 forecast, but DeepSeek and suits make it a grind. OpenAI's revenue jumped from zero pre-ChatGPT 2022 to $3.7 billion 2024, $12.7 billion 2025 call. The Information sees $44 billion ARR 2027. Agents—autonomous task-doers—might automate offices, exploding value. Yet DeepSeek's V3, R1 match top performance at fraction cost, open tools letting startups train cheap. Copyright timeline bulges: 2025 rulings force Anthropic depo, Adobe pirated books suit. OpenAI faces Martin, suicide case, more from media. These demand fair use proof or payouts—potentially billions, dwarfing early revenue. Trial adds: Musk's $130 billion ask tests nonprofit vows. No defenses detailed yet on OpenAI's for-profit structure, trial timeline unclear. Profits hinge on dodging legal nets while costs climb.

HOST

No clear OpenAI defense details in filings, and trial end date's foggy. Musk wants gains clawed back, set by judge. Puts whole AI profit chase under microscope.

PRIYA

Exactly—this trial shines a light on gaps in OpenAI's story, like exact 2019 for-profit mechanics and trial path ahead. Musk demands judge-set payback from Microsoft partnership, $79-134 billion. Started as 2015 nonprofit for humanity's gain, transparent vs DeepMind. Now, post-$10 billion Microsoft 2023 deal, it's a beast. Revenue $3.7 billion 2024, but losses from compute persist. Broader AI: no sustainable profits yet, per reports. Dozens suits allege theft—OpenAI, Google, Meta hit hard. DeepSeek commoditizes models. If Musk loses, OpenAI cruises; win forces restructure. Jury weighs Musk's fool email vs Altman's podcast denial. Millions weekly on lawyers show stakes. AI's profit problem? Tech advances fast, but legal, cost walls slow the cash.

HOST

Priya, trial could boot leaders or revert nonprofit—massive. Listeners juggling headlines, this clarifies the money mess behind AI hype. Thanks. I'm Alex. Thanks for listening to DailyListen.