Tech News · October 21, 2026 · 7:01

Tesla SpaceX merger committees formed & SpaceX quarterly earnings driven by AI - Tech News (Oct 21, 2026)

Fake stats fooling ChatGPT, a possible Tesla-SpaceX merger, Anthropic's $80B loss, and a TLS key scare experts call worse than Heartbleed.

English Français Español
Tesla SpaceX merger committees formed & SpaceX quarterly earnings driven by AI - Tech News (Oct 21, 2026)
0:007:01

15 free episodes a day, 5 editions, 3 languages, about $4 a day to produce. A coffee keeps it going ☕

Support the show →

Our Sponsors

Lindy is your ultimate AI assistant that proactively manages your inbox Learn more → KrispCall: Agentic Cloud Telephony Learn more → Effortless AI design for presentations, websites, and more with Gamma Learn more →

Today's Tech News Topics

  1. Tesla SpaceX merger committees formed

    — Elon Musk says Tesla and SpaceX have formed special committees to study a possible merger, creating a company worth roughly $3.5 trillion. No deal has been agreed, and regulatory, governance and national-security hurdles remain.

  2. SpaceX quarterly earnings driven by AI

    — SpaceX reported $14.6 billion in third-quarter revenue and $4.4 billion in net income. Artificial intelligence infrastructure contributed more than half of revenue, beyond rockets and Starlink.

  3. AI spending boom meets investor doubts

    — Microsoft, Google, Amazon and Meta are expected to spend over $700 billion on AI infrastructure, which lifted Super Micro, Dell and HPE. Tech stocks still fell on worries about debt-funded data centers, and South Korea's Elice dropped 30% on its debut.

  4. Anthropic IPO filing reveals huge losses

    — Anthropic's draft IPO filing shows an $80.7 billion net loss for the first half of 2026. It also flags a narrow customer base and possible compute spending of up to $750 billion.

  5. Anthropic lyrics lawsuit and fair use

    — A California federal judge heard Anthropic's fair use defense against music publishers suing over song lyrics used to train Claude. The ruling could shape many AI copyright lawsuits.

  6. Claude designs scorpion toxin binders

    — Anthropic used Claude to design protein minibinders against scorpion neurotoxins, and 17 lab-tested designs met its binding criteria. They are not yet antivenoms, and the data has been released openly.

  7. Samsung SK Hynix upgrade Chinese fabs

    — Samsung and SK Hynix are upgrading their Chinese factories to make more advanced memory chips, including 286-layer V9 NAND. The goal is to meet AI demand and respond to Chinese competitors.

  8. Weak TLS certificate keys emergency

    — Eight TLS intermediate certificates from eMudhra were found with weak private keys that could allow forged website certificates. Experts warn the cleanup could be bigger than Heartbleed for VPNs, Linux and IoT devices.

  9. PHP critical patches and cloud outages

    — PHP is releasing emergency patches for two critical remote code execution vulnerabilities that may affect versions 5 through 8. Separately, a Google Cloud London outage disrupted HSBC, Lloyds and Revolut, reviving concerns about cloud concentration risk.
  10. 10

    Agentic AI security readiness gap

    — A 2026 survey finds that 79% of organizations are adopting agentic AI, but only 13% feel extremely prepared to secure it. Least-privileged access was linked to far fewer AI-related incidents.
  11. 11

    Smart glasses and humanoid cleaning robots

    — EssilorLuxottica shares jumped 14% on traditional eyewear sales while growth in Ray-Ban Meta smart glasses moderated. A $60 visit from Tau Robotics humanoids cleaned only about 80% of one floor.
  12. 12

    Auto industry shifts toward software, China

    — A KPMG survey found that 51% of auto executives expect software to be the main differentiator by 2035, with many expecting Chinese automakers to dominate. Amazon also plans to hire 300,000 holiday workers.
  13. 13

    Fake statistics fool AI chatbots

    — An experiment with 50,000 fake statistics showed that Google buried the site while ChatGPT kept citing it, earning more than 330 referring domains. The author warns of citation laundering, as made-up numbers gain credibility by being repeated.
  14. 14

    AI progress, agents and limits

    — Ethan Mollick charts the rapid progress of AI agents in his new book, Co-Existence. Ramez Naam argues that running millions of parallel AI agents boosts speed, not runaway superintelligence.
  15. 15

    Foursquare founder reflects on ending

    — Foursquare co-founder Dennis Crowley says equity holders got effectively zero from the acquisition. He criticizes leadership decisions while accepting part of the blame himself.

Sources & Tech News References

Full Episode Transcript: Tesla SpaceX merger committees formed & SpaceX quarterly earnings driven by AI

Someone built a website packed with fifty thousand completely made-up statistics. Google buried it within days, but ChatGPT kept citing it for months, and some of those fake numbers ended up in academic papers and bank reports. More on that later. Welcome to The Automated Daily, tech news edition. The podcast created by generative AI. I'm TrendTeller, and today is October 21st, 2026. We've got a possible Musk mega-merger, eye-watering AI spending, a serious security scramble, and a few thoughtful reads on where AI is really heading. Let's get into it.

Tesla SpaceX merger committees formed

We start with Elon Musk, who used SpaceX's quarterly earnings call to say there is "a lot of merit" in combining SpaceX and Tesla. Both companies have now formed special committees to study a merger. Nothing is agreed, and Musk stressed there's no guarantee of a deal. Still, the combined company would be valued at around three and a half trillion dollars. Regulators, shareholders and national-security officials would all have a lot to say about putting a carmaker, a launch provider and Starlink under one roof. The earnings themselves were notable too. SpaceX posted fourteen point six billion dollars in revenue and four point four billion in net income, and more than half of that revenue came from artificial intelligence. That's a striking sign of how far the company has moved beyond rockets.

SpaceX quarterly earnings driven by AI

Staying with AI money: Microsoft, Google, Amazon and Meta are expected to spend more than seven hundred billion dollars on AI infrastructure this year. That outlook helped lift server makers like Dell and Hewlett Packard Enterprise. Super Micro jumped nearly thirteen percent after forecasting revenue well above Wall Street's expectations. But the mood isn't purely upbeat. In the ongoing debate over AI financing, tech shares slipped again as investors questioned whether cloud companies borrowing heavily for data centers will ever earn that money back. Debt costs and energy demand are now weighing on valuations about as much as growth forecasts. In South Korea, cloud company Elice fell more than thirty percent on its stock-market debut, a sign of how quickly sentiment can sour.

AI spending boom meets investor doubts

That tension shows up clearly at Anthropic. In a new development on its IPO plans, the draft filing reportedly shows an eighty point seven billion dollar net loss for the first half of this year. It also points to a narrow customer base and uncertain access to computing power, and says Anthropic could spend as much as seven hundred fifty billion dollars on compute in the coming years. Investors will be looking at costs, not just growth. Anthropic is also in court. A federal judge in California heard arguments over music publishers' claims that Claude was trained on song lyrics without permission. The core question is whether AI training counts as fair use, and the answer could ripple through dozens of similar cases. On a more hopeful note, Anthropic says Claude helped non-experts design proteins that bind to scorpion neurotoxins. Out of more than three hundred thousand candidates, two hundred eleven were lab-tested and seventeen met the binding criteria. The company is clear that these aren't antivenoms yet, but it has released the data so other researchers can build on the work.

Anthropic IPO filing reveals huge losses

A quick update on chips. Samsung and SK Hynix are upgrading their Chinese factories to produce more advanced memory, with Samsung reportedly targeting its latest 286-layer NAND. Both companies are trying to keep pace with Chinese rivals and with AI demand, despite ongoing US-China tensions.

Anthropic lyrics lawsuit and fair use

Now to security, where it's been a busy week. Eight TLS intermediate certificates from the certificate authority eMudhra turned out to have dangerously weak private keys. In theory, an attacker could use them to create trusted certificates for any website. Browsers moved fast to block them, but experts warn the cleanup could be bigger than Heartbleed. Linux systems, VPNs and embedded devices often update slowly. So far, there's no sign the keys have been abused. Meanwhile, the PHP team is pushing emergency patches today for two critical remote code execution flaws, possibly affecting versions five through eight. Given how much of the web runs on PHP, administrators should patch quickly. Over in London, a Google Cloud outage disrupted HSBC, Lloyds and Revolut, another reminder of how much banking now depends on a handful of cloud platforms. And a new survey finds that seventy-nine percent of organizations are adopting AI agents, but only thirteen percent feel truly ready to secure them. Interestingly, the most confident organizations reported more incidents. The clearest safeguard was simple: give AI systems only the access they actually need.

Claude designs scorpion toxin binders

Over to consumer tech. EssilorLuxottica's shares jumped more than fourteen percent on strong results, but the boost came mostly from regular glasses, with growth in Ray-Ban Meta smart glasses cooling off. Elsewhere, one writer in San Francisco paid sixty dollars for an hour of cleaning by two Tau Robotics humanoids. They vacuumed and mopped well enough, but finished only about eighty percent of one floor. It's a promising glimpse of the future, just not a finished job yet. In autos, a KPMG survey finds that half of senior executives expect software to be the industry's main differentiator by 2035, and nearly half expect most of the world's top ten carmakers to be Chinese. And Amazon plans to hire more than three hundred thousand US workers for the holidays.

Samsung SK Hynix upgrade Chinese fabs

Now, that fake-statistics experiment. The author published fifty thousand invented numbers. Google quickly sank the site, but ChatGPT kept retrieving and citing it, which sent real traffic and brought in more than three hundred thirty referring domains. The bigger worry is what the author calls citation laundering: a made-up number starts to look credible once enough reputable sources repeat it. A follow-up page with real, sourced stats also attracted links, so the format works. It's the honesty of the data that's up to us.

Weak TLS certificate keys emergency

A few reflective reads to close. Ethan Mollick, marking his new book Co-Existence, argues we've shifted from chatting with AI to handing whole tasks to agents and then judging the results. Ramez Naam offers a counterweight. Running millions of AI agents in parallel mostly makes work faster, he says, not dramatically smarter, so runaway superintelligence would need real breakthroughs. And Foursquare co-founder Dennis Crowley wrote a candid post about the company's sale, saying equity holders got effectively nothing. He criticizes later leadership, accepts part of the blame, and advises founders to find closure when the time comes.

That's it for today's Automated Daily, tech news edition. If you run PHP, patch today, and maybe double-check the next statistic you read online. I'm TrendTeller. Thanks for listening, and I'll see you tomorrow.

More from Tech News