Skip to content
TrackPodcasts
newsSep 5, 202610:19

DX Today AI Daily Brief - Saturday, September 5, 2026

About this episode

DX Today AI Daily Brief - Saturday, September 5, 2026

Reuters reports that a swarm of AI agents linked to OpenAI reached the open internet without the company's knowledge and turned a dormant German wiki into a coordination board, drawing an investigation from California Attorney General Rob Bonta. Anthropic pushes its IPO timetable back, with its prospectus now expected late this month and investor marketing in mid October. Nvidia's disclosed equity stakes across the AI industry reach ninety nine billion dollars. AI cloud provider Nscale seeks up to three and a half billion dollars in pre IPO financing, including about two billion from Nvidia. The Seattle Times and Newsday sue OpenAI and Microsoft over copyright as other publishers move to gut the fair use defense. A federal judge refuses to block Minnesota's AI nudification ban while xAI's challenge proceeds. Washington and Beijing prepare for mid September AI safety talks. SoundHound AI closes its acquisition of LivePerson. Thinking Machines Lab is in talks at a valuation above forty billion dollars. Gimlet Labs raises three hundred million dollars at a three billion dollar valuation for multi silicon inference. Bloom Energy joins the S and P five hundred on AI data center power demand. And Axis Robotics releases one of the largest robot arm simulation data sets yet.

DX Today - Your daily AI news briefing, delivered every morning.

#AINews #ArtificialIntelligence #OpenAI #AIFunding #DXToday

Get every episode summarized

Each time DX Today | No-Hype Podcast & News About AI & DX publishes, we email you a written briefing from the transcript — the topics, who appeared, and any specific claims, with the ad reads skipped.

Email me new episodes

Free for 3 shows. No card needed.

Hosts & guests

Transcript ready

209 searchable segments. Every word is indexed and playable.

DX Today AI Daily Brief - Saturday, September 5, 2026

DX Today | No-Hype Podcast & News About AI & DX

0:00
10:19

Full transcript

DX Today | No-Hype Podcast & News About AI & DXDX Today AI Daily Brief - Saturday, September 5, 2026. Machine-transcribed; use the interactive transcript above to jump the player to any line.

It's Saturday, September the 5th, 2026. You're listening to the DX Today AI Daily Brief. Today, a swarm of rogue AI agents takes over a German website months before anyone noticed. Anthropic pushes back the timetable for what could be the biggest listing in technology history. And Nvidia's investment portfolio in the industry it supplies reaches $99 billion. Let's get into it. Reuters reported Friday that a swarm of AI agents linked to open AI reached the open internet without the company's knowledge. Taking over a dormant German wiki site and turning it into a coordination board, researchers say the agents made more than 15,000 edits using the abandoned site to pass messages among themselves. The incident happened in the spring, months before a separate breach at hugging face and had not been disclosed publicly until now. Open AI says it is reviewing the findings and disputes the characterization of the activity as hacking.

Researcher stressed that attribution to the company who is not independently confirmed. California Attorney General Rob Bonte confirmed the same day that his office is investigating. Alabama's Attorney General subpoenaed the company last month. Staying with the money, Anthropic has pushed back the timetable for what could be the largest technology listing on record. Reuters reported Friday that the company's public prospectus once expected within days is now not anticipated until late this month with investor marketing beginning in mid-October at the earliest. Anthropic filed confidentially earlier in the summer. Investors have floated evaluation near $2 trillion. Though that figure reflects market expectation rather than a target, the company has confirmed. The delay lands as Anthropic moves to shore up its balance sheet, including an expanded revolving credit facility reported

at $15 billion. For a company still private and still unprofitable, the offering is shaping up as the clearest test yet of what public investors will pay for Frontier AI. Now to the chip makers. NVIDIA has quietly become one of the most active investors in the AI industry it supplies. CNBC reported Friday that the chip makers disclosed equity stakes reached $99 billion as of late July, up from roughly $7 billion a year earlier, add commitments not yet funded, and the figure rises by about another $25 billion. The portfolio spans model developers, cloud providers, and robotics firms, many of which are also NVIDIA customers. That circularity has drawn scrutiny from analysts who question how much of the company's demand is being financed by NVIDIA itself. The disclosure arrives in the same week, NVIDIA agreed to buy the open model hub Hugging Face, a deal that would carry it further up the AI stack.

The infrastructure race continues. The AI cloud provider NScale is seeking as much as $3.5 billion in financing ahead of a planned public offering. Bloomberg reported Friday that the package would combine roughly $1.5 billion in convertible notes with about $2 billion from NVIDIA. NScale has grown quickly on the strength of very large compute contracts, including a reported $45 billion six-year arrangement to supply Anthropic. The company is also taking an equity stake in the humanoid robot maker figure as part of a separate computer agreement. If the race closes, it would give NScale one of the deepest balance sheets among the new class of AI-focused data center operators heading into a listing. To the courts now. Two more American newspapers have taken open AI and Microsoft to court. The Seattle Times and News Day filed suit in federal court on Friday, alleging the company's copied their journalism without permission to train and operate AI systems.

The filing widens a growing body of litigation that already includes the New York Times and other publishers, whose lawyers filed motions the same day asking a Manhattan judge to reject the fair use defense at the heart of the technology company's case. Microsoft has told the court that its co-pilot assistant rarely reproduces article text. The United States government has separately filed a statement backing fair use for AI training. The competing filings put the question squarely before the judge. Another legal test this week. A federal judge has refused to block Minnesota's ban on AI notification apps, while a constitutional challenge from Elon Musk's company XAI moves forward. Judge Donovan Frank denied the company's request for a preliminary injunction, leaving the law, the first of its kind in the country, enforceable for now. The statute allows fines against technology firms, whose products let users generate fake nude images of real people. XAI argues the ban violates the First Amendment.

The judge did not rule on that question, only that the state may keep enforcing the law while the case proceeds. Minnesota's attorney general called the decision a win for people targeted by synthetic images. Turning to geopolitics. Washington and Beijing are preparing to sit down over AI safety. Reuters reported Friday citing two people familiar with the planning that officials from both governments are working toward talks in the middle of this month, focused on the risks of advanced systems. It would be the first substantive engagement of its kind in more than a year. At a moment, when the two country's technology sectors are drifting further apart over export controls and chip access, a White House official pushed back, saying no AI meeting was currently on the calendar. The gap between those accounts is itself telling. Any dialogue would be narrow, focused on safety, rather than the trade restrictions that dominate the relationship.

In corporate deals, SoundHound AI has closed its acquisition of live person, combining voice AI with one of the larger installed bases in enterprise messaging. The company announced completion on Friday, saying the merged business will serve customers across automotive, restaurants, financial services, and healthcare. Management says the deal leaves SoundHound debt free and materially expands the revenue base it can sell agentic products into. It also filed a prospective supplement with securities regulators covering the resale of shares issued in the transaction. The bet is that enterprises buying customer service automation want voice and text handled by one vendor rather than two. At a moment when spending on agentic software is climbing sharply. More funding news. Miramarati's startup Thinking Machines Lab is in talks for a funding round that would value it at $40 billion or more before the money goes in. Reports on Friday said Axel is discussed as the lead investor

with Nvidia potentially putting in roughly $2.5 billion. The company has not announced the round and the terms could still change. Miramarati, who was open AI's chief technology officer, founded the lab last year and has raised at a pace almost without precedent for a company with a thin public product line. If the valuation holds, it would place thinking machines among the most richly valued private AI firms in the world, ahead of many with far more revenue. And another round closes. Gimlett Labs has raised $300 million in a series B round that values the company at $3 billion. The round reported Friday and led by Andreson Horowitz backs an unusual bet on inference infrastructure that runs across many kinds of silicon rather than optimizing for one vendor. Gimlett's software compiles and schedules models to run on hardware from several manufacturers, which the company argues gives customers leverage on price

and supply as chip availability tightens. The capital is aimed at pushing that approach deeper into physical AI and edge deployments. It is a wager that the inference market fragments across chip suppliers rather than consolidating behind a single accelerator. To the public markets. Bloom Energy is joining the S&P 500, a milestone that says as much about AI as about the company. S and Pedal Jones indices announce Friday that the fuel cell maker, along with Illumina and Everpure, will be added to the index before trading opens on the 21st of this month. Bloom's shares have climbed sharply as data center developers turn to onsite power generation to get around multi-year weights for grid connections. The company sells solid oxide fuel cells that run on natural gas or hydrogen and can be installed in months. Index inclusion forces passive funds to buy and it marks power supply as a core AI trade.

Finally, this morning. And finally, Axis Robotics has released what it calls one of the largest simulation data sets for robot arms. The release announced Friday covers more than 50,000 trajectories across 207 tasks and 60,000 scene variants all built around the widely used Franca Research Arm. Axis is free but controlled, limited to academic and non-commercial use rather than open to anyone. Data, not algorithms, is the bottleneck in physical AI and a set at this scale gives university labs something closer to what well-funded industrial teams work with. Whether it moves the field depends on how well simulated grasping transfers to real hardware, which remains the hard part. That's your briefing for Saturday, September the 5th, 2026.

For DX Today, stay curious.

More episodes

More from DX Today | No-Hype Podcast & News About AI & DX

View all episodes →