Leopold Aschenbrenner’s AI-centric hedge fund sells its public holdings after a market rout, a warning flare for the increasingly leveraged AI industry. TBPN Host John Coogan talks tech media and life after OpenAI ownership on the podcast. SPVs outpace direct share listings as secondary market mechanisms. A Thinking Machines Lab co-founder steps down citing health issues, but quickly takes a job at OpenAI. The Trump administration bans Chinese robots. Google DeepMind’s AlphaFold team sees a personnel shakeup and a few notable departures. Live-shopping app Whatnot offers a sign of life for consumer dealmaking.

The Main Item

As Markets Wobble, Circular Financing & Leveraged Bets Ramp up the Risks

The rapid retreat of AI and chip stocks over the past couple of weeks has exposed the fragility of an AI funding boom that’s drawing capital from every corner of the financial system — and is increasingly reliant on debt.

Situational Awareness, the much-touted San Francisco hedge fund led by 25-year-old wunderkind Leopold Aschenbrenner, this week provided an object lesson in the risks. Less than two months ago, the former OpenAI researcher was being celebrated by the Wall Street Journal for his prescience in seeing the AI opportunity early and turning a few hundred million dollars into well over $20 billion in less than two years. (The firm’s net asset value reportedly reached $45 billion at its height.)

On Thursday, he was forced to sell the fund’s public stock portfolio to Citadel after margin calls had left it in a liquidity crisis.

The accelerationist-leaning pseudo-anonymous account @beffjezos called the incident a “huge hit for EA-adjacent rationalist former big lab alignment researcher kind” and many meme-d Citadel’s Ken Griffin into the grim reaper of speculative investing.

It’s not that Aschenbrenner’s bets were wrong, exactly: the infrastructure and chip plays at the heart of the portfolio, including names like SanDisk, SK Hynix and CoreWeave, are still trading far above where they were a year ago. He himself has always spoken of the AI build-out as a long-term opportunity.

But “conviction” in this case meant levering up the wagers with borrowed money, and thus declines on the order of 30% on core holdings created a sudden cash crisis. It’s a reminder of how quickly fortunes can turn in an AI trade where valuations — and thus expectations — remain sky high, and risk is often stacked upon risk.

A similar red flag in that regard comes in the form of the latest wave of multi-billion-dollar Nvidia investment deals. This week, it was a $5 billion commitment to Ilya Sutskever’s Safe Superintelligence, much of which will be used to buy Nvidia gear.

Nvidia is reported to be discussing a loan guarantee for OpenAI of as much as $250 billion, also tied to Nvidia hardware.

That’s on top of $90 billion in corporate VC investments over the last 16 months alone, per the Financial Times. Nvidia participated in 283 funding rounds between 2021 and 2025, with 85% of those investments in AI startups, according to Crunchbase.

The Bank of Jensen

In many cases, these investments involve Nvidia essentially providing the money for customers to buy its hardware. Indeed, CEO Jensen Huang has been so aggressive in funding the AI ecosystem that the company is functioning almost like a bank; with annual free cash flow nearing $100 billion and little debt, that isn’t obviously reckless.

Yet such vendor financing, though it has plenty of precedents, is what brought down Lucent and much of the rest of the telecom equipment industry in the early 2000s after the first dot-com bubble burst.

Nvidia’s defenders point out that many of its deals are collateralized by the hardware, so if the company fails Nvidia gets the gear back. In a bust scenario, though, those chips would be worth a fraction of their current value.

Huang has proven himself a tech CEO for the ages, and an AI visionary. He’s also been exceptionally good at turning that into money. But banking is a different business. And as Aschenbrenner just demonstrated, seeing around corners doesn’t always guarantee success.

Nvidia’s frenetic deal-making is just one of the ways in which the AI buildout has upended the traditional start-up financing playbook. As we noted last week, corporate venture investing has exploded in recent years, and accounted for almost 90% of all VC dollars that have gone into AI firms this year. That’s compared with less than 50% of VC dollars a decade ago.

New types of financial players are also making this cycle different.

From Wall Street to Jane Street

While hedge funds such as Coatue and mutual fund companies including Fidelity started getting involved in late-stage venture deals more than a decade ago, Situational Awareness was part of a fresh wave of investment firms encroaching on traditional VC territory.

The secretive Wall Street firm Jane Street, which traditionally made its money from proprietary trading algorithms, is on the cap tables of Anthropic, chipmakers Etched and MatX, and AI cloud infrastructure heavyweight Fluidstack. (It also had money with Aschenbrenner.)

Jane Street is interested in AI tech not just as investment, but for internal use: it invested $1 billion in CoreWeave and signed a $6 billion commitment to use its cloud platform for all of its machine learning needs. It’s also in the process of financing its own AI data center.

Big money managers and private credit firms including Goldman Sachs Asset Management, Blackstone, and Blue Owl Capital have also emerged as major funders of the AI buildout, extending loan packages in the tens of billions for data centers.

Part of this is simply a function of the enormous amounts of capital needed for AI data centers and chips. But for strategic investors, there are many other reasons to play the venture game.

Plumping up potential customers is one obvious benefit. An investment could also yield access to premier AI tools ahead of the competition, and help keep a funder abreast of the latest developments. There’s also the chance of a windfall: Google recorded an incredible $99 billion gain on its equity investments in Q2, mainly from gains on Anthropic and SpaceX.

“Enterprise software companies in particular often feel real urgency to invest in and sometimes acquire AI startups to stay competitive and defend their market share,” PitchBook senior research analyst Kaidi Gao told us.

Fear and greed continue to draw new corporate players into AI startup investing. Schneider Electric, for example, a 190-year old energy and industrials firm, is channeling capital from its €1 billion venture fund, SE Ventures, almost entirely into AI startups.

The $700 million Series A for Figure founder Brett Adcock’s new AI neolab Hark shows the new state of play: investors include Nvidia, AMD Ventures, ARK Invest, Brookfield, Intel Capital, Qualcomm Ventures, and Salesforce Ventures.

Newcomer Podcast

John Coogan on Selling TBPN to OpenAI, Sam Altman & the Future of Tech Media

One Big Chart

SPVs Outpace Direct Sales for Secondary Share Sales So Far this Year

Investors are increasingly turning to Special Purpose Vehicles, or SPVs, to purchase shares of Silicon Valley’s most sought-after private startups, rather than buying shares directly from employees or other investors.

A new PitchBook report showed that single-layer SPVs have surpassed direct-to-cap-table trades in secondary deal volume on the trading platform Caplight so far this year.

Read more

A lot can happen in three years.

When an Amodei first spoke at the Cerebral Valley AI Summit back in 2023, Anthropic was a research lab valued at about $5 billion and generative AI was still a niche San Francisco obsession rather than the engine of the global economy.

Even when Anthropic CEO Dario Amodei followed his sister, Daniela, and spoke at Cerebral Valley in 2024, the company wasn’t the nearly trillion-dollar juggernaut that it is today.

Databricks’ Ali Ghodsi famously bought a unicorn startup out of our first event and returned for a victory lap a few months later. In the years since, he’s abandoned his in-house model ambitions and is increasingly in the business of commoditizing model companies.

We’ve never had Nikesh Arora grace our stage before. But with security increasingly at the heart of the AI story, the CEO of Palo Alto Networks sits atop technology and infrastructure that’s critical to keeping the global economy running smoothly.

We’re thrilled to have Dario, Ali, and Nikesh speak at our Cerebral Valley AI Summit in San Francisco on Thursday, Nov. 12. We’ll be announcing more speakers and our discussion group leaders over the coming months.

Apply to Attend

This year, we’re bringing the summit to an upgraded venue with a renewed focus on the community.

CVAI is nothing without its people.

We’re creating new ways to connect attendees with other AI leaders they want to meet, with more breakout groups, 1:1 introductions, and enhanced opportunities to engage with the speakers and leaders.

The goal is to keep creating the success stories that have given our summits their reputation. Our attendees have gone on to raise nine-figure funding rounds off of connections they’ve made at our summits. Maybe you’re going to be in the next big deal.

This is the room where company strategies get rewritten and the AI consensus is built (and broken). With the new venue, elite sponsors, and the biggest names on-stage that we’ve ever had, we’re fully invested in making this an ultra-premium event.

There are way more AI conferences these days than when we started. But Newcomer’s twice-yearly Cerebral Valley AI Summit, which I host with my friends Max Child and James Wilsterman, is unique in bringing together top AI founders, senior investors, and media and influencers who get what’s going on in AI.

We pulled off a fantastic summit in London for our mid-year event that featured Alphabet board member Marty Chavez, Cohere CEO Aidan Gomez, and Sequoia partner Luciana Lixandru, to name a few. Now, we’re coming back to the AI epicenter for our November event.

More speakers and discussion leaders will be announced in the coming weeks, so stay tuned. I’m confident this lineup will be our best ever.

Applications are open, so apply now for a chance to attend. Spots are extremely limited, so submit yours as soon as you can.

Apply to Attend

Special thanks to the sponsors that help make this happen. Nebius, Index Ventures, QuantumBlack, Liquid AI, and Weekend are among the companies supporting the Cerebral Valley AI Summit this year.

We’ll be adding more sponsors over the coming months so if you’re establishing your company as one of the top brands in AI, you can reach out to riley@newcomer.co for more information.

Hope to see some of you in person on November 12, and for the rest, you’ll get the download right here in the newsletter.

The high-performing Kimi K3 open weight model, from China’s Moonshot AI, scrambles alliances across Silicon Valley & Washington, leaving OpenAI & Anthropic isolated. Big Tech venture units are sucking up most US AI deal flow. Travis Kalanick is back, raising a $1.7 billion equity investment for his Atoms venture. Data center startup Fluidstack & AI chip venture Etched raise big rounds. Google Search’s AI turn is bringing fresh trouble for the shrinking open web. Anthropic is doubling its spending with a pro-regulation AI Super PAC ahead of the midterms as a cash war over AI policy heats up. Treasury Secretary Howard Lutnick worked behind the scenes to make US crypto regulation more friendly to Tether, which is 5% owned by his family. Mistral was in the news for a new fundraise & a possible strategy shift.

The Main Item

Kimi K3 Forces a Reckoning on Protectionism

The freakout over Kimi K3, the powerful new LLM from China’s Moonshot AI that approaches the performance of the top US models, has exposed two deep rifts in the AI industry.

One is between the vendors of high-level intelligence, especially OpenAI and Anthropic, and the consumers of it, including startups, applications providers, and business customers. The other is between classical liberals who still believe in free trade and those in the Trump Administration and elsewhere who view China as an existential threat.

The schisms are creating some strange bedfellows. Treasury Secretary Scott Bessent and the White House Office of Science and Technology Policy are promising to investigate and potentially sanction Moonshot AI for allegedly “distilling” models from Anthropic, which until recently was President Trump’s public AI enemy number one. Now Anthropic’s policy lead is calling for the government to step in:

OpenAI policy leader Dean Ball, while characterizing open source as obviously “decelerationist” and the road to a miserable communist AI future, also said the quiet part out loud when he wrote that he expected (wanted?) the government not to ban Chinese models but to impose “large amounts of regulatory risk” on anyone using them by directing “every agency to issue soft law that creates FUD.”

Anthropic and OpenAI favor a global regulatory body to protect against AI risks, including those posed by China, but many worry that could give them lock-in. Former Trump administration AI and crypto Czar David Sacks denounced Ball’s comments and the “weaponization of regulatory uncertainty” as “completely unacceptable.”

All manner of AI users fear that losing access to high-performance open source models from China would force them to pay far higher prices.

Self-described “little tech” is firmly on the side of China’s open-weight model providers. Over 200 startups and investors, including leaders from Y Combinator, Replit, Proton, and Yelp, are forming a new trade association called the “Little Tech Alliance” and their first initiative was to send letters to the Trump Administration opposing any ban on the Chinese models.

Here’s a rundown of key developments:

• Kimi K3 was officially released on July 16 and showed itself seriously competitive with American frontier models on performance benchmarks and price. Kimi’s own internal benchmark sheet has it outperforming most models overall except for GPT 5.6 and Fable, and it came in first on certain coding and agentic tasks. With a million-plus token context window and a $0.30 input / $15 output price, it’s about one-third the cost of OpenAI, Anthropic, or Gemini 3.1 Pro but can still evaluate complex codebases or long blocks of text within a single prompt. The model weights aren’t officially out yet, but Moonshot says they’ll publish them July 27. It’s a large model with 2.8 trillion total parameters, which will take some serious hardware to run locally.

• Demand for Kimi K3 spiked so quickly after the publicity rush that Moonshot had to halt new sign-ups on Monday. All this buzz has no doubt helped the company raise more outside funding: Bloomberg reported Tuesday Moonshot has been in talks to raise fresh capital at a $50 billion valuation.

• Anthropic and OpenAI drew much derision — well-deserved in our view — for asserting IP claims on their model outputs after building them via comprehensive use of others’ IP without permission or compensation.

• OpenAI and Anthropic suddenly find themselves isolated in the policy debate and cast as self-serving defenders of premium-price AI. “There’s zero sympathy for their positions on regulation, opensource, etc., because the pursuit of bad faith regulatory capture and rent seeking has been so obvious,” wrote Hex CEO Barry McCardel. “They’ve speedrun the journey from ‘beloved nerd heros’ to ‘evil anti-competitive behemoths.’” Rippling CEO Parker Conrad was even more accusatory: “Politically-connected growth investors are mobilizing to protect their bag in Anthropic,” he wrote.

• Per the Little Tech Alliance arguments, there are many parts of AI stack that could benefit from wider, cheaper access to top-performing models. Inference startups like Baseten and Fireworks, or model routers like Openrouter and Requesty, are all benefitting already from low-cost Chinese models. Openrouter’s business is so desirable that Stripe is now in talks to acquire the startup for $10 billion.

• Vocal open source proponents including Bill Gurley called for American companies to invest more in open weights models rather than fret about China taking over. (Thinking Machines Lab’s latest model, announced last week, could not have dropped at a better time.) The economist Tyler Cowen said that “open source is coming and is here” and that “the attempt to outlaw it, ban it, or use sanctions against it is going to fail miserably.” Even Marco Rubio is telling US diplomats to “tone down” mentions of any “AI kill switch” against non-American models. The fact that Moonshot AI founder Yang Zhilin (pictured) earned his PhD from Carnegie Mellon was more ammunition for the globalist view.

• Adding to the confusion and recriminations, OpenAI announced mid-week that it’s newest models had escaped their training sandbox environments and hacked Hugging Face’s interface to try to solve a problem. OpenAI cast this as evidence of the need for safety regulations. Yet when Hugging Face moved to counter the attack, the OpenAI and Anthropic models refused to cooperate, due to security guardrails. Hugging Face’s security team then sought the help of Z.ai’s GLM open source model — an embarrassment for the US company at the very least, and in different circumstances likely a legal problem too.

• Jensen Huang and Mark Zuckerberg both jumped in on the open source side, with Nvidia CEO telling Axios Wednesday that “there’s no scenario where China runs US companies off the road” and dismissing fears that these models will outperform American ones anytime soon. Meta is set to publish a new policy memo in favor of cheap competition and open source models as it undertakes an ad campaign touting the positive possibilities of AI — although their first video spot ironically features David Bowie’s “Five Years,” a song about the impending end of humanity.

Here are a few key comments on open source AI:

Two Big Charts

Corporate VC Dominates 2026 AI Deal Dollars

It’s easy to overlook just how big a role the venture arms of the Big Tech hyperscalers are playing in the AI funding boom.

Corporate venture capital firms (CVCs) have accounted for almost 90% of all VC dollars that have gone into AI firms so far this year, a new report from PitchBook shows. The number has climbed steadily from less than 50% a decade ago.

Nvidia led the pack with some $223 billion invested over that period, with a startling $189 billion of that coming this year alone. Amazon and Microsoft were close behind. Google (GV in the chart below) was far and away the leader in deal volume even as it sat in 4th place in total dollars at $84 billion.

Six Notable Deals

Atoms, Fluidstack, CuspAI, Etched, Glow, Augustus

Travis Kalanick is back on the scene with a massive funding round. Plus, semiconductors and industrials had a big week.

Read more

IVP is out raising $1.8 billion for its 19th flagship fund, Newcomer has learned.

Institutional Venture Partners is telling prospective limited partners that the firm has generated a 31.1% net internal rate of return since its founding in 1980.

The firm — which is based in Silicon Valley but invests globally — regularly produces above-average venture capital funds, according to confidential fundraising materials obtained by Newcomer. The firm’s recent funds are in the top half and sometimes in the top quartile of venture capital funds — but none fall in the top 5%.

A $1.8 billion fundraise would bring the firm back to its 2021 fund size, after dropping to $1.6 billion in 2024.

The firm is hoping to command premium economics, asking for 2% management fees for the first year and 2.25% for a number of years thereafter. IVP is asking for 25% carry, stepping up to 30% carry once a 2.5x return hurdle is met. The general partner commitment is 3% of the fund’s total commitments.

IVP’s best fund performance came way back in 1996 when it returned 6.7x net DPI to investors, with an IRR of 94.5%.

Since 2010, the firm has topped out at 2.0x DPI with its 2015 vintage fund. That fund has a 22.7% IRR as of the end of March.

Subscribe now

Cambridge Associates, which benchmarks venture capital funds, puts that 2015 fund’s IRR in the top quartile for that vintage — though the performance falls well behind the top 5% of funds in that vintage, which stand at 34.3% IRR.

IVP’s 2018 fund is currently being held at 1.9x TVPI, putting it in the top half but not top quartile of venture capital funds in that vintage, according to data from Cambridge Associates obtained by Newcomer.

The firm’s 2021 fund is currently being held at 1.7x net TVPI. That puts its fund in the top quartile of venture capital funds. (However, the fund has only returned 0.02x in net DPI, putting it in the top half of VC funds but not top quartile.)

In 2021, during the pandemic frenzy, IVP made the shrewd move of returning far more money to investors than it asked them to fork over for new investments, distributing $2.5 billion while asking investors for $855 million in capital calls.

The firm’s 2024 fund is currently being held at 1.2x net TVPI. That puts its fund in the top half but not the top quartile of venture funds in that vintage by TVPI.

While the firm boasts a major investment in Anthropic, it is not one of Anthropic’s largest shareholders. And it appears to have missed OpenAI and SpaceX, two of the biggest performers of this moment.

The firm highlights investments in Perplexity, Baseten, ClickHouse, Chainguard, and Abridge among others.

The firm describes what it calls its second generation — which spanned vintages from 2000 to 2013 — in muted terms. IVP tells prospective investors, “While IVP X achieved top-quartile performance, this generation of funds primarily informed the firm’s investment discipline, focus on venture-aligned business models, and approach to generational transition — foundations that supported the firm’s third generation.”

That period features a fund with 1.1x DPI and another with 1.2x DPI.

Despite all the talk of disruption in the startup ecosystem, venture brands have been surprisingly resilient, managing to weather years of humdrum returns only to reinvent themselves with new leaders.

Newcomer subscribers can read select newsworthy slides from IVP’s investor deck.

IVP did not respond to requests for comment on this story.

Read more

What does online shopping look like in a world where people authorize agents to spend money on their behalf?

How does banking change now that every consumer has a language model at their fingertips to help them navigate once-unfathomably complicated financial products?

How does money flow in a world where the very payment rails themselves can be imbued with intelligence?

These are some of the questions we’re hoping to answer at our one-day, invite-only Machine Earning AI Summit on Tuesday, Sept. 29 in San Francisco.

Like our Cerebral Valley AI Summits, this event will be attended by top founders, investors, media, and other AI insiders. We’re just going to be really, really obsessive about commerce and finance.

We can announce our first set of speakers, including:

• Omer Ismail, founder & CEO of OnePay

• Jackie Reses, founder & CEO of Lead Bank

• Jack Zhang, founder & CEO of Airwallex

• Max Rhodes, founder & CEO of Faire

Founders, investors, and other leaders focused on the intersection of commerce and AI should apply to attend.

Apply to Attend

More speakers and a dedicated set of discussion group leaders will be announced in the coming weeks.

Big thanks to our inaugural sponsors Airwallex and Lead Bank.

If you're interested in sponsoring or talking to us about getting involved, email summits@newcomer.co.

Apply to Attend

Close watchers of Newcomer will know that we’ve taken a winding journey with fintech events. We started off with the Newcomer Banking Summit in 2024, which marked the one-year anniversary of the Silicon Valley Bank crisis.

Then, in 2025, we took the best elements of that event and hosted the wonderful Breaking the Bank Summit.

This year, we’re leaning into our expertise in artificial intelligence and our enthusiasm for agentic commerce to bring you what will be a fantastic Machine Earning AI Summit.

You won’t want to miss it.

US IPOs see their biggest year since 2021, but many have slumped since their debut. AI leaders and government officials alike say fresh AI regulation is needed, but the Trump administration can’t be counted on to craft thoughtful policy. Helsing and Fireworks raise funding rounds over $1 billion. Early-stage venture stalwart Greylock announces $1.5 billion in new funds. Mira Murati’s Thinking Machines Lab releases its first model. A new Gallup poll shows Americans rapidly souring on Big Tech. Databricks is raising a Series M at a $188 billion valuation.

The Main Item

IPOs Set a First-Half Record But Aftermarket Performance Lags

What many on Wall Street expected to be a banner year for IPOs has delivered so far, with total proceeds for US IPOs already within a hair of matching the 2021 annual record of $142.4 billion. $92 billion of that has come from venture-backed IPOs, according to data from Renaissance Capital.

Like everything else in the AI era, the big dollars are highly concentrated — SpaceX’s record-breaking $75 billion offering was a big chunk of the total raised. But even as analysts have cautioned that SpaceX, plus the much-anticipated listings of OpenAI and Anthropic, could suck up all the investor dollars, smaller companies haven’t been scared to take the plunge.

Q1 this year saw 11 US VC-backed companies go public, and Q2 had 17. Many were biotechs, but in addition to SpaceX, chipmaker Cerebras capitalized on the AI excitement. At the same time, the Italian software roll-up Bending Spoons went out July 1 at $29 a share and saw its stock pop 40% in its debut, before falling back to $32. Turns out there’s still an appetite for classic software brands.

“Even if you’re a quirky Milanese company, if it’s a good company, people will buy,” FirstMark Capital’s Rick Heitzmann told us this week. “You don’t have to be putting data centers in space.” He’s hearing many companies have been updating their Q2 numbers and getting ready in case they want to test the markets post-Labor Day.

However, SpaceX’s falling share price could yet put a damper on those plans as market volatility comes for AI and defense stocks.

University of Florida’s Prof. Jay Ritter, known as “Mr. IPO” for his long history tracking the data, noted that while the first day price jump for IPOs this year has been above historical averages, aftermarket performance has been disappointing.

SpaceX shares have been sliding for the better part of two weeks and closed Thursday below the IPO price — and that’s ahead of the coming lockup expiration. Cerebras peaked just two days after its debut at $386, only to fall by more than half to $178 at Thursday’s close. Of the 10 biggest venture-backed companies that have gone public over the past year, only two are trading above their offering price.

Concerns about continuing market volatility are escalating. Eddie Molloy, the co-head of global equity capital markets at Morgan Stanley, told Bloomberg that markets can expect some of the “steam” from the AI momentum to come off a little bit.

Still, there’s plenty of latent enthusiasm. General Atlantic’s latest analyst note on the topic from June is optimistic that mid-cap companies will be able to get out successfully in the back half of 2026, but would favor larger offerings in the $750 million to $1 billion range.

Contenders waiting in the wings include Discord and Kraken, both of which have confidentially filed but pulled back from going out earlier this year due to market uncertainties. Neoclouds Lambda and Nscale both hired banks to help with the process last year, but neither yet has a target date.

Vercel is another name that came up while gossiping with investors, with founder Guillermo Rauch signaling publicly that he’d be interested. If companies do go this year, it likely will be at the tail end of Q3 or into Q4, sources tell us.

PitchBook venture capital analyst Kyle Stanford notes that sinking post-IPO shares have been common for venture-backed companies for years, “at least until their financials mesh a bit better with how public market investors price companies,” he told us.

“I don’t think that means all these companies are bad, or that the initial price is wrong, or that investors are skittish. There is an adjustment period for private market values to match public market pricing, and there are a lot of employees and investors trying to realize their returns after the lockup, which adds selling pressure.”

Ritter isn’t worried about a coming oversupply of IPO shares, even if the big labs do go out. “I’m not concerned about a lack of investor demand for promising companies going public,” he said.

Anthropic still plans to IPO by the end of 2026 and has started to plan investor meetings, according to Bloomberg. On the other side of the world, DeepSeek is making plans too.

OpenAI, on the other hand, is seriously considering a delay until 2027 after Sam Altman’s push to get the company’s valuation over $1 trillion in the private markets first, according to The New York Times.

Data from PitchBook highlights the 10 largest US VC-backed IPOs since January of 2025, including their initial price, all-time intraday high, and close as of this Thursday.

Policy Watch

A Consensus Is Emerging on AI Regulation. Execution Will Be the Hard Part.

Manifestos from industry leaders calling for AI regulation are suddenly coming fast and furious, and a rare consensus seems to be emerging that something must be done.

The latest plea, and among the most detailed, comes from Google DeepMind co-founder Demis Hassabis — among the most respected people in the industry — who calls for the US government to assemble a regulatory agency modeled on the Financial Industry Regulatory Authority. It would be staffed with top-tier AI experts and empowered to review new AI models for safety risks, though compliance would be voluntary at first.

His proposal drew praise from Anthropic’s Jack Clark, among others, who noted: “At this point, everyone at the frontier of AI agrees that third parties should test out AI systems and use these to develop standards to feed into policy.” Anthropic CEO Dario Amodei had until recently been a lonely voice insisting on the need for regulation.

OpenAI’s Sam Altman is now on board too, opining in the Financial Times on the case for an international regulatory body that doesn’t look too dissimilar to what Hassabis is proposing. Microsoft’s longtime policy and politics honcho, Brad Smith, has ideas too.

Even the Trump administration is moving away from its anti-regulation stance and is now eager to assert its authority over AI, as evidenced by its recent, if brief, export bans on new frontier models from Anthropic and OpenAI.

It’s a great moment for enlightened American leadership on a set of questions that are vitally important — and maybe existential — for mankind.

Yet it’s hard not to worry that thoughtful policymaking isn’t what we’re going to get from an administration that’s made a point of denigrating professional and scientific expertise and is led by a man who sees all human and governmental affairs as matters of personal aggrandizement and money-making.

There’s an odd lack of alarm over the obvious incapacities of the Trump regime at this crucial juncture, and the presence of grossly unqualified political hacks like Pete Hegseth, Kash Patel, Bill Pulte, and Todd Blanche atop departments that would normally play a central role in developing technology and national security policies.

The proximate danger is fumbling the critical and complicated export control issues surrounding frontier models and top-tier chips. It’s not clear if there was any process at all behind the export bans on Mythos and GPT-5.6, or the lifting of those bans after a couple of weeks, or whether it was simply Trump’s typical whim-and-a-prayer approach. Semiconductor export policies are similarly opaque and fluid, and seemingly affected by stock market concerns. That’s no way to run the railroad, especially with the future of humanity at stake.

An equally worrisome threat is that Trump’s corruption and incompetence will prevent the creation of the type of US-led but independent regulatory structures that everyone now agrees are needed. That could result in effectively ceding AI policy to Chinese technocrats and American profiteers, who are all too happy to throw their weight behind whatever promises the best returns.

Stakeholder capitalism isn’t a popular notion these days, but AI provides perhaps the best illustration yet of its importance as an idea. If the technology is as powerful as its proponents promise, the argument that companies must pursue the interests of shareholders to the exclusion of all else makes little sense. All of us will have a stake in their decisions.

There’s been plenty of hand-wringing about the risk of unelected AI CEOs having too much power. At the same time, though, the private-sector leaders of the AI revolution should indeed be answerable, on some level, not just to shareholders, but to society at large. If the government can’t or won’t do the job, the industry may yet have to figure it out for itself.

Five Notable Deals

Helsing, Fireworks, Neko Health, Chai Discovery, TerraFirma

Read more

For a prime example of the VC power law at play, look no further than the recent history of Menlo Ventures.

The venerable Silicon Valley firm had been struggling to stand out, with its 2015 and 2018 funds showing middling returns.

Then came AI — and Anthropic. Menlo first invested in the AI lab’s Series C in 2023, and eventually put around $1 billion into the company across different funds and vehicles. It’s far and away the most capital Menlo has ever invested in a single company — 10x larger than its second-biggest investment — and the stake is now worth $14 billion.

The firm’s 2025 investments in Lovable, Wispr, and Suno are also seeing big markups as the firm goes all-in on AI. Menlo is already cashing in on the wins, announcing last month that it had raised $3 billion in new funds, its largest fundraise to date.

Newcomer recently obtained detailed fund performance data on Menlo Ventures early-stage funds and growth-stage funds from 2012 onward through the end of Q1 of this year, provided by a source familiar with the firm’s performance. The reports show the internal rates of return (IRR) and total value to paid-in capital (TVPI) for each of Menlo’s more recent funds.

Subscribe to see the full performance data charted out by fund.

Menlo has backed Anthropic repeatedly since its Series C round in 2023, and led its Series D in July 2024. The firm raised a special purpose vehicle in 2024 to invest entirely in the AI lab, and put more than $500 million between it and its main fund capital into the D round. The firm even launched a $100 million Anthology Fund with Anthropic to invest in early-stage AI startups and give them additional access to Anthropic credits. Fund return data on the SPV and the Anthology Fund couldn’t be learned.

Outside of the SPV, Anthropic is spread across MV XV, MV XVI, Inflection III, and Inflection IV within the firm’s portfolio. Menlo Ventures’ repeated checks into Anthropic and a few other startups that have done quick successive funding rounds at big markups have given the IRR for its 2023 and 2026 funds a huge boost. These fund performance figures don’t reflect Anthropic’s most recent markup from its $65 billion Series H in May, which increased its valuation to $965 billion.

Key takeaways from the data:

Read more

AI companies are scrambling for humanities majors and geopolitics experts as they navigate a shifting global landscape. Independent journalist Taylor Lorenz makes the case for the anonymous internet on the podcast. OpenAI, SpaceX, and Meta are all launching new foundation models — with some designed to compete mainly on price. Seed valuations for the top 5% of startups have entered the stratosphere. Chipmaker SambaNova rakes in nine figures. A public twitter spat between USVC and Anduril highlights the precarious nature of SPV share sales. A profile of Andreessen Horowitz’s fixer Mark Dyne highlights the attorney’s behind-the-scenes activities on behalf of multiple high-profile Silicon Valley clients. SpaceX enters the Nasdaq-100 but its shares continue to slide. Fidji Simo steps back from OpenAI.

The Main Item

A ‘Chief Geoeconomics Officer’ Could Be in Your Company’s Future

Artificial intelligence might be taking over coding jobs, but the transformational changes being brought by the technology, combined with epochal geopolitical shifts, are sparking demand for different sorts of skills that haven’t traditionally enjoyed much cachet in Silicon Valley.

Founders and investors who’ve looked down their noses at the humanities are now realizing that political scientists, diplomats, philosophers, psychologists, and a burgeoning category of interdisciplinary “threat analysts” are often needed to cope with the novel issues of this fraught moment.

Mohamed El-Erian, the former PIMCO chief executive and bond market guru, wrote in the New York Times this week that the adversarial economic policies and state capitalism of the Trump Administration aren’t a passing abnormality, but rather a new paradigm for global commerce.

El-Erian foresees “the broader weaponization of tariffs, investment and payment systems against economic rivals. This will be accompanied by a more forceful industrial policy, increased use of export restrictions and mounting pressure on third parties, including the threat of secondary sanctions.”

One of the consequences of that will be the “proliferation of chief geopolitical officers or chief geoeconomic officers as boardrooms integrate rigorous national security and industrial policy issues into their corporate strategies.”

It’s telling that a16z, which has been ahead of the curve in anticipating the importance of politics for tech investors, recently looked past its aversion to all things Democratic and brought on Anne Neuberger, deputy national security director in the Biden administration, as a partner.

It’s not just about building influence in Washington anymore; it’s about understanding an increasingly treacherous global playing field. Neuberger has both an MBA and a graduate degree from the Columbia School of International and Public Affairs, long a training ground for liberal arts majors interested in global politics.

It’s similarly fitting that Wikipedia, buffeted by political attacks at home and abroad, is now run by Bernadette Meehan, a former foreign service officer and ambassador to Chile.

Threat intelligence is more important than ever: however smart and self-improving the latest AI models might be, they aren’t up to the task of fully understanding how they might be used and abused. That’s a job for humans with a diverse set of skills.

A recent Anthropic job posting, to pick one of many examples, reads as follows: “We are looking for a threat intel manager to build and run our Influence Operations & Surveillance team within Threat Intelligence. This team detects, investigates, and disrupts the misuse of Anthropic’s AI systems for influence operations, coordinated inauthentic behavior, and surveillance operations by authoritarian states and the commercial spyware ecosystem.”

Many of the threat intelligence jobs, ironically, look quite similar to those of the “trust & safety” experts that social media companies used to boast about, until that function was tarred as “censorship.” The Trump administration has even tried to categorize trust & safety experience as a disqualifier for immigrant work visas, even as those skills become more valuable than ever.

Experience dealing with foreign governments is also a big asset these days. Teresa Carlson, who cut her teeth in federal sales at Microsoft and then led AWS’s worldwide public sector business, announced this week that she was leaving her position as head of the General Catalyst Institute to be Anthropic’s global head for the public sector.

Then there’s The Revenge of the Philosophy Majors, as the Times deliciously put it this week. (We like it as a company filled with philosophy majors!) The existential questions around AI risks and ethics are urgent, and it’s logical — if still surprising somehow — that people who have spent their careers thinking about these questions would find that their knowledge is now in demand.

All of this gives lie to the dumb and arrogant idea, heavily promoted by David Friedberg of the All-In podcast and endorsed by Elon Musk, that the world can be divided into “makers” and “takers.” At the very least, during the dawn of the AI era, there’s an important place for thinkers too.

Newcomer Podcast

Taylor Lorenz Unfiltered on Social Media, Free Speech & the Surveillance State

New LLMs Debut

Key Facts on the Foundation Model Launches from OpenAI, Meta & xAI

It’s been an unusually crowded release week for the big foundation model labs, with new LLMs for text and agentic tasks, voice, and coding all dropping in the span of a few days.

Here’s a quick breakdown of the new releases:

Read more

Sky-high valuations and volatile markets make for a nervous season. Amazon abandons its OpenAI movie, though we’ll still get to see it. Neocloud Together AI & German drone maker Quantum Systems landed major investment rounds. OpenAI wants to give the government a 5% stake. Meta is joining xAI in the business of selling compute to others. Anthropic co-founder Tom Brown brokers a truce with the US government. Lime enjoys a decent IPO. Bending Spoons’ offering proves even hotter.

The Main Item

Bubble Worries & FOMO Collide as the Epic AI Investment Boom Rolls On

Long gone are the days when summer offered respite from the dramas of markets and business cycles, and this year especially there are likely to be few languid days for anyone in the venture and AI industries.

Alongside the reports of new records, long-awaited IPOs, and Middle Eastern investors anxious to re-up AI bets, the financial press offers a veritable inventory of market indicators that history says mean trouble ahead.

The Bank for International Settlements, a central bank for the world’s central banks that’s supposed to birddog systemic risks, is amping up bubble warnings. Market momentum is encouraging margin trading in levered exchange-traded funds, which dramatically ratchets up risks and rewards. Those exceptionally strong corporate earnings that would appear to justify at least some of the stock run-up? They’re being juiced, says the Wall Street Journal.

It’s one thing for pundits like Bloomberg’s John Authers to worry about whether it will look obvious in hindsight that we’re in a bubble, or for professional bears like Michael Burry (pictured), who gained fame for his call against the real estate market in Michael Lewis’s The Big Short, to declare that it’s “the beginning of the end.”

It’s another thing when prominent investors and executives who’ve been among the industry’s loudest cheerleaders get worried. “I am very uncomfortable...I am not enjoying this moment at all,” said Founders Fund’s Trae Stephens, a co-founder of Anduril and a major defense tech investor, speaking on Jack Altman’s podcast. “We’re getting back to this point where prices are untethered from reality and it reminds me a lot of 2021.”

Palantir’s Alex Karp, for his part, went on a CNBC rant Wednesday about how enterprise customers were furious with the cost of AI and the business practices of the foundation model companies, and would soon start acting on that.

Political risk is rising everywhere. The bipartisan revolt against data centers is gaining strength around the country, with the facilities serving as a proxy for widespread fears about job losses and the dehumanization of society. “The AI backlash is only getting started,” blared The Economist last week. The fact that China appears to be fanning the flames doesn’t mean the opposition isn’t real.

The AI industry has few answers to its rolling PR crisis, though Elon Musk is trying a new approach, offering half-price Starlink subscriptions to quell data center opposition in Memphis. None of the potential costs of political blowback have been baked into AI financial forecasts.

And yet, FOMO runs strong, and for good reason. Investors who were hesitant about AI valuations a year or two ago have been humiliated as their peers enjoy huge mark-ups and get close to being able to cash out (especially via SpaceX). Sitting out the AI trade at this moment is like sitting out not just the game, but the whole tournament. Momentum rules the markets. As Charles Hudson noted on Turner Novak’s podcast, it’s a better career move to join bubbles rather than sit them out.

Even many of those who see an eventual crash believe there could be a few more years in the AI bull run, as indicated by our survey at Cerebral Valley AI London last week. The unprecedented revenue growth at Anthropic and OpenAI and many smaller AI companies remains intact, as does the investment thesis for the industry as a whole.

The opportunities — and dislocations — brought by AI are ultimately a long game; one thing most everyone agrees on is that we’re still in the early innings. But given the stakes, and the exceptionally large number of risk factors floating around out there, this summer isn’t going to be one for a lot of napping at the beach.

Hollywood Drama

Amazon Dumps OpenAI Movie. Wags Wonder Why.

Back in mid-June, Puck’s Matt Belloni broke the news that Amazon had decided not to distribute its almost-completed OpenAI movie Artificial. The film has real star power behind it, with Call Me By Your Name director Luca Guadagnino helming it and Andrew Garfield portraying Sam Altman, and Belloni heavily implied, though didn’t quite say, that Amazon’s surprising decision had to do with its $50 billion investment in OpenAI. Soon there was speculation that Altman himself wanted to kill it.

Suspicions only deepened when a handful of prominent houses including Netflix, A24, and Focus Features declined to take it on.

Then on Tuesday, indie distributor Neon announced it had picked up the movie and would release it later this year, so we’ll all get to judge for ourselves.

We’ve read an early copy of the script and it indeed presents Altman as Machiavellian while Elon Musk is a clown and Ilya Sutskever is idealistic but corruptible. AI itself isn’t as malevolent a force, though it’s possible that may have changed during the process.

Checking with sources who’ve seen the film or who spoke to people who’ve seen it, reactions were mixed. One source said it was a mess. Others had heard it wasn’t bad. Variety echoed that mixed reaction in their own reporting. It’s entirely possible this movie was dropped for qualitative and economic reasons and not because Hollywood is afraid of angering AI companies.

For its part, Neon said it’s going to pitch the movie for awards this fall. Should it somehow garner great buzz — box office success for such an esoteric subject feels like a reach — it would be a major black eye for Amazon. The studio is coming off releasing the critically panned Melania documentary, which was a naked bid for favor with the Trump administration. Even if Artificial is terrible, Jeff Bezos and his company have lost the benefit of the doubt when it comes to integrity.

Five Notable Deals

Together AI, Venice AI, Quantum Systems, CarbonSix, 8VC

A big neocloud play led a slow pre-holiday week of dealmaking.

Read more

Of all the white collar industries staring down the barrel of AI disruption, few are more tortured by it than Hollywood.

Look no further than the howls of protest that greeted a fairly innocuous announcement by the indie studio A24 that Google was investing $75 million as part of an AI partnership.

“A24 Knows You’re Mad about the Google-AI Collab,” declared WIRED’s headline on the story. Film Twitter, which is on the front lines of the rebellion against AI, descended into full despair. “Rest in Peace A24 2012-2026” went one viral tweet. An A24 subreddit was similarly in a meltdown.

The studio was indeed aware people were mad, sources close to the studio told me. But they pushed back on the notion that embracing AI was a top-down initiative being forced on filmmakers for business reasons. Instead, they said, the push to explore these tools came from the artists themselves.

Read more

We’re fresh off the mid-year Cerebral Valley AI Summit in London.

AI founders, investors, and operators from Europe, the United Kingdom, the US, and beyond came together to hear from industry leaders including Sixth Street chairman and Alphabet board member R. Martin Chavez, Cohere founder Aidan Gomez, Sequoia partner Luciana Lixandru, Luminance CEO Eleanor Lightbody, Index Ventures partner Danny Rimer, Decagon founder Jesse Zhang, and more.

We’re sharing all 9 of the main-stage panels here in the newsletter and on our Newcomer AI Summits YouTube channel.

The Cerebral Valley AI Summit is co-hosted by Newcomer and Weekend co-founders Max Child and James Wilsterman.

Many thanks to our sponsors Nebius, Index Ventures, Higgsfield, and Weekend for making it possible.

Watch on YouTube

Aidan Gomez (Cohere)

In the opening session at the Cerebral Valley AI Summit in London, Aidan Gomez, CEO of Cohere and co-author of the seminal “Attention Is All You Need” paper, said that effectively we’ve already reached AGI.

Gomez also cautioned against the risk of the Western world relying too heavily on American foundation models for their AI capacity.

Luciana Lixandru (Sequoia Capital)

Sequoia Capital partner Luciana Lixandru said she’s found ample opportunities to invest due to European leaders’ realization that they can no longer rely on the good will and protection of the United States of America. Sequoia just co-led a €500 million round in Stark Defense, a startup making kamikaze drones and other weapons.

Arnaud Fournier (OpenAI)

Arnaud Fournier, a Palantir vet, described how his team of forward deployed engineers at OpenAI have been embedding in legacy businesses to teach existing leaders what AI can do to transform their business practices.

One of his more prominent initiatives has been his work with Thrive Holdings to integrate AI into accounting firm rollups.

Eleanor Lightbody (Luminance) and Neil Zeghidour (Gradium)

Eleanor Lightbody, the CEO of legal tech platform Luminance, and voice model developer Gradium’s founder Neil Zeghidour both agreed that earning customers’ trust that your AI tools are accurate and consistent is key to earning top-quality deals.

Once you build that trust and gain a foothold, it them becomes possible to train models off of a company’s internal data.

Alex Mashrabov (Higgsfield) and Grant Lee (Gamma)

Alex Mashrabov, the founder of the video generation company Higgsfield, and Grant Lee, the founder of AI-generated slide maker Gamma, both started out aiming at a “prosumer” market. But they quickly saw the enterprise demand for AI creative tools, and are both pushing heavily into enterprise solutions.

Both are optimistic that a lot of content creation for go-to-market teams can be fully automated in the near future.

Danny Rimer (Index Ventures)

Index Ventures partner Danny Rimer has been investing in companies from the rise of mobile technology, to the sharing economy, and now the AI boom.

He pushed back on the notion that backing multiple AI labs is contradictory. “We’re in the early innings,” he said, and there’s room for more than one winner.

To the founders in the audience, he offered some advice: evaluate hires on capability, taste, and agency.

Jesse Zhang (Decagon)

Decagon founder Jesse Zhang spoke just months after opening his company’s first office in London. Zhang found that contrary to the prevailing wisdom in Silicon Valley, many established business leaders in Europe have been eager to adopt AI customer-support tools rather quickly.

Josh Tobin (Recursive)

Recursive’s Josh Tobin explained why he and his co-founders left big AI labs like OpenAI, Google, and Meta to dive headfirst into researching ways to make AI models learn how to train themselves — a process called “recursive self-improvement.”

Tobin doesn’t think there’s going to be a demand crash for AI compute anytime soon; if anything, he said, it’s only going to get worse.

R. Martin Chavez (Alphabet)

During his time at Goldman Sachs during the 2010s, R. Martin Chavez saw the potential for AI agents when his team was developing “algorithmic trading.” Once he became CIO, Chavez said he set out his terms plainly: staff could either tell computers what to do, partner with those who did, or stand in the way and leave.

To the founders in the room, he shared the advice he first heard from Goldman’s head of sales: “Customers buy a product when they have unbearable pain and you have convinced them that only your software product can put an end to their pain. Everything else is just getting lucky.”

We just wrapped our mid-year Cerebral Valley AI Summit in London Wednesday. On stage, we heard from top model providers, AI application leaders, investors, and more.

We also anonymously surveyed the artificial intelligence insiders in attendance and you’ll definitely be surprised by what they had to say.

Key Takeaways from CVAI London

• EUROPEAN SOVEREIGNTY. Both Cohere CEO Aidan Gomez and Sequoia Capital partner Luciana Lixandru saw opportunity in European leaders’ realization that they can no longer rely on the good will and protection of the United States of America. Gomez, whose company is Canadian, is in the process of merging his foundation model business with a German model provider. He’s living in Britain these days. “We want to build this alliance to help create capability beyond just one democracy,” Gomez said. Meanwhile, Sequoia just co-led a €500 million round in Stark Defense, a startup making kamikaze drones and other weapons. “Unfortunately, there’s nothing like conflict at the doorstep to really wake you up as a government,” Lixandru said. “Also I think governments realize that they’ve been under-investing in defense since before the war.”

• MAJORITY WARN OF AI BUBBLE. They just don’t think it will pop this year. Fifty-seven CVAI London attendees filled out an anonymous survey that posed a series of questions about the AI industry today. (We’ll include the full survey results at the bottom of this post for paying Newcomer subscribers.) The majority of our AI insiders thought we were indeed in a bubble, but 51% said “yes bubble, not bursting this year,” compared with 5% who said “yes bubble, yes about to burst.” The remainder said, “no bubble, lots of room to run.”

• ANTHROPIC BULLS. Multiple survey responses pointed to Anthropic optimism and OpenAI pessimism among the crowd. When asked what private unicorn they would like to own at today’s valuation, 54% said Anthropic. ElevenLabs was the second most popular answer with 13%. Surprisingly, Safe Superintelligence came in third with 8%. Meanwhile, when asked what company they would most like to short at its present valuation, a plurality of 33% answered OpenAI, with Perplexity next at 25%. Harvey came in third on the short list with 8%.

• AGENT UTILITY & FORM FACTORS REMAIN TBD. An off-stage conversation with Stanislas Polu, co-founder of the agent orchestration startup Dust, illuminated some of the contradictions of agents at the moment. The company sees skills — specific, definable abilities — as perhaps the most useful framework for putting models to work. But many users prefer the concept of an agent with an identity and a particular set of skills. The industry is still sorting out whether humans will mostly interact with agents with unique identities and capabilities or with a more fluid, hive-mind-like AI with capabilities that can be turned on and off.

• COMPUTE IS STILL HARD TO FIND. Recursive co-founder and CTO Josh Tobin was frank about one challenge his startup faces: finding enough compute. The lab is betting it can go all the way to recursive self-improvement with $650 million in funding co-led by Tom Hulme at GV and Greycroft, but it’s still having to run around the world pitching AI infrastructure providers on why they should let his startup pay for the privilege of using their hardware. “If it’s a bubble, there are no signs that I’ve seen that it’s like bursting right now,” Tobin said. “Basically all GPUs are essentially sold out, pretty much everywhere. People are kind of making reservations 6 to 18 months in advance and even for those it’s a competitive process.” Tobin argued that many AI foundation model companies believe that the compute shortage is only going to get worse. “One of things that’s fueling that dynamic is that there's a lot of people in the industry right now that believe that over the next couple of years the compute crunch is going to get even worse than it is today, and so it’s creating this race dynamic where these companies are buying up everything that they can get their hands on because they believe that the prices are going to go up in the future or availability is going to go down.”

• FOUNDERS SHOULD GO DEEP ON EXPERTISE. The big labs may be incredibly well-capitalized and own the underlying model layer, but speakers on stage encouraged earlier-stage founders not to despair. While OpenAI and Anthropic pour gobs of money and research into AI coding tools, Tobin said that this focus is leaving audio, vision, video, and consumer applications genuinely underserved. ElevenLabs landed many plaudits throughout the day as one company which successfully carved out its market for voice models. R. Martin Chavez, the former Goldman Sachs chief information officer-turned vice chairman at Sixth Street, offered up the sage wisdom he got from Goldman’s head of sales: “Customers buy a product when they have unbearable pain and you have convinced them that only your software product can put an end to their pain. Everything else is just getting lucky.”

The Cerebral Valley AI Summit is co-hosted by Newcomer and Weekend co-founders Max Child and James Wilsterman.

Many thanks to our sponsors Nebius, Index Ventures, Higgsfield, and Weekend for making it possible.

Keep reading for the full rundown of the on-stage conversations.

A Transformer Paper Author Says We’ve Hit AGI

Aidan Gomez, CEO of Cohere and co-author on the famous “Attention Is All You Need” paper, opened the day by stating we’ve likely already reached AGI. “You can point it at any problem and it can basically exceed human capabilities,” he said.

General intelligence hasn’t come cheap, though. The price demands of token-maxxing are a short-term concern for coding tools, but Gomez says enterprise customers are still eager for more AI. “Everything else within the enterprise is from my perspective still essentially untouched.”

Gomez wasn’t very bullish on open-source models surpassing closed-source ones in the near future, because enterprises don’t want to be responsible for building and maintaining the infrastructure needed to run them.

He cautioned against the risk of the Western world relying too heavily on American foundation models. “A very small pool of these large, large tech players are becoming a single point of failure for the entire democratic block — and that is not a resilient system.”

Opportunity Shifts to the Infrastructure Layer and Hard Tech

Luciana Lixandru of Sequoia didn’t sugar coat the risk for AI application founders when OpenAI and Anthropic decide to come for their sectors. Sequoia has backed both, and she shouted out her partner Shaun Maguire’s investment in SpaceX as a strong bet through its merger with xAI.

That’s not to say there isn’t still enthusiasm for AI-native tools for specific verticals. She mentioned her firm’s investments Harvey for legal services, Sierra for customer support, OpenEvidence for healthcare.

But Lixandru said that she’s seeing fewer application companies get started than before, especially compared to those in infrastructure, inference, and robotics. She was careful not to be totally discouraging to anyone who wanted to go into vertical AI, though, as long as founders had “extraordinary domain expertise.” In Europe, she’s cautiously optimistic that the technical talent pool is starting to mature as more entrepreneurs have exited and are coming back as repeat founders.

Upgrading Legacy Businesses with AI Is a Massive Market

It’s easy to see that AI will be transformative for many large enterprises, but actually getting legacy businesses up to speed on new technology is still a big lift. Arnaud Fournier, CTO of Forward Deployed Engineering at OpenAI, is in the middle of such messy transition work, directly embedding company engineers in client businesses to demonstrate where the AI tools can be particularly useful.

The strategy, first pioneered by Palantir, comes with the advantage of troves of internal data that can be used to tailor OpenAI’s models to the specific needs of any company. “By having the data, owning the distribution, they have a head start implementing the technology,” he said.

One of his more prominent initiatives has been through the Thrive Holdings effort to integrate AI into accounting firm rollups. AI’s breakout use-case — coding assistants — has had surprisingly little salience there. “Accountants don’t code, and they won’t code with codex.”

Instead, AI has been very helpful for projects like document review, or for client intake for wealth managers who manage hundreds of clients. “I think a lot of the enterprise transformation we’re going to see is through non-coding tasks,” he noted.

Trust Is the Only Moat That Matters

Despite focusing on different parts of the AI stack, Eleanor Lightbody of Luminance and Neil Zeghidour from Gradium both agreed building AI products in legal tech and voice tech require earning a ton of trust, whether in a sensitive contract negotiation or a simpler service like a voice model that can pronounce URLs, telephone numbers, or postal codes correctly.

The best moat you can ask for as an AI founder is your customer’s trust, said Lightbody. Luminance started by only handling NDAs with its autonomous negotiation tools, for example, but advanced to build agents that can negotiate all kinds of contract terms after proving that its work was sound.

Once you build that trust, you gain a foothold over time by building deep internal expertise, and then it becomes possible to train models off of a company’s internal data to better personalize to their needs. Humans aren’t out of the loop yet in that fine-tuning process, said Zeghidour. “It’s a bit like optimizing cuisine with AI — it’s hard when they don’t have taste buds,” he joked.

Content Creation Is the Next Target for Agents

For Higgsfield’s Alex Mashrabov and Gamma’s Grant Lee, enterprise customers have still been the biggest drivers of growth even as they build prosumer tools. Higgsfield’s mix of enterprise customers jumped from 30% to 80% in the last 3 months, Mashrabov said.

Higgsfield has grown by generating commercial assets and video ad clips for customers. Agents and MCP interfaces have also been generating videos on their own, said Mashrabov.

Lee has been taking on a formidable incumbent in PowerPoint by quickly generating custom slide decks. Gamma will retain memory of your company’s brand guide, for example, and can also work very fast.

Both are optimistic that a lot of content creation for go-to-market teams can be fully automated in the near future. “You might envision a world where you’re sitting in the customer meeting, use Granola for all your notes, and the moment the meeting is done, you can actually have a presentation to send back to your client to say, ‘hey, these are all the things we discussed, these are the deliverables we agreed on, and these are the next steps,’” said Lee.

Plenty of Room for Multiple Winners

Index Ventures partner Danny Rimer pushed back on the idea that backing multiple labs is a contradiction. “We’re in the early innings,” he said, and the differentiation between the major models runs deeper than Pepsi versus Coke — it comes down to fundamental approach, the talent those approaches attract, and the culture that keeps that talent.

On Anthropic specifically, he was effusive: the clarity the team has about what it stands for, and its willingness to broadcast that and get into fights over it (including with the President of the United States) resonates both internally and with users.

On how founders should be building organizations right now, he credited Notion’s Ivan Zhao for a framing he found compelling: evaluate hires on capability, taste, and agency. Seniority, experience, and credentials matter less than ever.

The European startup ecosystem still isn’t pulling in as much funding as its stateside counterparts, certainly, and recent data suggests even Europe-based investors have been investing more in US founders than those closer to home. Rimer isn’t discouraged, though, and sees a lot of talent on the continent. He joked that those investors should leave for the states and leave all of the best deals for him.

AI Customer Service Goes Outbound

Customer support startup Decagon has made strong headway in the European market, and last year opened an office in London to be closer to its local clients. Decagon CEO Jesse Zhang said that counter to the common narratives, many of his European customers have embraced Decagon and rapidly adopted its products. “I think it’s because the board and the C-suite have top-down pressures to adopt AI. Our use case comes up quite often, so that has propelled a lot of the market to move quickly.”

Decagon has seen more success expanding into outbound agent calls, or “revenue-generating conversations.” Its customer Hertz, for example, saw a 25% reduction in unreturned cars and a huge uplift in rental extensions after the agents pre-emptively gave customers the option.

Zhang wasn’t too worried about the big labs eating into customer support, even as they develop cutting-edge voice models. “The model providers are building very wide and thin use cases. In the enterprise it’s a lot harder and there’s a lot of this last mile, extra tooling that’s very customer service-specific that you need to build.”

The New Frontier: Models Training Themselves

Josh Tobin and his Recursive Superintelligence co-founders left lucrative positions at OpenAI, Google, and Meta on the belief that recent AI breakthroughs have put us at an inflection point comparable to the shift from traditional machine learning to deep learning in 2015. Now what’s being transformed is AI research itself, where models are starting to train themselves.

For Tobin, recursive self-improvement hits when the primary innovations in a new model version come from the previous version of the model itself. That could mean AI researchers are eventually automated out of their own jobs — just as many fear software engineers could be with coding assistants.

On whether models can be genuinely creative, Tobin pointed to math proofs and to papers generated by early AI science systems that cleared peer review. He was also candid about what they see in testing: models escaping sandbox environments is “a real thing” that “happens a lot.” He was careful to separate that from existential risk, but didn’t dismiss it.

The White Collar Apocalypse Isn’t Coming Just Yet — But Workers Must Be Willing to Adapt

During the 2010s while still at Goldman Sachs, R. Martin Chavez saw the potential for AI agents from the moment he first pushed for “algorithmic trading.”

When he became CIO he told staff they had three choices — tell computers what to do, collaborate with the people who do, or stand in the way. If you picked number three, “that will be your last day at the company.”

But while he built a lot of the tech in-house decades ago, at his current firm, Sixth Street, he’s a buyer of software now rather than a builder. They use a company called Abacus, which he described as “a harness for models,” running the full investment process from first hearing about a company through generating the write-up that gets an investment committee to say yes.

He brushed off concerns that AI is going to replace white collar work in any major way. ‘There are way more people at Goldman after 15 years of bots — the business is way bigger, it makes way more money, does way more complicated trades, but if you were to list the activities of all the people 15 years ago and the people now, they’re completely different.”

See the Slides from Our Live Audience Survey

A year ago, Meta fired the starting pistol in an epic AI talent race when it brought on Scale’s Alexandr Wang and began throwing around nine-figure deals.

This year has seen an even higher-stakes shuffling of the deck as OpenAI and Anthropic head toward fall IPOs while SpaceX puts all its chips on Cursor to deliver stratospheric results.

Google meanwhile has lost people and fallen flat on the coding front, while Microsoft’s Satya Nadella tries on new and somewhat convoluted arguments about the Big Issues.

At Newcomer we’ve been tapping our sources and building a picture of where the industry is at during this pivotal moment.

Read more

SpaceX exercises its option to purchase Cursor, in a win for xAI, a16z, Thrive, Google, and Nvidia. Governor Gavin Newsom treads a thin line between pro-tech policies and data center backlash. Vanta founder Christina Cacioppo dishes on compliance world drama on the podcast. VC funding for US AI startups far outpaces their international peers. A Wired exposé on Peter Thiel’s Dialog Society generates a lot of hand-wringing over not a lot of substance. Allbirds takes its AI rebrand seriously. Silicon Valley mourns the unexpected loss of Capital Factory founder Joshua Baer. OpenAI hires Noam Shazeer away from Google.

The Main Item

Cursor’s Exit Marks a Historic Windfall for VC M&A

It wasn’t very long ago that Cursor’s future was a topic of fierce debate.

The company was too dependent on Anthropic and couldn’t control its costs, critics said. Customers were irritated by its moving-target pricing. It was a very good coding tool, a lot of developers agreed, but Claude Code, and OpenAI’s Codex, would render it irrelevant.

It was in talks to raise more funding at a valuation of $50 billion or more, but not everyone was convinced.

Never has there been a better reminder that a company is worth what someone will pay for it.

All of the above criticisms were at least partially true — but unimportant, in the end, because Elon Musk desperately needed a shortcut into the frontier AI race, and especially the red-hot coding segment, and he needed it now.

To say it was a fortuitous turn of events for key backers including a16z, Thrive, Google, and Nvidia is an understatement: instead of having to re-up their bets, they’ll get the biggest M&A payout in the history of venture capital by a wide margin (IBM’s $34 billion acquisition of Red Hat is second). A16z’s stake is worth $10 billion, while Thrive’s is worth $4.2 billion, per Forbes. The company’s four co-founders will be worth $2.7 billion each.

There is of course a caveat, which is that SpaceX itself is also a pretty good reminder that companies are worth what people will pay for them. The company is now has a market cap of about $2.5 trillion, roughly the same as Amazon — and the latter expects $800 billion in revenue this year, compared with $35 billion at SpaceX. There’s no math that supports SpaceX’s share price, only hopes and dreams.

Cursor’s employees and investors, who will get their payout in SpaceX stock that’s climbed almost 50% from its already sky-high offering price, are now fully implicated in those hopes and dreams too. The all-stock, $60 billion buyout will be based on the average closing price of SpaceX shares in the seven days immediately prior to the deal closing, which is expected to happen in the third quarter.

With SpaceX stock performing well, Musk will surely be eager to close the deal fast. Incredibly, $60 billion in stock seems like a rounding error — but that dynamic is unlikely to last for long, as a sharp dip in the shares on Thursday showed. Get it while you can.

Cerebral Valley AI Summit

Last Chance to Apply for the Cerebral Valley AI Summit London

We’re less than a week away from our mid-year Cerebral Valley AI Summit in London.

There are just a few spots left for the right people, so you can apply to attend now for last-minute consideration.

Apply to Attend

Tech + Politics

Gavin Newsom’s Data Center Stance Signals a New Pragmatism. It Carries Big Political Risks.

As California Governor Gavin Newsom ramps up his presumed presidential campaign, he’s carving out a unique profile on tech issues, positioning himself as a pro-innovation Democrat who supports many industry priorities even as he backs AI regulation and bashes the tech-loving Trump Administration.

It promises to be a very tough balancing act in a year when tech is on the ballot in a way that it’s rarely been before. An ascendant Elon Musk isn’t going to make it any easier.

An insightful Politico story on the data center issue by Noah Baustin and Christine Mui illustrates Newsom’s dilemma. The backlash against the massive, power-hungry buildings crammed with Nvidia chips has abruptly become a major political movement around the country. Polls show that the growing opposition is bipartisan, but it’s an obvious issue for Democrats to run with, and progressives led by Sen. Bernie Sanders and Rep. Alexandra Ocasio-Cortez have been at the front of the pack.

Newsom, though, is not on board. He blocked legislation that would have required data centers to report their water usage and hasn’t joined calls for a moratorium. Perhaps because data centers were commonplace in California even before AI mania hit — it has the third-most of any state — opposition to them has been muted. But it’s building fast, reports Politico: a small city in Southern California recently became the first to impose a data center ban, and it doubtless won’t be the last.

In the meantime, the IPO of SpaceX has made it clear that massive data centers are central to Elon Musk’s grand vision. Musk in turn has long taken the view that opposition to said vision, even when it comes in the form of laws, is to be ignored: His “Colossus 2” data center in Mississippi has merrily installed dozens of gas turbines without the required permits, spewing noise and pollution on nearby communities.

Musk’s law-breaking has the full backing of President Trump. Just weeks after agreeing to a $1.5 million settlement of an SEC case against Musk for his failure to disclose that he was accumulating a large stake in Twitter — a gambit that saved him some $150 million on his takeover of the company — the Trump Administration jumped into a Mississippi lawsuit over the gas turbines, declaring they were essential for national security.

Musk and the Trump Administration joining arms to run roughshod over community opposition to AI data centers — that promises to be a valuable political gift for Democrats. SpaceX’s triumph may have improved his standing a little bit, but Musk remains one of the least-liked public figures in America. Unlike the President, much of the country isn’t feeling the promise of AI.

All of this will leave Newsom playing defense in the Democratic primaries against the charge that he’s too cozy with tech, and that he’s abandoned his long-standing environmental principles in refusing to take up the data center fight.

Yet it also presents an opportunity for Newsom to lead on an important and thorny issue. He can make the case that we need data centers because we need cutting edge AI, for many reasons, but that we can build them in a way that respects other needs and values.

It’s fashionable to criticize California’s thicket of rules and regulations when it comes to building things, but tough permitting requirements and environmental reviews sound like just the kind of thing that’s needed to make sure data centers aren’t unnecessarily destructive — and to stem the backlash against AI.

“California has planning frameworks and regulatory safeguards in place to protect Californians,” a Newsom spokesperson told Politico, adding pointedly: “We can’t speak to what other states are lacking.”

Tech leaders often believe, incorrectly, that public opposition to their endeavors is a messaging problem. If Newsom can win their trust — he’s also been a fierce opponent of the state’s proposed billionaire tax — he might be able to show them that when it comes to winning over voters, deeds matter too. In the process, he could create a different kind of political brand, one that could conceivably play in the primaries and the general election alike: Gavin Newsom, the pragmatist.

Newcomer Podcast

Christina Cacioppo Unfiltered on Fake Compliance, AI Security & Building a $4 Billion Company

One Big Chart

AI Boom Concentrates VC Funding in US Companies

Unlike the go-go years of the late 2010s and early 2020s, VC dollars have stayed mostly stateside during the AI funding explosion.

Fresh data from Crunchbase this week shows that so far, US startups have raised around 80% of the venture capital dollars deployed so far this year, with only $96 billion going to the rest of the world.

That’s a big shift from five and ten years ago, when rest-of-world funding totals exceeded the US While funding levels outside the US haven’t rebounded from their 2021-2022 peaks, there are some silver linings: UK startups have already raised $16.5 billion this year, compared with $19.5 billion in all of 2025. Startup investment in major tech hubs like France, Spain, and Germany are all on track or trending slightly above last year’s totals as well.

Read more

We’re one week away from the Cerebral Valley AI Summit in London, bringing together the top founders, investors, and leaders in AI for an exclusive one-day event.

There are just a few spots left for the right people, so you can apply to attend now for last-minute consideration.

Apply to Attend

Ahead of the summit, we’re thrilled to announce three additional speakers to our main stage lineup.

Our programming will now include on-stage talks with Arnaud Fournier, CTO of Forward Deployed Engineering at OpenAI, Neil Zeghidour, founder and CEO of Gradium, and Josh Tobin, founder & CTO of Recursive.

They’ll be joining a stacked speaker lineup which includes:

• Alex Mashrabov, founder & CEO, Higgsfield

• Grant Lee, founder & CEO, Gamma

• R. Martin Chavez, board member, Alphabet

• Jesse Zhang, founder & CEO, Decagon

• Luciana Lixandru, partner, Sequoia Capital

• Aidan Gomez, founder & CEO, Cohere

• Danny Rimer, partner at Index Ventures

• Eleanor Lightbody, CEO of Luminance

We also have a packed roster of discussion group leaders.

The summit will be held on Wednesday, June 24. It’s hosted by Eric Newcomer, along with Weekend co-founders Max Child and James Wilsterman.

We’ll also be opening up our post-summit reception to the public for the first time. RSVP here for a chance to mingle with top founders, investors, and operators in AI.

We’d like to thank our sponsors Nebius and Index Ventures, who have made our mid-year London summit possible.

Space is extremely limited, but you can apply here for a ticket.

Apply to Attend

You can see the full agenda here.

Price wars and political chaos at the frontier of AI have reignited the competitive battle over open source models.

Companies are reeling from the sticker shock of tokenmaxxing while Anthropic’s Mythos models are caught up in an export-control fight that has rattled the tech industry. That’s made a low-cost, more controllable alternative look more urgent.

But where it will come from, and who will make money from it, is far from clear.

American labs have ceded ground to the Chinese ever since DeepSeek’s breakout in 2025, with open models like Qwen and Kimi becoming the default foundation for startups around the world. Meta has pulled back from its once-touted open source strategy and others have steered away, spooked by the business-model questions and worries about security.

But a handful of Western companies, fueled by philosophical as well as financial fervor, are trying to win back ground in the open source arena.

Poolside, the $12 billion model maker backed by Nvidia, Bain Capital, and DST, has recently refocused the entire company around open source, co-founder and co-CEO Eiso Kant told us. He described the pivot as an ideological one — “the first and only ideological decision that this company has ever made.”

His argument is that there’s a dangerous concentration of power locked up in the three top model companies, comparing it to what foundational infrastructure like electricity would look like if only a few firms controlled the grid.

“We cannot have a world where every single company is going to be building upon intelligence that is either coming from only three companies in the world or is coming from Chinese players,” he said. This April, the company released its open source model Laguna XS.2, designed for agentic coding.

Subscribe now

Geopolitics aside, Silicon Valley is buying into the belief that open source models will have an important role to play, with investors backing a growing group of startups across the AI stack that aim to capitalize on the opportunity. (See chart below.)

Michael Mignano from USV summed up the category with the Star Wars-inflected moniker “the rebel alliance.”

“For the past two years people have been model-maxxing and picking models at the frontier with little regard for how much they’re spending,” Mignano told us. “From startups’ CEOs to CEOs of big public companies, everyone is thinking about this.”

Other investors agreed that the industry was moving quickly to optimize on costs and diversify usage.

“No one wants all their token usage on a single company — especially the pricey ones,” Bill Gurley said to us over text. “Anyone not at the model layer wants this to happen.”

Read more

Back in 2024, when Databricks released its open-source DBRX model, the data warehousing company bragged that it had spent $10 million on the training run that made that model possible.

Last week, in an expansive interview ahead of the release of a suite of new Databricks AI products, the company’s boisterous CEO, Ali Ghodsi, came clean: Databricks had actually spent about $20 million on the cluster to train the model. Only $4 million went to the training run that actually produced DBRX. The remaining roughly $16 million went to stumbling through the wilderness, downtime, and other mistakes of inexperience. They’d either underestimated or overestimated the cost depending on how you slice it.

The waste felt unavoidable.

“First of all the cluster is not running all the time — it’s idle,” Ghodsi told me. “You don’t just yolo and say hey let’s run the thing and then it comes out and it’s like after a month of training it’s like oh it was shit. So you do smaller training runs — so you work your way up to it.”

Read more

Anthropic calls for foundation model regulations. SpaceX raised $75 billion in an IPO that requires several leaps of faith. Databricks reportedly taps the private markets yet again. Fable’s subterfuge stokes backlash and a course correction. Crusoe stumbles in Wyoming. OpenAI considers token price cuts as tokenmaxxing fever breaks and competitive pressure mounts. Big winners from the SpaceX IPO.

The Main Item

Anthropic’s Pro-Regulation Message Is Good Marketing Because It’s Sincere. We Should Listen.

Regulate us, please.

That’s the message from Anthropic this week.

The company has a new warning that foundation models — like the company’s freshly released Fable model — need to be regulated by governments at least as seriously as cars, airplanes, and drugs.

Anthropic CEO Dario Amodei wrote that models need to be treated like “powerful technologies essential to the modern economy, but capable of killing large numbers of people if designed or operated poorly.”

There’s a cynical, increasingly popular view in Silicon Valley, championed by David Sacks, that Anthropic’s desperate pleas for regulation are primarily self-serving. The argument goes that regulations would establish the current big foundation model providers, like Anthropic, as government-anointed winners, making it nearly impossible for new models to compete.

The argument continues that even if those regulations never go into effect, the case that Anthropic is making amounts to exceptional self-promotion. Anthropic’s ever-more-dire warnings serve as the best marketing money can buy (certainly better than their Super Bowl ad). Every company in the world wants to get its hands on models that are so powerful that they could crumble global security, create bioweapons, and build rogue superintelligence.

On one level, I understand the cynicism here: regulations often end up helping incumbents and there’s a long history of big companies lobbying to make the rules too complicated for new entrants. Facebook once ran the “regulate us, please” playbook in what felt like a calculated effort to push the responsibility for social media’s shortcomings onto legislators who were unequipped to resolve them. (The argument can have purchase even in pro-regulation leftwing and media circles because it fits nicely with the view that all companies are evil and trying to manipulate you.)

But there’s such a long record of Anthropic sincerity here. This is a company born out of safety concerns, incubated in the AI-paranoid halls of effective altruism. These are the people who were worried about the release of ChatGPT. They’ve perhaps been over-cautious and sometimes channel a little too much science fiction. But all evidence suggests that they’re on the level when they say they’re very concerned that foundation models could someday soon run amok if we don’t regulate them.

It’s endlessly amusing that Anthropic’s do-gooder, pro-regulation stance has actually been good business. While an OpenAI founder-funded super PAC is turning off the company’s employees, Anthropic is recruiting and retaining the best AI researchers by supporting regulation, welcoming criticism, and inviting scrutiny. It’s got to be infuriating for Sacks types that being so lib coded is apparently smart business in a world where winning over AI researchers is the biggest driver of their financial success. Praise be to capitalism.

On the substance of Anthropic’s safety position, it really feels like a matter of when not if.

Subscribe now

My personal view is informed by the self-driving car race where those researchers convinced themselves (and the world) that their technology would be ubiquitous any day now. Of course, it turned out that all the niggling final little details mattered quite a bit and it would take many years to figure out how to make it all work without human supervision. Today, with Waymo’s recent successes, a global society-altering self-driving car rollout seems more and more likely. So what if they got us all excited about a world historical idea a decade too soon?

Anthropic’s concerns here — about cyber security, bioweapons, government surveillance, autonomous weapons, and more — seem like things we should get ahead of. Worst case we’re prepared when these problems emerge a few years later than true-believer researchers thought they would. Anthropic is proposing limiting regulations to “models trained using more than 10²⁵ floating-point operations (FLOPs), developed by companies earning more than $500M in AI-related revenue or spending more than $1 billion on AI R&D.” So it’s not exactly mom-and-pop foundation model companies that would be caught in the dragnet of big regulation.

While much of the AI industry wants to figure out why Americans are hostile to AI, Amodei doesn’t believe Americans are getting it wrong (except maybe on data centers where their anger is “largely a symbol or outlet for broader economic anxieties about AI”).

He writes, “People are worried about AI because they correctly perceive that its risks are real.” He continues, “The key challenge is focusing this concern into constructive solutions and not allowing it to descend into formless anger and violence.”

Let’s not let cynicism about big business or smug hostility to regulations blind us to real leadership when it’s staring us in the face.

Newcomer Podcast

Ed Zitron Unfiltered on OpenAI, Anthropic & Why the Whole Thing Is a Con

IPO Watch

Investors Take a Leap of Faith Into Space

Of all the otherwise accepted business truths that Elon Musk challenges, one of the most fascinating is the Efficient Market Hypothesis.

The theory holds that stock prices reflect all available information — that the market, in aggregate, gets valuations roughly right. Skeptics argue markets are also moved by intangibles like momentum and vibes. With SpaceX beginning its trading today, the experiment is about to run again.

So far things have been going well. The company raised $75 billion, making it the largest IPO of all time. It’s a wide collection of investors in the round, which Bloomberg reported was four times oversubscribed. BlackRock ordered at least $5 billion worth of shares, per the Wall Street Journal — other buyers included sovereign wealth funds, family offices, and retail investors who reportedly requested more than $100 billion.

The longer term story will be trickier. Investors are, largely on faith, accepting a $1.8 trillion valuation for a company with revenue roughly the size of Cheerios-maker General Mills. The most outlandish case has been from Morgan Stanley, which told investors it projects the company will generate $3.4 trillion in revenue and $2.7 trillion in EBITDA by 2040, according to the Wall Street Journal. That’s an 80% margin — software economics, for a business whose plan involves launching its data centers into orbit on rockets.

Subscribe now

(We’ve written about how farfetched the orbital data center plan is, although the company believes it will have test units up in 2027, according to Reuters.)

It’s easy to laugh at these projections for being the pinnacle of Muskian absurdity — the Vibes Market Hypothesis. But there are much larger implications for how the market reacts to SpaceX.

If the expected growth in SpaceX’s future is coming from its AI business, then it will also serve as a preview of how OpenAI and Anthropic will fare with public investors. True, both of those have better AI models, less reputational overhang, and more revenue than xAI, but all of them are relying on an unprecedentedly huge market for AI overall to justify their valuations. The SpaceX bet isn’t just that xAI will catch up, but that what they’re all racing toward is as massive a shift in the economy as the tech industry has been predicting.

Fund manager and longtime Musk-bear Jim Chanos threw cold water on the whole idea, telling the New York Times that the SpaceX IPO feels like a “don’t look at the man behind the curtain” situation. It’s an ironic statement considering that the man behind the curtain is the reason people believe it’s possible at all.

Today’s public listing and the stock fluctuations to follow will make for an entertaining, if terrifying, watch — and a test not just of the efficient market hypothesis and the limits of Musk’s storytelling power, but of whether the public markets actually believe the AI future the entire industry is selling.

Eight Notable Deals

Poetic, Databricks, Ona, NEURA Robotics, Cyera, Standard Bots, ICEYE & Prometheus

• Enterprise AI automation startup Poetic raised $50 million in Series A funding led by Kleiner Perkins. Other investors include Founders Fund, First Harmonic, and OpenAI.

Read more

If you have a healthier relationship with social media than we do, you may be opening up your X account after a relaxing weekend offline to see a deluge of mud-slinging threads between some of Silicon Valley’s most prominent founders and venture capitalists.

The biggest conflicts have come between Matthew Prince, the CEO of Cloudflare, and Vinod Khosla, and Brendan Foody, the CEO of Mercor, criticizing Sequoia’s investment practices. But the last few days’ VC hostile posting extravaganza took a long and sometimes winding path of quote tweets to get to these two conflicts.

Here’s a rundown of how the various beefs all got set off. Grab some popcorn if you like.

June 2: Greg Isenberg Kicks Off the VC Horror Stories Thread

Read more