Elon Musk, Marc Andreessen, and many other tech leaders are long on Trump and that’s not a good position. Rest of World falls victim to a more inward-looking era. Anthropic S-1 brings plenty of takes and a searing critique from Roger McNamee. Five races on Tuesday are especially interesting for Silicon Valley, starting with Texas. Thrive bought ADP, the venerable home-security business. Kalshi raised a billion. China’s AI bubble starts to deflate. Big tech earnings paint a mixed picture as Google and Amazon shares plummet while Microsoft jumps. Phone operator Jio celebrates India’s biggest-ever IPO. Revolut founder Nik Storonsky has choice words for VC.

The Main Item

Putting Half-a-Billion on Red Leaves Musk Open to Retribution

Odds are looking pretty good that 2026 will go down as the most epic year in the history of venture capital, with Anthropic, SpaceX, and a few others (Cerebras, OpenRouter, Hugging Face, Cursor) all delivering legacy-making returns for numerous investors.

It will also be remembered for unprecedented investment in a risky, volatile sector that venture capitalists and tech companies have historically avoided: electoral politics. Those bets are looking far less promising. On top of Trump’s disastrous policies and dismal approval ratings, opposition to AI has suddenly emerged as a major campaign issue and the GOP is on the wrong side of it.

Elon Musk, of course, has made by far the biggest play, spending almost half-a-billion dollars supporting Republicans in the midterms thus far, according to the latest campaign finance reports. Almost $200 million of that has gone to Ken Paxton, the noxious Texas attorney general who’s in danger of becoming the first Republican to lose a statewide race there in decades.

Musk isn’t alone among tech figures now outpacing longtime Republican megadonors like Miriam Adelson and the Koch brothers. Ben Horowitz and Marc Andreessen have kicked in $92 million to GOP super PACs, Jeff Yass of Susquehanna $110 million, and Greg Brockman of OpenAI $50 million. (On the Democratic side, George Soros gave $102 million and Michael Bloomberg $84 million, with no other donors exceeding the $40 million mark.)

Then there are the other forms of donation: appearing at the White House and showering the President in gifts and praise, and kicking in for the infamous ballroom, as Jeff Bezos, Tim Cook, Jensen Huang, Mark Zuckerberg, and even Marc Benioff have all done.

It’s no mystery why business leaders typically try to play both sides of the aisle: politicians and party dominance can come and go, but corporations are forever. Going back to David Packard, who served as deputy secretary of defense under President Richard Nixon, tech executives have certainly had their partisan loyalties at times. It wasn’t that long ago that there was something of a revolving door between Silicon Valley and the Obama Administration.

Billionaires Rally Against Billionaire Tax

But there’s no precedent for Musk putting such a massive pile of chips on red, or for the overall scale of tech spending on politics. It goes well beyond the battle for Congress: in California, tech figures led by Sergey Brin are also investing huge sums to beat back a proposed one-time 5% wealth tax on billionaires. The campaign against it has spent an astonishing $330 million so far, with Brin contributing more than $100 million on his own and Silicon Valley heavyweights including Eric Schmidt, John Doerr, and Patrick Collison accounting for most of the rest.

If team blue wins the Senate, and especially if they win the Texas Senate race and/ or the governorship, Musk’s investment is likely to look very foolish. Democrats are already salivating about how they might take their revenge, floating ideas like new laws that apply only to people with a net worth of more than $500 billion. The nationalization of parts of SpaceX could even be on the table.

It’s still possible that Musk’s money will help prevent the worst for Republicans. It’s also possible that in the event of defeat, he’ll simply switch his massive underwriting to Democrats, many of whom would be sorely tempted to take the money rather than go to war with the world’s richest man. A16z likely made this sort of calculation, having shown a willingness to hedge its bets when Horowitz said he’d also donate to Kamala Harris shortly after the firm endorsed Trump.

Indeed, as Politico reports, tech firms are already scrambling to hire lawyers and lobbyists as they brace for inquiries from a Democratic House and possibly the Senate.

Musk appears to be in a race to merge Tesla and SpaceX and lock in as much federal money as he can while he still has a friend in the White House. That dynamic is in itself… unhealthy! It’s also a dangerous gamble for Musk and his backers. Investors who want to bet on politics would be better off sticking with Kalshi.

Subscribe now

Deglobalization Watch

Waning Interest in Tech’s Global Impact Dooms Rest of World

The news this week that Sophie Schmidt’s Rest of World, which explores the global impact of technology, is likely to shut down doesn’t come as a huge surprise. Launched in 2020, it was clearly a difficult financial proposition from the start, an expensive exercise in high-quality journalism about important topics that lacked a well-defined target audience or monetization strategy. It was a non-profit, to be sure, and Schmidt, the daughter of the former Google chairman, personally committed up to $60 million over a decade for what she called her “life’s project.” But philanthropy is rarely enough by itself.

We don’t have any reason to believe that Schmidt lost interest or didn’t do everything she could — a common hazard when wealthy people underwrite journalism — but cratering traffic and lack of a viable growth path will ultimately do in a non-profit just as they would a business.

Yet the expected demise of the publication (Schmidt said it will close in January absent a white knight) is about much more than AI’s impact on Google search and the ever-present challenges of new media business models. Rest of World debuted at a moment when the immense and lightning-fast changes that smartphones and social media were bringing to all corners of the planet had become an important part of the technology conversation.

Your Newcomer scribe was running global technology coverage for Reuters at the time, based in Singapore, and we went hard after stories about, say, how Facebook was transforming the Vietnamese economy, or stoking ethnic violence in Myanmar, or employing armies of Indian contractors. Rest of World’s highlights ranged from a deep-dive on China as the world’s shopping cart to long-term beat coverage of the plight of gig workers around the globe. They were a very welcome addition to the mix.

Now, though, in the era of deglobalization and Trumpian hostility to journalism, idealistic non-profits, and anything that smacks of “rest of world,” there just isn’t the same interest in those kinds of stories. There’s no point in doing expensive, worthy reporting that isn’t able to build a meaningful audience. Yet it’s hard not to be a bit sad that a richly funded international reporting operation like Rest of World, doing ambitious work on major global stories, just couldn’t find a way.

Chatter

Six Takes on Anthropic’s IPO

The Anthropic S-1 has been a god-send for tech and finance take-havers this week. Here are a few thoughtful or amusing ones we’ve seen so far.

1. NYU Stern professor Aswath Damodaran broke through on Wednesday with his case that Anthropic should be valued at $1.475 trillion, not in its $1.5- $2 trillion current standing.

2. Roger McNamee calls the IPO “a very dangerous thing for investors,” and regards the S-1 as “speculative fiction.” (See his full anti-AI screed here.)

3. Ed Elson did a breakdown on which risks on the S-1 actually matter.

4. Matt Levine delivered a sharp observation on the S-1's blend of AI doomerism and banker verbiage.

   

5. A note about the unprecedented nature of the potential wealth creation from a semi-anonymous poster.

6. And finally, a good meme:

Midterm Elections

Five Races with High Stakes for Tech

1. Texas Governor and Senate Races: With Governor Greg Abbott fending off a strong challenge from Gina Hinojosa and Ken Paxton in a dead heat with James Talarico, Elon Musk has swooped in with large amounts of last-minute spending. He’s put in an additional $5 million into the Governor’s race and nearly $168 million into the Senate race. Talarico and Hinojosa have both leaned into anti-billionaire rhetoric while Abbott plays defense on data centers.

2. California Billionaire Tax: The one-time 5% wealth tax on California’s billionaires to cover healthcare deficits has led around 70 current and former billionaire residents to spend $612 million in statewide races so far, per Politico, with around half of that in direct opposition to the initiative. The latest polls show support for the tax falling but the outcome is far from predictable.

3. Washington State Millionaire Tax Repeal: The legislature passed a 9.9% tax on household income exceeding $1 million per year this March, scheduled to go into effect in 2028. Opponents rallied to get Ballot Initiative 645, which would repeal the tax, in front of voters.

4. Ohio Senate Race: The crypto industry PAC Fairshake has poured over $55 million into the race in an effort for defeat Democratic challenger and former Senator Sherrod Brown, a longtime critic of the industry. Brown has been attacking his opponent, incumbent Senator Jon Husted, for being too supportive of data center construction.

5. Florida Governor’s Race: Representative Byron Donalds has been running possibly the most pro-data center campaign in the country as Americans broadly sour on their construction. His opponent, David Jolly, has branded him “Data Center Donalds.”

Seven Notable Deals

Kalshi, Thrive Holdings + ADT, Re:Build, Bedrock Robotics, Serval, General Compute, Disruptive

Prediction markets platform Kalshi reaches a $40 billion valuation. Plus, hardware startups have VCs shelling out for manufacturing, robotics, and compute.

Read more

Mettle Capital, a new VC fund started by three former partners at Sequoia India after a contentious exit, has raised $350 million to invest in Indian tech startups, according to two sources with direct knowledge.

Mettle’s fundraise is an indication that LPs are not yet writing off India as an investment destination or technology hub, even though the investors and founders in the region are undergoing a moment of soul-searching, with many eyeing Silicon Valley as US AI startups widen the gap with the rest of the world.

Read more

Space data centers, once confined to sci-fi novels and ambitious pitch decks, are beginning to get real.

A SpaceX rocket earlier this month carried an experimental Google satellite that will make simple requests to Gemini while researchers test how its four TPUs hold up in space. The same launch carried the Reason-1 satellite from startup Cowboy Space, which is experimenting with beaming solar power back to Earth with a laser.

Cowboy Space, led by Robinhood cofounder Baiju Bhatt and backed by Jan Hammer at Index Ventures, among others, this week unveiled an even bigger swing: its own purpose-built data center rocket that can carry five racks of NVIDIA Vera Rubin GPUs into orbit. The startup also plans to use this rocket to compete with SpaceX in selling launch services to others under the name “Big Iron Express.”

Bhatt told Newcomer that his company’s “north star” is to make space compute cost-competitive with terrestrial compute, and that it’s vertically integrating to the extreme to achieve that end. Another well-funded player is Starcloud, backed by Benchmark and currently valued at $2.3 billion.

Read more

Anthropic’s S-1 arrived at a particularly fraught moment in the tech business, with anti-AI sentiment raging across the country and a highly consequential election about to take place. The numbers are extraordinary — and will likely polarize the debates even further.

We’ve pulled a handful of highlights from the filing, some of which were foreshadowed:

Read more

Anthropic’s pre-IPO fundraising forced VCs to get bigger and the stakes could not be higher for its Wall Street debut. OpenAI is in talks for a $30 billion fundraise as a bridge to an IPO. SemiAnalysis was challenged on its data center build estimates, prompting a lively debate. A ranking of AI VCs from Stanford’s Ilya Strebulaev and Ohio State’s Blake Jackson puts a16z on top. AI lab Flapping Airplanes raises a $1 billion Series A and AI legal startup Legora raises a $550 million Series E. Bison Capital and Inspired Ventures raise new early-stage funds. The merger of Tesla and SpaceX inches closer. A PHP security freakout sweeps the Web. Meta beats estimates but worries investors. Polymarket’s Coplan hits a rough patch.

The Main Item

AI Mania Pushes the VC Megafund Model to Its Limits

Does the big, well-known VC firm have enough money to do the deal?

That’s not a question that came up in Silicon Valley — until the AI revolution.

As we detailed yesterday in our tale about Anthropic’s fundraising, VCs including Menlo Ventures, Lightspeed, and ICONIQ effectively had to super-size themselves to be able to write the huge checks that Anthropic was seeking. They raised new funds, urged LPs to co-invest alongside them, and in some cases marketed special purpose vehicles widely to come up with the cash.

Anthropic’s S-1, expected any day, will show a cap table dominated by corporate investors and late-stage megafunds. A handful of smaller VCs hold more modest — albeit extraordinarily lucrative — stakes. We’ve written frequently about the growing power of a handful of Silicon Valley’s biggest firms, led by a16z and Sequoia, and Anthropic is something of a case study in the underlying forces driving that trend. Capital requirements have skyrocketed, alongside LP interest in going big on tech and AI.

OpenAI is poised to add another exclamation point on all this with a $30 billion fundraise at a $1.25 trillion valuation, according to a CNBC report Thursday.

Yet even as the rise of the VC megafunds reaches something of an apotheosis, there are reasons to think this capital concentration may have some limits.

Even assuming the Anthropic IPO goes well — if it doesn’t, all bets are off — there is a very large herd of unicorns out there that still has nowhere to go. The M&A business is reeling from high interest rates. Equity markets are receptive only to the best IPOs, and the AI-stock rout that hit late this week will close the window further.

A lot of LPs are sitting on big paper gains but haven’t gotten much cash back over the past decade, and thus don’t have the winnings to re-up into larger and larger funds indefinitely.

Subscribe now

One potential weakness of the megafund model, in fact, is that it’s dependent on a fairly small pool of very large institutional investors, and sooner or later they are going to expect the kind of returns that only a small slice of venture funds have delivered in recent years. All the big firms were able to raise large new funds over the past year to go after AI, but it’s unclear how much more money there is behind that.

On the other side of the industry’s barbell, there is a growing cadre of investors who made a lot of money in earlier tech waves — or even in other fields — and can write checks much more sizeable than traditional seed or early-stage investors. Sound Ventures, along with Factorial Funds, started by a former Cruise Automation engineer, are poised to be major winners in Anthropic after putting $35 million and $50 million, respectively, into the Series C. They got the chance when the giants of Sand Hill Road took a pass.

That dynamic is evident with SpaceX too, where low-profile players like Valor Equity Partners scored huge windfalls.

It’s certainly true that the fracturing of the traditional venture model and the rise of the megafunds has made life tough for many mid-tier firms and emerging managers, who are often locked out of the best deals (or can get in only at inflated prices in the multiple-valuation rounds that are becoming increasingly common).

Yet the sheer scale of the AI investing opportunity means there’s still room for shrewd new players with fresh ideas. Another way to think about the megafunds is that they arise from the VC industry’s famously herd-like behavior: everyone wants to be in the deal everyone else is in.

The biggest opportunities, then, may lie in going against the grain. That game remains wide open.

CVAI

Spots Are Filling Fast for This Year’s Cerebral Valley AI Summit. Apply Now to Secure Your Seat.

We’re three weeks out from the Cerebral Valley AI Summit, hosted in the heart of San Francisco on Nov. 12.

We’ve secured some of the top names in Silicon Valley to speak on our main stage, including Ali Ghodsi, Nikesh Arora, Igor Babuschkin, Lin Qiao, and Brendan Foody.

More speakers will be added to our lineup, so stay tuned.

Apply now to get a seat in the room.

Apply to Attend

Tweet Storm

SemiAnalysis Rebuked Over Estimates of Data Center Capacity

Everyone is rushing to build data center capacity, but how much will actually come online is still uncertain — and some executives weren’t convinced by recent SemiAnalysis predictions.

SemiAnalysis researchers forecast 50GW of data center capacity coming online in 2027, based on signed contracts and visible construction already underway.

Altimeter’s Brad Gerstner had disputed the estimate on the All-In podcast weeks ago, calling it “too aggressive” and asserting that capacity could come in around 25GW. SemiAnalysis followed up this data with satellite images of the aforementioned construction sites.

IREN CEO Daniel Roberts then spelled out how IREN pays for its GPUs, where customers prepay for about half, and then debt covers most of the rest. That means that the revenue is coming in to support the buildout — but the issue is whether or not the chips and data centers can be physically up and running by 2027.

And for what it’s worth, Gerstner found its latest position more compelling:

Cerebral Valley Show

Did Elizabeth Holmes Just Fool Everyone Again?

One Big Chart

A16z Leads Latest Ranking of Top AI VCs

A16z claimed the top spot on the respected ranking of the 50 AI-focused VCs put together by Stanford GSB professor Ilya Strebulaev and Ohio State University professor Blake Jackson. The list is built off data points including valuations, growth rates, dilution, capital distributions, and value-add from the firm.

Lightspeed, Menlo Ventures, Sequoia, and ICONIQ Growth rounded out the top five. While the megafirms also have significant investments outside of AI, firms like Menlo Ventures and Dragoneer scored very well thanks to their heavy concentration in AI and especially in major deals like Anthropic.

Nine Notable Deals

Flapping Airplanes, Legora, ASI, Pigment, Arda, Otter, Warhound + Inspired Capital, Bison Ventures

Legora reached a valuation of more than $8 billion while Pigment hit $2 billion in a busy week for outsized European funding rounds. Buzzy neolab Flapping Airplanes hit a $6 billion valuation. Plus, ex-OpenAI executive Bob McGrew secured backing for his manufacturing AI startup.

• AI neolab Flapping Airplanes raised $1 billion in Series A funding co-led by Lux Capital and Index Ventures. Other investors include Sequoia Capital, GV, Kleiner Perkins, Valor Equity Partners, Atreides, A*, and Nvidia.

• Legal AI startup Legora raised $550 million in Series E funding led by Atreides Management. Other investors include Baron Capital, Capital Group, Fidelity, Glade Brook Capital Partners, JP Morgan, Accel, Benchmark, General Catalyst, ICONIQ, and Salesforce Ventures.

• ASI, a startup co-founded by Bob McGrew building an AI system for the military and airlines to manage their assets, raised $325 million in Series C funding led by Spark Capital. Other investors include a16z, Craft Ventures, Meritech Capital, Energy Impact Partners, and Citadel Securities.

• AI enterprise-planning startup Pigment raised $250 million in Series E funding led by the Scaleup Europe Fund. Other investors include World Innovation Lab, Greenoaks, ICONIQ, IVP, and Meritech, with an additional planned investment from Nvidia.

• Manufacturing AI startup Arda raised $232 million in Series A funding co-led by Altimeter Capital, Khosla Ventures, and Accel. Other investors include Felicis, XYZ Venture Capital, and 8VC.

• Otter, the AI note-taker, raised $320 million from General Catalyst’s Customer Value Fund.

• Warhound, a startup co-founded by Colossal Bioscience’s Ben Lamm and Anduril’s Palmer Luckey to genetically engineer warfighting dogs, raised $200 million in seed funding from investors like In-Q-Tel, Sequoia Capital, ARK Invest, Franklin Templeton, 8VC, Lux Capital, USIT, Bob Nelsen, Breyer Capital, CAZ Investments, Neogenesis, and Builders VC

• Inspired Capital, the early-stage fund founded by Alexa von Tobel and Penny Pritzker, raised $450 million for its Fund IV. Past deals from the firm include Suno, Rho, Deeptune, and MosaicML.

• Bison Ventures raised $160 million for its second early-stage fund focused on deals for the “physical world.” Its past investments include Thinking Machines Lab, AI infra startup Zyphra, and robotics company Cobot.

Stories We’re Reading

SpaceX & Tesla Begin Their Dance, Companies Rush to Patch PHP, Meta Beats Estimates & Preps New Shares, Polymarket’s Copland Catches Heat

• SpaceX and Tesla reported earnings on back-to-back days this week and beat Wall Street’s expectations, though Tesla’s shares dipped as the firm’s lower-than-expected capex spooked investors looking for more aggressive investment in AI and robotics. But the big action in the Muskiverse was trying to sort out what the inevitable SpaceX-Tesla merger will mean after Musk publicly endorsed the idea and said both companies had set up special committees to consider a tie-up. In some ways, a merger would simply neaten up the boundaries of the Musk empire: SpaceX already resembles a sci-fi Berkshire Hathaway, and The Wall Street Journal reports that the two companies have been bedfellows for decades, sharing resources and personnel. That said, Musk is technically conflicted out of discussions as the CEO of both firms, and publicly weighing in has already thrown internal talks at both companies into chaos.

• Serious issues in the underlying structure of PHP, the scripting language used by more than 70% of websites globally, led to an international security freakout this week. On Wednesday, the PHP Foundation — the organization that manages and supports the language — published a release revealing the discovery of two serious flaws, designated as CVE-2026-109618 and CVE-2026-109619. These holes impact all PHP systems dating back to a version released in 2012, and could allow hackers to gain remote control of any server running PHP that hasn’t installed the fix. The security organizations Calif and Wiz independently discovered the flaws using AI models and reported them to PHP. Shortly after the announcement, cloud providers scrambled to secure their networks. Cloudflare CEO Matthew Prince posted on X Thursday morning that exploitation attempts on Cloudflare’s network rose from thousands in the 10 hours before the disclosure to millions afterward, and Cloudflare said in a blog post that it deployed firewall protections for its customers hours before the disclosure. Research published by Amazon Web Services on Wednesday indicated that at least one hacker had attempted to exploit the bugs on October 17, four days before the vulnerabilities were made public. The PHP Foundation is urging all users who could be affected to update with these patches as soon as possible to avoid hacks. Several government websites running on PHP, including those for Kenya, Finland, France, Germany, and the Netherlands, went preemptively offline Wednesday — and some remained down Thursday morning.

• Meta posted $65 billion in Q3 revenues to beat Wall Street’s estimate of $63 billion. Strong ad sales drove the numbers, and investors were initially enthusiastic about the company’s launch of its personal AI assistant Muse and subsequent Muse Wonder model. But after the stock initially jumped, sentiment soured as capex spending rose to $35 billion, free cash flow turned negative, operating margins plummeted, and almost $10 billion was charged for a teen mental health settlement. Meta also announced a new class of non-voting Class C shares that will let the company raise capital for AI spending without diluting CEO Mark Zuckerberg’s voting control. As of Thursday’s close, the stock had fallen 7%.

• Polymarket CEO Shayne Coplan has apparently had a rocky ride as Peter Thiel’s Founders Fund tries to mentor the enfant terrible founder. Bloomberg reported that Coplan used company money to pay for living expenses and private-jet flights before Polymarket had generated any revenue — money the investors insisted he repay before they would cut another check. Coplan and his allies pushed back hard against Bloomberg’s characterization of the situation, insisting he proactively accounted for and repaid every dollar as the business grew, and that his total expenses were less than a typical software engineer’s salary. Bloomberg’s piece comes at a poor time for Polymarket, as the former poster child for prediction markets continues to hemorrhage market share to archrival Kalshi. But the story also shows the tightrope VCs must walk to keep startups on the straight and narrow without looking unfriendly to founders.

If all goes to plan, Dario Amodei and his Anthropic co-founders will ring the NASDAQ bell next month to mark the company’s IPO. Their celebration will no doubt be greeted with the traditional high-fiving and back-patting from venture capitalists.

But the story of Anthropic’s rise from a small group of OpenAI defectors to an industry-leading juggernaut, poised for what will likely be the largest IPO of all time, is a far messier narrative than we normally get in Silicon Valley.

For years, VC firms passed on backing Anthropic, fearing its unusual structure and fixation on safety over commercialization made it a bad bet. Today no venture firm owns more than 2% of the company, according to investor documents and half-a-dozen sources with knowledge of the company’s multiple funding rounds. Keeping VCs below that line to ensure no one investor had outsized control was a goal Anthropic communicated to investors, one of these people said.

That’s a highly unusual circumstance for a tech firm on the brink of going public — even the widely held SpaceX had some institutional investors owning more than 3% of the company when it debuted.

Anthropic’s journey is instead marked by prestigious angel investors — many, though not all, associated with effective altruism — funding its early growth. The infamous Sam Bankman-Fried of the late FTX led the Series B, resulting in a big pot of shares being distributed to others after the crypto brokerage collapsed and Bankman-Fried was jailed. In subsequent rounds, Anthropic executives and its investors battled over what type of capital was or wasn’t allowed.

That isn’t to say there aren’t distinct VC winners in the Anthropic IPO. Menlo Ventures, Lightspeed, and Iconiq each led middle-stage funding rounds— and kept following on — building stakes that look to be worth tens of billions. Meanwhile, names like Spark Capital, Sound Ventures, and Factorial Funds got a piece of the company early and secured positions that are today also worth billions.

Anthropic’s massive appetite for capital has ended up having a gravity-warping effect on much of the VC industry, accelerating the trend towards megafunds as it all but remade the firms that backed it; investors accustomed to writing checks that rarely topped $100 million suddenly found themselves making nine- and ten-figure commitments to keep pace with the company’s growth. For nearly all of the venture players, Anthropic became the largest investment in their firms’ history.

Early Rounds a Tough Sell

Anthropic’s earliest rounds were a tough sell to the Silicon Valley establishment. Anjney Midha, an early backer and friend of the co-founders like Tom Brown, has said publicly that he helped pitch the company to 22 venture firms and that 21 passed. (He has never said which one didn’t.) So Amodei and his co-founders turned to wealthy individuals, many of them in or adjacent to the EA movement, including Skype co-founder Jaan Tallinn, who led the Series A, and Facebook co-founder Dustin Moskovitz. Eric Schmidt, not explicitly an EA acolyte, also invested in the A round and is poised to be one of the IPO’s biggest winners.

(Tallinn remembers it differently. He said in a recent podcast interview that before investing, he asked Amodei to line up offers from venture firms. He wanted to be sure Anthropic would get funded with or without him, so his money would simply take a VC’s place rather than bring another AI lab into existence. The logic is very EA, and sure enough Bankman-Fried led the company’s B round.)

Prominent tech figures who were friendly with the founders tried to help out in those early days, but also struck out selling it to Sand Hill Road.

The breakthrough finally came with Anthropic’s Series C in the spring of 2023, when Spark Capital partner Yasmin Razavi agreed to take the leap. Spark, until then best known for backing consumer companies like Twitter, Oculus and Affirm, put in $75 million to lead the $450 million round, which valued Anthropic at about $4.6 billion before the new money. Google joined, along with under-the-radar firms including Factorial Funds, which put in about $50 million, and Sound Ventures, which invested $35 million. Factorial is an AI-focused shop started by former Cruise Automation engineer Sol Bier. Sound was co-founded by Ashton Kutcher and talent manager Guy Oseary. Menlo Ventures put in around $10 million and Zoom and Salesforce participated through their corporate venture arms.

Spark kept investing in later rounds, and its stake was diluted significantly as the price climbed. But Razavi scored a trophy that no other VC could claim by joining the company’s board. Under the terms of the C round, Spark’s right to name Razavi to the board was set to expire once Anthropic topped $1 billion in annual revenue, at which point its institutional preferred shareholders would choose the director instead, according to a document reviewed by Newcomer. Anthropic blew past that mark, but Razavi remains the only VC on the board.

After the company got through the challenging C round, which was still thin on brand-name VCs, the Silicon Valley purse strings finally loosened a bit. In the fall of 2023 Amazon announced a multi-billion-dollar commitment; the next year the startup announced it closed a $750 million series D led by Menlo partner Matt Murphy that valued the company at $18.4 billion.

A Dust-Up Over SPVs

Menlo’s commitment of more than $500 million, the largest in the firm’s history, was mostly raised through a special purpose vehicle (SPV). Accounts of how that came together differ. One source described a sprawling effort where the firm sent generic emails about the vehicle to a diffuse set of investors, including people who already had a stake in the company. One such email reviewed by Newcomer included links to a private data room and a webinar the firm was hosting.

Another source said the SPV was run in close coordination with Anthropic, which had reviewed the material that was shared in the webinar.

During the funding process, Murphy also brokered introductions between Anthropic executives and other investors including Byron Deeter at Bessemer, said sources familiar with the matter. Lightspeed invested $100 million in the D round and both it and Bessemer would go on to play larger roles in subsequent rounds.

The matter of SPVs would become a point of contention across multiple rounds, coming to a head after the series D when Anthropic brought on Krishna Rao as its CFO. The former Airbnb executive wanted to cut down on the proliferation of SPVs to help the company better understand the provenance of all the dollars flowing in. Among other concerns, he and Amodei wanted to ensure that no money from Middle Eastern, Chinese or Russian sources ended up backing the Anthropic, according to a person familiar with the matter. (The Gulf ban didn’t last: in July 2025, Amodei told staff in a leaked memo that the company would seek investment from the UAE and Qatar to stay competitive.)

Rao ultimately told VCs that if a firm was given an allocation to invest in Anthropic, they had to be good for the money themselves rather than marketing the stake to a wide collection of suitors, according to one investor.

Yet Rao’s message that SPVs were on their way out turned out to be less than categorical. Anthropic’s series E in spring 2025, which valued the company at $61.5 billion was led by Lightspeed, which put in $1 billion. The commitment was spread across multiple funds but filled out partially with additional dollars from its top LPs, who were allowed to co-invest in the round, according to documents reviewed by Newcomer.

Menlo participated in the E round as well. Rao permitted it to syndicate the round, but only to its top five LPs, according to a person familiar with the matter.

The SPV issue came around again in the fall of 2025 when Anthropic held its series F at $183 billion. Iconiq, which led the round, pitched its wealthy network of investors as part of its $3 billion allocation, according to two people familiar with the matter.

Again there are differing accounts on the nature of the pitch. One source says it was strictly a co-investment opportunity for Iconiq’s LPs and others in their network, with the full sign-off from Anthropic. Another person who reviewed the Iconiq pitch said there were fees attached to the vehicle, making it different from a typical no-fee LP co-investment.

Sequoia Elbows In, Finally

Read more

We’re a little over three weeks away from the Cerebral Valley AI Summit, which we’re hosting in the heart of San Francisco on Nov. 12.

We’ve secured some of the top names in Silicon Valley to speak on our main stage, including Igor Babuschkin, Nikesh Arora, and Ali Ghodsi.

Now, we’re thrilled to announce that Brendan Foody, the CEO of Mercor, will join the lineup for our flagship event.

Additionally, we are excited to share details on the terrific group of founders, investors, and AI leaders who are leading our discussion groups. These small, curated conversations have become a highlight of our summits.

Discussion group leaders will include:

• Joon Sung Park, founder & CEO of Simile

• Emmett Shear, founder & CEO of Softmax

• Rodrigo Liang, founder & CEO of SambaNova

• Michael Grinich, founder & CEO of WorkOS

• Keith Peiris, founder & CEO of Lightfield

• Zoë Weil, founder & CEO of Sequen AI

• Jai Das, founder, president & partner of Sapphire Ventures

• Eugenia Kuyda, founder & CEO of Wabi

• Wesley Chan, founder & managing partner of FPV

• Dmytro Dzhulgakov, co-founder & CTO of Fireworks

• Harjot Gill, founder & CEO of CodeRabbit

• Hooman Reza Nezhad, founder & CEO of Solcoa

• Andy Vitus, partner at Scale Venture Partners

• Kambiz Aghili, chief product officer of IREN

• James Detweiler, general partner at Felicis

• Laura Modiano, head of founder experience, global at OpenAI

• Jay Jackson, senior vice president of Global AI and Strategic Customer Operations at Oracle

We’d like to note one change to our agenda: With Anthropic bearing down on an IPO that’s likely to happen around the time of CVAI, we weren’t surprised that CEO Dario Amodei had to pull out of his speaking slot. We’re sad to miss him.

We will continue to add additional speakers to our lineup in the coming weeks, so stay tuned.

The Summit is filling up fast. Apply now to secure your seat!

Tickets are $1,299. Founder tickets are $199. Professional Services and Business Development tickets are $2,999.

Apply to Attend

A hearty thanks to all of our sponsors: Fireworks, Oracle, IREN, Index Ventures, Exa, Nebius, WorkOS, ElevenLabs, Higgsfield, Felicis, SambaNova, Sapphire Ventures, Bessemer Venture Partners, QuantumBlack, Liquid AI, and Weekend.

We’re excited to see you in San Francisco in November!

MVP Ventures, an investor in three of Palmer Luckey’s startups, is in talks to raise $100 million to increase its stake in its best-performing companies, according to a source with knowledge and a presentation seen by Newcomer.

MVP, a five-year-old firm that has raised two venture funds, isn’t well-known in Silion Valley, but it holds early — albeit small — positions in defense technology firm Anduril, financial services firm Erebor, and gaming venture ModRetro, all of which were set up by Luckey. Other bets include AI-training startup Turing, voice AI firm Wispr Flow, and rocket-launch startup Stoke Space.

Read more

We just published my conversation with Mercury CEO Immad Akhund on the Newcomer Podcast. You can watch it on YouTube and Spotify, or listen along on Apple or wherever you get your podcasts.

Akhund shared some exclusive milestones for his fintech business: Mercury just hit $1 billion in annualized revenue, up around 50% year over year. The company now has more than 450,000 customers. It’s also launching a payroll service.

But it wasn’t just a brag-fest: We talked about AI agents spending money, how he thinks we are in a bubbly moment for startups, and the cooldown in fintech startups after the pandemic frenzy.

Give it a listen.

I wanted to use this as a moment to remind Newcomer newsletter readers about the great conversations we’re having on the Newcomer Podcast and the fiery debates on my Ringer-style AI chat show, the Cerebral Valley Show, that I host with my friends Max Child and James Wilsterman.

The Newcomer Podcast

We’ve featured guests like Amanda Askell, Reid Hoffman, Vinod Khosla, Keith Rabois, Tony Fadell, Ed Zitron, Ben Smith, Taylor Lorenz, and Katie Drummond.

I’m looking for guests that aren’t afraid to chop it up, debate, and face probing questions from an independent thinker who believes in the tech industry and capitalism.

If you’re interested in coming on the Newcomer podcast, reach out to my producer Dave Heuff. If you’re interested in sponsoring the podcast, email AJ Frucci.

The Cerebral Valley Show

The hidden gem of the Newcomer media empire is the fortnightly Cerebral Valley Show with my Cerebral Valley AI Summit co-hosts and Weekend co-founders Max Child and James Wilsterman.

There are a lot of podcasts for AI haters and AI zealots. We’re somewhere in between — enthusiasts who poke and prod at the industry. We’re The Ringer for tech podcasts, somewhere you can get an entertaining gloss on the insider conversation in the AI industry even if you don’t sit on X all day.

You can listen on YouTube, Spotify and Apple podcasts.

Rising interest rates are hitting neoclouds hard and threatening AI momentum. Valar Atomics feuds with its own investor, Masha Bucher and Day One. Latest Ramp data shows AI token spend leveling out. Bumble taken private by Ithaca Holdings, a buyout vehicle for top VCs. Ramp and Modal raise big new rounds. Bending Spoons brings in Berkshire Hathaway money, lays off Airtable team. Meta has a big week with Muse models. Ron Conway plays for Team Blue in the midterms. Bloomberg finds a sad story in Marc Andreessen’s hometown.

The Main Item

CoreWeave & the AI Infrastructure Build-Out Face a Moment of Truth

CoreWeave is an archetypal company of the AI boom, for better and for worse. Now the markets are testing its mettle.

Launched by a group of finance and bitcoin guys, CoreWeave saw early that there’d be a huge need for AI compute and built a highly leveraged vehicle that deployed massive amounts of capital to serve that need.

The risks were obvious from the get-go and CoreWeave’s IPO in the spring of 2025, though rocky out of the gate, had investors quickly send the shares soaring and set the whole neocloud category on fire. Companies including Nebius, Nscale, and Firmus ramped up debt-financed projects backed by Nvidia (both its chips and its equity in some cases) and private credit lenders such as Apollo and Blackstone.

Now the grim reaper of rising interest rates, alongside political backlash and general market jitters over the AI trade, is coming for the neoclouds. Anthropic financials that were released this week didn’t provide much comfort.

The betting here is that there’s carnage to come.

Firmus Grid, an Australian firm, may prove an early indicator. A former Bitcoin miner with close ties to Nvidia and also backed by Coatue, Jane Street, and Blackstone, it tried to go public last week at a valuation of $30 billion, though it only had $59 million in revenue in FY 2026 and has only delivered 46MW of its 912MW data center pipeline. Oh and its CEO had done time for insider trading. That’s circa-2000 bubble vibes for sure.

The Firmus IPO was pulled after investors balked at the valuation, and what looked like arrogant overreach, and they ultimately weren’t willing to back it even at a lower price. The company insists it still plans to go public in New York next year and in the meantime will raise private funds, though there’s been no announcement yet.

Nscale, which filed to go public last month, has a similar profile: a $35 billion valuation based on contracts with Microsoft and Anthropic, and a big ratio of promised megawatts to actual capacity online. That deal is already looking shaky.

CoreWeave, for its part, is battling a bond market that has abruptly become far more skeptical of the AI buildout; its debt was trading at distress levels mid-week before recovering somewhat, but the bonds still pay a 15% yield as of Thursday. Yet the company needs to be able to borrow at investment-grade rates — i.e. more like 6% — for its math to work.

Equity investors, still drawn to the upside of the AI trade, are hanging in there; CoreWeave shares have fallen 30% over recent weeks but still sit 60% above its IPO offering price 18 months ago. Yet the company is clearly in the hot seat: CEO Mike Intrator was forced to plead “irrational markets” in explaining the bond selloff, which was amusing given that the whole category is arguably the product of irrational exuberance.

Nvidia’s Jensen Huang — deeply invested, both literally and figuratively, in the success of CoreWeave and Firmus — offered a vote of confidence Thursday, insisting that CoreWeave’s cost of capital was actually declining and all would be fine. But a Bloomberg story late in the day showing that CoreWeave founders have sold $3.3 billion in stock since the IPO isn’t likely to improve the mood of shareholders.

Nor did the Anthropic revenues released this week provide a clean narrative for data center bulls.

In a meeting for potential public market investors, the company disclosed that it had hit nearly $17 billion in revenue for the third quarter. That’s up 46% from the previous quarter, which is impressive, but well below the explosive 100% growth it showed between the first and second quarters. Around 5% to 7% of Anthropic’s revenue came from “distillation,” a process where rival AI model makers train their models on the outputs of more advanced ones, per the Financial Times. That makes the $2 trillion valuation Anthropic was targeting at its IPO feel like a reach; on Thursday, Reuters reported that some investors were discussing a valuation of around $1.5 trillion.

The company is projecting it will spend upwards of $500 billion on data center capacity over the next decade. And cloud providers like Coreweave and the rest are taking on expensive debt in anticipation of Anthropic being able to make good on that projection.

It's a risky bet even if you believe in AI’s transformative power.

Pro Rata Rights

Valar Atomics’ Battle with Day One Spotlights the Valley’s Unwritten Rules

A key principle of Silicon Valley today is being “founder friendly,” and there are a lot of tacit understandings about what that means. Isaiah Taylor, founder and CEO of Valar Atomics, is trying to enforce the unwritten rules with a lawsuit against his own seed investor, Day One Ventures.

When Sequoia invested in the nuclear-reactor maker earlier this year, other investors waived pro-rata rights and “spring back” rights that entitled them to invest more and would have diluted Taylor. But Masha Bucher, Day One’s founder, refused, and after several testy exchanges, Taylor says Bucher called on Quinn Emanuel to send legal threats. You can read Taylor’s version here and Bucher’s response here.

Publicly, most people backed Taylor, and some termed the situation “tough”; Silicon Valley founders are untouchable at the moment. But investors we spoke to were privately more sympathetic to Day One’s predicament.

The tussle also exposes the Valley’s seesawing founder-investor dynamic. Taylor raised money from Bucher — not a top-tier VC, with ties to suspect Russian investors — when he did not have better options, but after a billion dollars in funding, he felt empowered to rearrange his cap table.

Bucher’s side believes she alone was told to give up future pro-rata rights, while other investors were not, according to a source with knowledge. Taylor also gave other investors some allocation, despite their waiving, the person said. These details have not been previously reported. Valar did not respond to a request for comment.

“Masha wasn’t some elite VC, but god, what founder will now take money from her after this? How will it look? Can they trust her? And Day One’s name in a press release will now have a wildly different association,” one VC told us.

Investors like to say that if like to say if things work as they should, none of the details in complex deal agreements ultimately matter. If Valar becomes a “multi-trillion-dollar company,” as Taylor says it can be, Bucher will make spectacular amounts of money, although the pro-rata might have made the difference between, say, a 100x and a 150x.

While the episode in some ways underscores the importance of the fine print, sources told us it was unlikely to spur legal tweaks or challenges at other startups because if VCs over-negotiate, founders will go elsewhere.

Newcomer Podcast

Alex Kantrowitz on OpenAI, Nvidia & What Could Derail the AI Boom

Two Big Charts

Enterprise AI Token Spending Levels Out, Ramp Data Shows

The growth in new AI customers has dropped significantly, according to fresh business spending data from Ramp.

OpenAI managed to narrow Anthropic’s gain and xAI saw a slight uptick due to the popularity of GrokBot, but it wasn’t enough to reverse the trend that began this summer.

Meanwhile, AI token spending growth has been basically flat for two months in a row, likely due to price declines at the frontier and the models becoming more efficient overall, writes Ramp chief economist Ara Kharazian.

Read more

For 20 years, Anil Advani, a Palo Alto-based lawyer, has worked on the niche project of helping Indian startups incorporate in the United States. Suddenly, though, Advani’s field is a lot less niche: Every week, a dozen Indian founders seek his counsel on moving to the US, more than triple the rate from two years ago. They’re coming for the AI opportunity and all they can learn about the US market.

Hot on their heels are Indian VCs. Multiple firms have opened offices in the Bay Area and hired executives just in the last few months. Over the past two years, Indian VCs have made over a hundred investments in US-based companies, a Newcomer calculation shows — a number that would have been unthinkable a few years ago.

Read more

OpenAI’s annual recurring revenue is $20 billion less than news reports had indicated, spotlighting fuzzy math issues. Open-source models from the US and Europe launched to media fanfare this week and vowed to take on China, but national origin doesn’t matter much to the founders using them. Startup funding so far this year shatters 2021 records but slows in latest quarter. Singapore startup Manus raises after its forced split from Meta. Google’s AI push in the classroom proves a nightmare. Nvidia CEO Jensen Huang’s son-in-law takes on a bigger role inside the company. USV among the VC firms raising new funds. A new lawsuit over Groq’s deal with Nvidia could throw a wrench in the reverse acquihire strategy. Tech moguls get the Hollywood treatment.

The Main Item

OpenAI Wanted Its Revenue Run-Rate to Match Anthropic’s. Cue the Chaos

AI stocks got hammered Thursday after the Financial Times reported that OpenAI was spinning its annualized revenue figure to investors in order to match Anthropic, and that its ARR was actually $20 billion below a widely cited $70 billion figure.

On the one hand, it was simply a confusion about gross versus net revenue.

On the other hand, OpenAI, Anthropic and the scoop-hungry media have been complicit in the confusion. OpenAI has long been complaining that its revenue figure looks feeble compared to Anthropic's because it’s been using a net figure while Anthropic has been using gross (ie Anthropic's number includes the sales of models by partners like Amazon and Microsoft, even though the partners take a percentage of those sales).

But when the FT got its hands on OpenAI's latest revenue figure — circulated as part of the company's current funding round — the paper blew it up into a headline saying it was $20 billion below earlier reports. The story explained this, but stocks took a hit anyway. Bloomberg later reported that OpenAI is expecting to reach $70 billion annualized revenue by the end of the year — with the assumption that this is net and not gross. Investors and customers believe that Anthropic and OpenAI are running more or less neck-and-neck.

The reality here is that sophisticated investors looking at these figures should understand the gross versus net discrepancy. But Dan Primack put it best:

A lack of transparency is also at the heart of the tranched-round issue that we’ve been writing about, as Natasha Mascarenhas at Bloomberg points out:

Deals at multiple valuations are becoming normalized, and the optics are one of the reasons. As journalists we’re pro-transparency in general, and there are real risks for investors, employees, and founders alike if the true valuation figures aren’t made clear.

China Challenge

Reflection AI, Mistral & Germany’s Aleph Alpha Aim to Invent Open-Source Business Models

Until now, the policy debate over the possible risks of open-source AI models has been wrapped up in the China question. Not anymore.

American and European companies this week took aim at red-hot Chinese rivals such as DeepSeek with the announcements of three new open models: Reflection’s Beam in the US, Mistral’s Large 4 in France, and Aleph Alpha’s Kolibri in Germany. All three startups were openly positioning their new models as the ones to pick if you’re concerned about AI sovereignty.

Reflection’s new model enjoyed a blitz of coverage across Axios, Reuters, and Fortune, with the company press release explicitly saying it “advances the frontier for the Western open ecosystem.” Beam’s core claim is that it’s even more efficient on inference than Chinese model maker Z.ai’s highly touted GLM 5.2, though that isn’t Z.ai’s latest offering.

Absent government pressure or new regulations, though, the sovereignty pitch may prove a tough one. Investors we spoke with this week said few of their portfolio companies were concerned about the national origins of their open-source models at this point, other than those working with sensitive trade or government data. They’d prefer a Western model in theory, but many will gladly use a Chinese one hosted by a US inference provider like Baseten, Fireworks, or Modal.

Thus far, proponents of open source have prevailed in arguing against regulation. But the obvious risks of cutting-edge AI in the wrong hands, combined with the business threat that cheap open-source intelligence could pose to OpenAI and Anthropic, mean the question isn’t going away.

Bill Gurley, a vocal open source backer, took to Substack this week to defend open source and pitch it as an ideal AI business strategy. Open models, he argues, are more secure than proprietary systems because the wider community stress-tests them — a concept that proved true in an earlier era of computing. Any regulation based on the potential dangers of open source will likely manufacture a monopoly for frontier labs, he says. “IBM was not allowed to outlaw Dell.”

Benchmark even backed Singapore-based AI provider Manus and cashed in when it was bought by Meta, though that deal was unwound by Chinese regulators. Manus this week raised more than $500 million in fresh capital after its founders bought it back (see below).

Menlo Ventures partner Deedy Das noted that cheaper, high-performance open models are great for the application-layer companies, and not such a big threat to the frontier model-makers.

“I find it all very strange that people pit them against each other because I think there’s clearly a world where the pie grows faster and everyone grows,” he said.

Is Open Source Cheaper?

Reflection claims it can run comparable queries on a fraction of the compute needed before, so cost is certainly part of the pitch. Yet it’s not clear whether open-source models will ultimately retain a massive pricing advantage.

Another investor we spoke with said implementing open-source models can get expensive quickly, given how much time and infrastructure spending is needed to customize them for specific tasks. Outsourcing it to an inference provider can be cheaper. But the total cost is still a function of how much you call on it, and the tokens add up.

When it comes to underlying performance, the new Western rivals are comparable to the Chinese models on some benchmarks, but behind on overall capabilities.

They haven’t yet dented the rankings when it comes to demand: Top Chinese models like GLM 5.3, Kimi K3, and DeepSeek V4.1 remain the most popular on the cloud platforms Baseten and Fireworks.

And of course nothing is slowing down in China: DeepSeek is close to raising $12 billion in funding co-led by Tencent and battery company Contemporary Amperex Technology Co. per Bloomberg.

How the business model for open-source LLMs will shake out remains an open question. Low prices are good for application-layer founders, of course, but LLM training remains very expensive.

We have already seen some companies step away from this race. Databricks CEO Ali Ghodsi told Newcomer this summer that by the time Databricks released its own open-weight foundation model in 2024, he was ready to move on from such an expensive and laborious task that could only produce a product that quickly became obsolete.

Tech giant Meta is still building open-source models, though it did pivot with its first-ever closed-source release this year, Muse Spark. The new open-source startups will need to be creative and nimble to avoid a similar change of course.

AI 100

Newcomer Partners with Bessemer for AI 100

We’re thrilled to be producing this year’s AI 100 list alongside Bessemer Venture Partners to highlight the most important AI companies in Silicon Valley.

Submissions for the list are due Oct. 15. We’ll share the full list on-stage at the Cerebral Valley AI Summit on Nov. 12.

Submit the companies you want to see represented here.

Submissions

Two Big Charts

Deal Value in 2026 Smashes 2021 Record, Q3 Slows to More Normal Pace

It’s been a year of massive, concentrated bets in venture, pushing dollar totals to new heights.

Deal value for US startups has already surpassed 2021’s record-breaking total by around 44%, according to the latest quarterly PitchBook-NVCA Venture Monitor report.

Read more

On a Monday morning in April of 2023, Lightspeed Venture Partner’s co-leaders Ravi Mhatre (pictured) and Bejul Somaia directed partners around the world to clear their calendars for the day. Get out of everything, they said. Even board meetings.

ChatGPT had come out a few months earlier and the two executives didn’t believe the firm was primed to meet the moment. Khosla Ventures, a direct rival, had an early stake in OpenAI — Lightspeed did not. It had also missed other breakout AI companies, including ElevenLabs and Hugging Face.

Somaia led the meeting and it went for seven hours. The partners broke into subgroups and reconvened. Those in Europe and Israel stayed on past midnight. The question on the table was, if AI is going to change everything, what would it mean to act as though that were true?

Read more

We were fascinated to hear from the founders and investors building a new generation of companies in finance and commerce at our Machine Earning AI Summit in San Francisco last week.

Industry leaders including Jean-Denis Greze, the co-founder of Town and former CTO of Plaid; Katie Haun, the founder of Haun Ventures; Ryan King, the co-founder of Chime; and Ilan Zerbib, the co-founder of Sapiom, offered insights on where fintech is going in the AI era.

We’re sharing all nine on-stage conversations here in the newsletter and on our Newcomer AI Summits YouTube channel.

Big thanks to our sponsors Airwallex, Exa, Nebius, Lead Bank, and OnePay.

Jean-Denis Greze (Town)

Town co-founder and CEO Jean-Denis Greze told the crowd that he isn’t losing sleep over agents going rogue. He said humans will approve purchases at first, and the mistakes that slip through will get budgeted in, as with credit card fraud today.

He does view personal AI agents as a business tool, calling the consumer business “a market to zero” because regular people rarely pay for software.

Jackie Reses (Lead Bank)

Lead Bank CEO Jackie Reses had spent weeks in Washington, D.C., ahead of the summit, meeting with stakeholders about her new 30-page proposal for how the financial system should treat AI agents.

“If a model hallucinates, whose responsibility is that?” she asked. She also reminded the room how early it still is for agentic commerce.

Katie Haun (Haun Ventures)

Katie Haun began the day emphasizing how much the systems for smart payments have improved in just a few short years.

She compared the AI industry’s rapid growth and regulatory hurdles to what crypto companies went through. She noted that, as with crypto startups, AI companies seem unprepared for the public backlash they are now facing.

Anish Acharya (Andreessen Horowitz)

Andreessen Horowitz general partner Anish Acharya named the three areas he thinks will produce the next wave of unicorns: industry-specific personal agents, coding agents for small businesses, and AI for entertainment. “The horizontal agents are going to hire the vertical agents to do things. Not enough people are building there,” he said.

Omer Ismail (OnePay)

Omer Ismail, the co-founder and CEO of the consumer fintech OnePay, made the case that “walled gardens” on the internet will soon become a thing of the past, thanks to AI agents. Safety will come in the form of new kinds of limits or guardrails on the agents — OnePay has “spend pockets” that cap what an agent can do.

Ilan Zerbib (Sapiom) & Paul Klein IV (Browserbase)

Sapiom co-founder and CEO Ilan Zerbib and Browserbase co-founder and CEO Paul Klein IV made the case that the big opportunity in agentic commerce is agents that buy on behalf of businesses. Browserbase gives agents the tools to navigate websites, and Sapiom builds the rails that let them pay. Klein said agents will take over back-office work like accounts payable and accounts receivable. One Browserbase customer’s agent recently made a $300,000 purchase on a credit card, Klein said.

Pranav Sood (Airwallex) & Mike Taylor (Gusto)

Gusto CFO Mike Taylor and Airwallex CFO Pranav Sood explained how they have integrated AI tools at their companies, with Taylor saying Claude has “completely changed” his role as a CFO. He now reviews Claude’s summaries and reports instead of raw spreadsheets.

Gusto, he said, has passed $1 billion in trailing revenue and has been profitable for about three years. Sood said Airwallex now operates in more than 20 countries and has reached $1.4 billion in annualized revenue; he also pushed back on recent scrutiny of Airwallex’s China ties.

Max Rhodes (Faire)

Faire co-founder and CEO Max Rhodes told the crowd at the Machine Earning AI Summit that he isn’t worried about AI agents cutting marketplaces out of the loop.

He argued that what LLMs do well, like back-and-forth conversation and personalization, is exactly what marketplaces can build into their own shopping experiences.

Ryan King (Chime)

Chime co-founder Ryan King said AI agents will make switching banks much easier, which is a big threat to how traditional banks have kept customers.

Still, he said regular people aren’t ready to let AI run their money. Chime’s AI assistant Jade helps customers track their budgets and spending habits but stops short of making direct purchases for them just yet.

Anthropic is preparing for its impending IPO while the markets are increasingly uncertain. AMD agrees to acquire Fei-Fei Li’s World Labs for $8.2 billion in a strategic move to assert its relevance in the AI race. Reid Hoffman returns to the podcast to pitch the benefits of the AI boom. Data from Odin’s Dan Gray illustrates how the VCs with the most capital behind them have typically backed the most unicorns. Salesforce agrees to buy Listen Labs for reportedly $2 billion. Meta poaches MongoDB’s CEO. OpenAI decides to scrap plans for its Astra rollout over safety concerns. Vinod Khosla fires off against a founder he invested in after two of his portfolio companies, Factory and Cognition, brawl over a poached leader. Instinct rakes in fresh funding from Coatue at a $10 billion valuation.

The Main Item

Anthropic's IPO Is About to Test the AI Boom

Over the past decade, it’s never felt like there’s been a perfect time to go public — the conditions are almost always rocky and the private markets have been so comparatively cozy.

But when you need as much capital as the frontier labs do, at some point you have to plug your nose and jump.

The last week has been a whirlwind of leaks, scoops, and announcements from Anthropic and OpenAI as they try to figure out how to land their next tranche of desperately needed capital.

On Monday Reuters pulled off a downright impressive scoop, getting its hands on what appears to be a draft of Anthropic’s prospectus. The numbers in it showed its staggering rate of spending. Last year Anthropic reported an operating loss of around $8 billion on revenue of $4.6 billion. It also was committed to spending $518 billion on cloud, computing, and infrastructure in the coming years.

The caveat here is these numbers are all before Anthropic’s Opus 4.6 model came out in February 2026; that was the monster model that kicked off the company’s agentic coding dominance and the subsequent “tokenmaxxing” craze that rocket-powered its revenue. Judging Anthropic without Opus 4.6 would be like evaluating Marvel before Iron Man.

Still, there have been swirling questions about Anthropic’s revenue growth rate and whether the company would be able to hit a $2 trillion valuation in an IPO. It has already postponed an expected October debut and now is looking to start trading in November before Thanksgiving.

The public markets in general don’t seem too welcoming. SB Energy, the SoftBank subsidiary that builds AI data centers, delayed its IPO in September. And this week Oura postponed its IPO indefinitely, citing market conditions.

OpenAI, meanwhile, is trying to bide its time as long as possible. The company is in early talks to raise a new private round, targeting at least $30 billion at a $1.4 trillion valuation, which is supposed to last the company until a 2027 IPO. Sam Altman has been spinning the delay as a reaction to concerns over model safety rather than anything wrong with the underlying business — to that end, several outlets reported this week that the company’s annualized revenue had neared $70 billion.

The next few weeks will be very instructive in assessing the health of the AI business (and the economy writ large). If the latest Anthropic financials, when they get released in full, show a company with still-booming revenue growth, then there’s much more runway to this cycle and we’re likely to see a downstream effect on the rest of the industry. If we start to see a slowdown, then things will be rough for the foreseeable future.

AI Mega Deals

AMD Is a Dark Horse. Is it Gaining?

Chipmaker AMD announced this week that it would be acquiring computer vision pioneer Dr. Fei-Fei Li’s AI startup World Labs for $8.2 billion in an all-stock deal. It is yet another shrewd move by AMD’s longtime CEO Dr. Lisa Su as the firm seeks to both win hyperscaler market share and expand into new verticals.

World Labs, fittingly, creates world models — simulated multi-modal and interactive environments that generate themselves in a consistent manner. The other big players in this space include Odyssey and former Meta chief scientist Yann LeCun’s AMI Labs. AMD says it is buying the firm to help it better understand how AI workloads are evolving as “demands on compute infrastructure become more diverse.”

In a Substack post discussing the acquisition, Li noted that Su was a “great friend” and an early investor in the firm. She also noted that the two companies had teamed up back in 2025 to conduct “model training and inference optimization on AMD GPUs.”

In such a male-dominated industry, it is both striking and exciting that AMD now has two women in its most senior technical roles as the firm continues to rebuild its claim to being a serious NVIDIA competitor.

The World Labs acquisition is the news of the week, but the background story which bubbles along is that initial shipments of AMD’s first rack-scale AI solution, Helios, will begin by the year’s end. Investors are excited to see these systems out in the world; HPE stock rose sharply this week after it announced a $1.2 billion deal to provide Helios racks to cloud infrastructure company Vultr.

AMD’s enterprise-grade GPU product line, called Instinct, also boasts exceedingly healthy demand, with 14 gigawatts of demand for these chips from OpenAI, Meta and Anthropic alone. For reference, McKinsey projects that there is 62 gigawatts of total global AI workload capacity in 2026. Anthropic’s two gigawatt partnership will also take the form of Helios racks.

This is not to overstate AMD’s position; it lags well behind NVIDIA at a mere $1 trillion market cap, has a much higher P/E multiple and lacks the CUDA ecosystem moat. Nonetheless, AMD has handily beat analyst expectations four quarters in a row, and Dr. Su can take a deserved victory lap.

Newcomer Podcast

Reid Hoffman on ChatGPT, Data Centers & Why America Needs AI

One Big Chart

The Biggest VC Firms Backed the Most Unicorns

If you have a big enough war chest, it’s easier to pay up for the best deals.

Megafunds that secured the most capital from 2010 to 2020 also tended to back the most future unicorns, according to data compiled by Dan Gray over at Odin.

Accel is the one outlier at the top — it raised less than $10 billion but invested in close to the same number of unicorns as a16z and Sequoia, which both brought in well over $10 billion during the same time frame.

Five Notable Deals

Listen Labs, EliseAI, Armadin, Varda Space, Hertha Metals

Read more

Benchmark has won out in a hot competition to lead a new funding round for early-stage chip startup Tendrils Compute, according to multiple people familiar with the matter.

Sources say that the company is already discussing a fast follow-up raise that would potentially value the firm at more than $1 billion. Neither round has previously been reported. Benchmark declined to comment.

The Tendrils deal is the latest in a wave of big VC bets on chips, marking the revival of the sector that gave Silicon Valley its name but has long been out of favor as investors focused on software and networking. Startups are betting that specialized chips for specific workloads will have an important place in the infrastructure build-out — in contrast to previous generations of the technology that were dominated by a small handful of chipmakers.

The small Tendrils team, which is led by Nils Cremer and based in Cambridge in the U.K., is betting that general-purpose chips based on an obscure graph-based computing model called interaction nets will let it scale without the coordination and cost problems associated with contemporary multi-core CPUs. As AI agents complete multiple steps of inference and tool calls, CPU speed may become a new bottleneck.

Subscribe to Newcomer to see our market map of hot chip alternatives.

Benchmark has been on the leading edge of the new chip enthusiasm. Back in 2016, Eric Vishria co-led the Series A round in Cerebras Systems and saw a 20.5x multiple on its check when the company went public this spring. It was Benchmark’s first hardware investment in 10 years, and Vishria almost skipped the meeting, he told TechCrunch.

That success, combined with the relentless demand for AI compute, has convinced VCs that there are plenty of returns to go around on chip startups. In August, Etched, which builds specialized chips and racks for AI inference, doubled its valuation to $21 billion off a fresh $700 million fundraise. Fellow inference-focused chipmaker Fractile is in talks to raise at a $6.5 billion valuation after landing a deal to supply chips to Anthropic. OLIX, a U.K.-based inference chip startup, raised $312 million at a $3.3 billion valuation back in August. Two other companies each announced nine-figure capital hauls in September: inference chip startup Euclyd and Delos Data, which builds software and networking equipment to help information travel more quickly around data centers. Just this morning, Volantis Semiconductors raised $88 million from Lachy Groom and Abstract Ventures for another inference-focused chip.

“They have a massive tailwind from capex build-out,” said Darian Shirazi, managing partner at Gradient. “They probably will all do well.”

Still, chips are notoriously capital-intensive and cyclical, with long development timelines, and are not for the faint of heart. Cerebras, for example, went public in May at $185 and has traded as high as $386, but now sits close to its offering price.

VCs have been getting creative to share the risk and cost of chip development. Coatue and startup chipmaker MatX have been in talks to form a joint venture to the tune of billions of dollars to help the company finance memory purchases and capacity with manufacturers, per The Information.

We’ve heard from VCs that the emerging chip ecosystem is noisy and difficult to parse. We’ve created a market map that helps clarify the state of play, focusing on three different types of chip startups.

You can also get insight on how investors are thinking about this sector from Newcomer’s VC Sentiment Report — a brand-new, quarterly study featuring quantitative questions and no-BS quotes from 25 anonymous VCs representing many of the best-known firms. Check it out here.

Read more

Meta launches its personal assistant Muse with national advertisements, a cute logo, and lots of ambition. Insurance startup Corgi goes big on outlandish marketing techniques but hits pause on cafés. A16z partner Josh Elman talks Apple and consumer AI on the podcast. VCs share their opinions on the top venture firms in Newcomer’s VC Sentiment Report. Lightspeed backs big AI rounds. Bessemer brings in billions for its latest funds. Investor China trips are back in vogue. Oracle cites “force majeure” in its New Mexico data center buildout. A16z launches a new educational venture with the Horowitz Andreessen Academy. Top AI CEOs push back against the AI doomers. The QuantumJane newsletter launches from a friend of Newcomer.

The Main Item

Meta’s Muse Jumps to an Early Lead in the Personal Agent Race

Personal agents have abruptly arrived as mass-market consumer products, led by Meta’s Muse.

They promise to be a godsend, lightening our administrative load in a big way and potentially bringing a welcome shift in the balance of power away from the big platforms and toward individuals. Investors are certainly convinced: startup Instinct, barely done with a round valuing it at $2.5 billion, is now reportedly looking for new money at $10 billion, the competition from Muse be damned.

It’s almost certainly early days for a category that could be the Next Big Thing, with many more entrants to come.

The form factor could grow beyond your cell phone, too — Mark Zuckerberg debuted Meta’s plans for a Tamagotchi-like personal Muse device on Wednesday, dubbed “Muse Charm.” The company is targeting a rollout of the device in time for the holidays.

Partnerships are popping up fast. Shopify CEO Tobi Lütke announced a deal with Muse on Monday to allow the AI assistant to check out across shops on its platform; Shopify stock jumped more than 10% on the news. Its deal likely won’t remain exclusive as Shopify aims to play with all parties, according to a source familiar with the matter.

Yet a lot has to happen for personal agents to become ubiquitous utilities. The truth is, neither the plumbing of online commerce nor its business norms are sufficient to handle an army of bots replacing humans. There will need to be rebuilding, and re-imagining.

Early adopters of Instinct complain that it keeps bugging them to fill out CAPTCHAs — to prove that it’s human. Amazon announced that it would block Muse from shopping on its platform.

Popular restaurants are seeing agents swarm their reservation systems, thanks in part to agent marketing pitches that cite painless restaurant booking as a feature. It’s easy to see the eateries or the reservation platforms simply banning them.

Partnerships won’t be a silver bullet for companies whose business model could be under threat. Travel platform Expedia announced an integration with Muse, but that didn’t stop its stock from sinking, along with its travel booking peers, on fears that it could be disintermediated by an agent army.

Indeed, a new version of the SaaS-pocalypse developed this week as whole categories of companies — dubbed the “consumer inertia” sector — saw their shares tank on the perceived agent threat. Banks, brokerages, and mobile phone carriers were down on the prospect of consumers switching effortlessly for better terms. Third-party services like OpenTable and DoorDash, which could be bypassed, got hit. So did subscription-based media businesses that could be vulnerable to agents rationalizing purchases; New York Times Co. shares are down 15% since late last week.

As with the SaaS-pocalypse, though, this is likely an overreaction; for consumers and businesses alike, software and services can be a lot stickier than they look.

It’s hard to imagine that Amazon, for example, will be able to maintain a policy of “no outside agents allowed” if its customers want to use them. But it intuitively seems reasonable that the company would want to treat humans and agents differently, for technical reasons if nothing else. It’s also intuitive, though maybe less reasonable, that the company will do everything in its power to prevent agents from taking over its relationships with customers.

Will Muse function less well with travel sites other than Expedia? Will it ultimately need them at all if it can access the underlying data via Amadeus and Sabre? If Muse is your agent, will Meta decide which booking service you use? Will that choice be good for the customer?

The questions pile up from there. But it’s exciting to think of the possibility of tools that could extricate us from the tedium created by their predecessors. We’ll certainly be pleased to never again have to pass a Completely Automated Public Turing test to tell Computers and Humans Apart.

Marketing Bites Back

New Dog, Old Tricks? Corgi Insurance Makes a Run

Corgi CEO Nico Laqua is very good at making headlines.

He’s celebrated a seven-days-a-week office culture and said he’d rather die at 50 than live until 80 if it ensured Corgi became a $1 trillion company. The histrionics invited blowback, which was undoubtedly the point.

When Corgi’s head of community was mocked in the New York Post for writing about her romantic non-negotiables on her X account, Laqua posted an extended defense of the company culture on X. Corgi made fresh headlines by buying a San Francisco building from basketball legend Steph Curry, who’d hoped to use it as a headquarters for his business ventures before getting into a tiff with local labor unions. Local real estate pros said the $11 million price was high.

Behind the bravado, though, strains are showing. One of Corgi’s central marketing initiatives is a chain of 24/7 cafés, and it had plans to open hundreds around the world. The company green-lit 16 new leases in San Francisco alone just two weeks ago, according to the San Francisco Business Times, before abruptly sacking the head of café expansion last week. One of the cafés was also shut down for lacking a valid health permit. Though a handful of cafés are in operation in San Francisco and elsewhere, the dramatic expansion appears to be on ice.

The Cute Animal Tradition

Corgi purports to be a new kind of insurance company, and Head of Underwriting Josh Shih told us the company’s ambition is to build a financial infrastructure company. It has more than doubled its value from $1.3 billion to $3.4 billion since January, according to PitchBook.

But in many ways it’s running a traditional playbook. Anna Martin, who works in European partnerships for Corgi, describes the company’s value as providing “normal insurance to the same audience at a different pace.” In an industry as old and strait-laced as insurance, delivering faster can be enough.

The marketing flash is also a long-standing industry staple. The GEICO Gecko was all but ubiquitous for years; celebrity-studded advertising comedy skits are a staple for the likes of Allstate and State Farm. It’s one of the only ways to stand out when selling a generic product that most people use rarely.

As Corgi expands into reinsurance, trucking, liquor, and real estate insurance, the firm seems to have recognized that an edgy social media presence might not be a trust-builder in specialty businesses.

Mario Conde, a partner at Bain & Company, told us that complex commercial insurance markets like cyber liability revolve around specialist individual underwriters. It’s unclear whether Laqua will be able to attract this talent from an industry he described as being “run on fax machines and with boomers.”

Erika Lee, Corgi’s head of brand, has a simple response to those concerns. The company hires an industry veteran to lead each new vertical, and “some of these executives have been underwriting longer than other staff members have been alive.”

Newcomer Podcast

Josh Elman on Apple, AI Agents & Why Startups Can Still Win

Two Big Charts

From Newcomer’s VC Sentiment Report: Founders Fund and Conviction Top Investors’ Fund Wish Lists

This week, we launched Newcomer’s VC Sentiment Report, the first installment in a quarterly, in-depth study of the state of the startup ecosystem based on interviews with 25 leading venture capitalists. Our team conducted 30- to 45-minute video interviews with top Silicon Valley investors — ranging from partners at large multi-stage VC firms to those who’ve gone on to raise their own funds — who were granted anonymity to speak freely. In addition to our qualitative and quantitative data, we have 344 unfiltered quotes from our interviewees.

Here’s a preview of two of our findings from behind the paywall:

We asked our panelists to tell us the multi-stage firm and emerging venture fund they’d invest in outside of their own. For established VCs, Founders Fund was the clear pick with nine mentions. Investors cited the firm’s contrarian strategy as a winner. “They’re kind of back to taking these huge swings on moonshot-oriented companies. They’re pretty fucking good at nailing those, at getting one to really work,” one confidant told us.

Sequoia came in second with six votes, though some of those votes were grudging, given how obvious a pick it is.

For emerging managers, Conviction was the runaway hit with 12 votes — no other fund got more than 2.

Become a paying subscriber to Newcomer to read the full report here.

Read the Report

Five Notable Deals

Snorkel AI, Cyera, Tekever, Enveda, Rightway

Lightspeed backed two big rounds this week — one in AI training and one in AI drug discovery.

Read more

The AI industry is growing faster than ever, and its future has never looked less certain.

No company embodies that contradiction better than Crusoe. The data center builder and cloud provider just raised $3.9 billion at a $30.9 billion valuation, led by Atreides Management, Valor Equity Partners, and Mubadala. And financials obtained by Newcomer show it’s on track for $2 billion in revenue this year.

Read more