Techin' Geek

The team occasionally shifts. Three years ago, I interviewed four partners from Benchmark, and currently, three of them have departed from the firm. The financial commitments have altered as well. This summer, Benchmark revealed $2 billion across two funds, marking only the second time in its 31-year journey that it has deviated from its usual practice of raising much smaller funds.

However, at this week’s TechCrunch Disrupt, the five partners of the venture firm effectively presented themselves as a cohesive group that isn’t rushing, unlike many of their competitors. While the world seems to be picking up speed, they are investing at the traditional pace Benchmark has maintained: just two or three new investments per year for each partner. Despite the overload of information globally, their Monday meetings among partners lack a formal agenda. And though economic signals suggest challenges ahead, they appear unfazed, interpreting many warning signs as results of X-fueled frenzy.

Here’s what the five general partners of Benchmark—convened on our stage—shared at the event. Some comments have been abridged for clarity; the full interview can be viewed at the conclusion of the article.

Benchmark has progressed in several ways, such as having a larger core fund and a new growth fund. How do you navigate change without compromising what defines Benchmark?

Peter Fenton: The foundation that remains unchanged is the core fund: collaborating intimately with founders right from the start. This typically involves a 10-to-15-year commitment, [working] closely, which doesn’t scale. There’s been inflation as well. We maintained the core fund at $500 million for two decades, and it has now been adjusted to align with the current economy, which we hadn’t previously done, and that was likely a miscalculation on our part.

[Regarding the growth fund] it represents an expansion of our mission. In some aspects, we’re authorized by our ecosystem to move beyond just our core early-stage activities. [Not having this previously] was a limitation for us. A range of opportunities presented to Benchmark could have yielded returns between 10 and 1,000x for us, but we were limited by our early-stage restrictions.

Are you considering making a few bets with that $1.25 billion growth fund?

Fenton: We all have varying perspectives, which is the advantage of a partnership. I’d personally love for us to experience, as that famous J.P. Morgan paper suggests, the agony and ecstasy of a concentrated portfolio. If the fund excels, one or two investments will account for 90% of the returns. So, I think it’ll be around four to six, maybe seven to eight.

Ev, you joined Benchmark late last year. You’ve been with Founders Fund, Bond, and Kleiner Perkins. How does Benchmark stand out?

Ev Randall: Every firm fundamentally reflects the individuals within it and their collaborative dynamics. Some operate with a highly hierarchical structure, containing a 50- or 60-member investment team with tiered processes, while others are small partnerships. Early in my career, I was attracted to small partnerships for their capacity for greater throughput and clarity in discussions, allowing for decision-making without convoluted corporate hierarchies. It’s akin to being part of a sports team.

Moreover, it’s essential to connect with your people. You must find those you not only like but also respect. At Benchmark, with the current group of GPs, it was apparent that everyone got along, but there was much more to it—it’s a team that respects and educates one another. This combination of qualities is rare, and once I recognized it, I was eager to join, and I believe Jack felt similarly as a newcomer.

Jack, you came on board in February, but just a year ago, you raised a second, $275 million fund for your own firm in what you described to the Wall Street Journal as a week. You host a highly popular podcast. What motivated you to join Benchmark?

Jack Altman: It’s amusing. I recorded a podcast with these guys earlier this year, during which we discussed the commitment to remaining a small fund. More than that, I perceived their desire to be the first call for founders, to maintain focus, and to genuinely care. Much of this industry is entangled in finances and reputations, yet what struck me about the Benchmark team, then and even more so as I got to know them, was their authentic concern for founders and helping companies. While I appreciated being an individual investor, in such a chaotic market, I found it challenging navigating alone. Having a group of smart, dedicated individuals to learn alongside has been fantastic. There’s truly no better place for growth.

I hear you gather on Mondays without a set agenda.

Chethan Puttagunta: Jack attempted to create an agenda. [Laughs.]

Altman: I’ve got an agenda ready for next Monday.

Eric Vishria: It’s quite arbitrary, actually. It serves as an ongoing dialogue. Amidst rapid change, we explore what’s becoming achievable that wasn’t possible half a year ago, and who’s shifting dynamics. We discuss companies we’re meeting, what has captured our interest and why, plus insights from board meetings. While trying to systematize it is tempting, the beauty of our group of five lies in the fact that we don’t have to. It’s a prolonged conversation only viable if we’re small and genuinely enjoy each other’s company. Following this session, we’re having dinner with one of the few entrepreneurs who has built a multi-hundred-billion-dollar enterprise, someone we’re not invested in—just to broaden our perspectives.

Puttagunta: Not having an agenda allows room to be responsive. If a founder needs assistance securing an executive or forming a strategic partnership, the five of us can step in to help.

You operate as an equal partnership. Can one person veto a deal?

Puttagunta: It’s unclear. We don’t have established rules regarding that.

Altman: I honestly don’t know. I was waiting for someone else to say something.

That’s never occurred?

Altman: Due to the firm’s design around few deals, having rigid protocols isn’t actually desirable. If we invested in 100 or 200 firms annually, you’d want a CRM and an official voting approach. Since that’s infrequent, the minimal structure is a significant advantage.

Fenton: One of our brand’s unique features is that we don’t rush to judge winners or losers. It’s not Shark Tank. We feel a sense of wonder, for lack of a better term, at all the potential . . . The rare occasion when someone does come in that’s polarizing — the entrepreneurs notice that. If one member of the group doesn’t resonate with you, we may not be the best partner.

Businesses are experiencing growth at unprecedented rates. It seems easier to feel like you’re missing out on prospects if you’re not investing more frequently. Have you increased your speed?

Puttagunta: I believe our pace remains steady. A core truth of venture capital in our field is that we won’t be involved with every thriving business, and that’s perfectly fine. There are companies we recognize, where we see the entrepreneur has specific strengths, and we aspire to enhance that strength. Our role is to discover how to collaborate with that entrepreneur. It doesn’t matter how much funding they aim to raise, what structure, or their current status. You simply must navigate through everything.

One trend that founders are keen on is VCs backing rival firms. Has the VC-founder relationship shifted due to rapid business expansion and encroachment into each other’s markets?

Vishria: We believe strongly in committing to an entrepreneur, and with that commitment, we affirm that we will engage closely with them in whatever endeavor they pursue, which should matter to them just as it matters to us. The harsh reality today is that every company is progressing rapidly, leading to natural shifts in business focus, pivots, or transformations, which will result in conflicts. That’s simply part of the current landscape. . .

Naturally, we maintain confidentiality and take necessary precautions, and we have been fortunate to avoid direct confrontations [as seen recently between rival companies Factory and Cognition]. . . that commitment to exclusivity in specific domains is truly vital. But circumstances are changing. There’s no doubt about it.

Puttagunta: Additionally, individuals are exceptionally ambitious in this AI age. With the costs of developing new products dropping significantly, founders recognize new opportunities and are eager to exploit them, suggesting that overlaps between competitors will likely increase. Nonetheless, it’s possible to maintain healthy relationships and mutual respect.

Jack, have you made any investments for Benchmark yet, and if yes, did you feel a lot of pressure with that initial deal?

Altman: I have made several, some of which are fairly recent and not all are public yet. My background is in software, but since joining, I have also ventured into chips, robotics, and infrastructure. When I joined, someone asked how my onboarding was, to which I responded, ‘It was perfect. There was no onboarding.’ Either Chethan or Eric mentioned early on, ‘Hey, you might feel pressured with your first deal. Don’t. You’ll be making numerous investments while you’re here. This industry involves unpredictable power dynamics. The chances of your first one being a major success are low.’ . . . I trusted they understood my approach as an individual and as an investor, and I aimed to remain true to that.

The dynamics of these decisions are heavily influenced by the group. The investments I made, I doubt we could have secured without the support of my partners. I’m not sure I would have known how to say yes without them.

Eric, investing in the AI chip company Cerebras was an unusual choice for Benchmark when it occurred. What led you to proceed with it?

Vishria: We invested in the spring of 2016 in their inaugural round. We hadn’t engaged in hardware investments in a decade. The founders appeared capable, their thesis was straightforward, and I struggled to grasp most of the rest. The Sunday night prior to their pitch to the partners, Peter shared with me that it contradicted everything he had learned in venture. However, after the Monday pitch, he had changed his mind. A significant part of the decision-making was understanding each other’s instincts. This marked the start of a journey filled with numerous highs and lows and several potential company failures along the way, and it’s far from finished. The passion for connecting with a founder and their vision is why we pursue this work.

Fenton: Andrew and his co-founders are outliers. What they proposed was unconventional, and this is evident today, as people react negatively to the stock due to its polarizing nature. That’s precisely where we need to position ourselves as investors. If everyone agrees, why would there be a startup?

On that subject, Cerebras was priced at $185 when it went public earlier this year, rose to $350, and has now returned near $185. What are your thoughts on this?

Vishria: About six weeks prior to the IPO, I presented multiple stock performance graphs: CoreWeave, Datadog, DoorDash, Amazon. The fluctuations they experience, from peak to trough, are astounding—often moving 500% within a year. So, who can say? You manage what lies within your control, and update the public markets every 12 weeks. Being a publicly traded company requires a steel stomach, and employees need to comprehend that.

Your former colleague, Bill Gurley, has long claimed that the conventional IPO process is flawed, with the IPO ‘pop’ benefiting bankers and their clients rather than the company. He advocates for direct listings. Does anyone share that perspective?

Vishria: I believe I may be the sole one here with direct listing experience. It can absolutely make sense in the right context. However, for this wave of AI firms, it’s unlikely since they require substantial capital. Ultimately, IPOs are about raising funds. Cerebras secured $6 billion. That’s significant, hence I think it’s improbable for this group mostly due to their capital needs.

Could a serious correction be on the horizon? Many AI firms have depleted their free cash flow. Borrowing costs are substantially high. Recently, we also witnessed the first AI bankruptcy with Character AI.

Randall: Predicting macro events with precision is always a challenge. Since I began investing, the most substantial change has been in the information landscape. Character.AI was the remnant of a company whose founder and team were acqui-hired by Google a year and a half ago, and it previously sold for $2.3 billion; its Chapter 11 filing perceived as an AI chill effect illustrates how the data environment for public tech equities is captured in this frenetic X echo chamber.

When you observe the market today and analyze CoreWeave and Nebius alongside others, those values correspond to the ARR figures of two private companies likely to go public within the next year . . . There’s a significant amount of capital expenditure and investment flowing through the anticipated health of OpenAI and Anthropic.

Our firms are inundated with demand and struggle to meet the computational needs of customers willing to pay them with favorable gross margins. There’s nothing fundamentally wrong with the AI ecosystem. What is troubling is the volatility stemming from often misleading or exaggerated data meant to attract clicks. I genuinely believe that OpenAI and Anthropic going public would stabilize the market, preventing shifts of 5% to 10% every time an anonymous account alters an estimate.

Those two companies have also significantly shifted the center of gravity back to San Francisco. Must a founder be situated in San Francisco now? Are you considering companies beyond California?

Altman: I wouldn’t say being in San Francisco is a requisite for founders, but it undeniably is currently the easiest place to secure talent, capital, and clients. Everyone [on the team] engages with firms outside of San Francisco, yet we’re also in San Francisco daily. So, while it’s not mandatory, I’d argue there’s a strong bias given our situation.

Fenton: Historically, over 90% of capital returns have originated from companies established here. There are compelling stories outside of this concentrated trend — Shopify is an exemplary case, as is Ramp — yet the ecosystem that develops when a successful company emerges fosters a backdrop of exceptional executives [and] ambitious individuals, raising the bar for success. I don’t believe this systemic occurrence has transpired elsewhere like it does here.

Randall: I’m reflecting on the non-SF-based firms I’ve encountered in the past 18 months that did not express interest in engaging with San Francisco. They are either opening an office here, spending a significant portion of their time in the city, or planning to relocate to San Francisco. Iron sharpens iron. If you aim to associate with peers that will refine you into your best self, it’s hard to find a locality that accomplishes that better than San Francisco in this current AI era.

[embedded content]

“You Can See Everything” has already become a box office hit, opening with $15 million during its debut weekend (according to Sunday projections). This marks the second-largest opening weekend for a documentary film in history.

The nearly three-hour film — revealed last month at a clandestine screening at the Telluride Film Festival —  is only showing in 855 cinemas this weekend. (In comparison, the top movie of the weekend, “Street Fighter,” is being screened in 3,575 venues.) Nevertheless, it reportedly had 2,000 sellout shows, with reviews being extremely favorable; “You Can See Everything” now boasts scores of 99% on both Rotten Tomatoes and Metacritic.

The narrative of Elizabeth Holmes’ ascent and downfall — establishing the blood-testing startup Theranos, garnering backing from many prominent figures in Silicon Valley, and ultimately receiving an 11-year prison sentence for defrauding investors — has been recounted in books, podcasts, television programs, and various documentaries.

However, the filmmakers of “You Can See Everything” were granted significant access to Holmes and her family during the month leading up to her incarceration, and the participation of “The Rehearsal” creator Nathan Fielder — who co-directed the film with Lance Oppenheim — adds an extra layer of intrigue.

While I haven’t personally viewed the film, I can confirm the challenge of securing tickets in Manhattan: I spent considerable time yesterday seeking a showing for tonight and managed to find only one with available seats that weren’t in the front row. (Yes, I am definitely going.)

TechCrunch’s Amanda Silberling managed to catch it on Thursday. She mentioned that the enigmatic Holmes seems to have finally encountered her match in Fielder, a deadpan comic who is arguably just as peculiar and enigmatic as she is. Fielder’s month spent with Holmes is reportedly just the start of what appears to be an increasingly surreal, metatextual journey.

The buzz surrounding “You Can See Everything” might be making it tougher for other technology-focused films to gain traction. Aaron Sorkin’s follow-up to “The Social Network,” titled “The Social Reckoning,” debuted to a disappointing $5.5 million last weekend and has already dropped out of the top 10 entirely. Meanwhile, Alex Gibney’s four-hour documentary “Musk” — possibly hindered by pushback from social media platforms and its lengthy runtime — only earned $306,702 this weekend.

Welcome back to TechCrunch Mobility, your destination for the future of transportation and the significant role AI is playing now more than ever. To receive this in your mailbox, subscribe here for free — just click TechCrunch Mobility!

As I concluded my interview with Rivian CEO and founder RJ Scaringe at TechCrunch Disrupt, one inquiry kept coming to mind: What vision does Scaringe have for Rivian? Observing Rivian’s actions over the last year, it’s evident that Scaringe desires the company to be much more than merely an electric vehicle manufacturer. Yet, he also intends to maintain a connection to that original purpose.

Rivian has become significantly more vertically integrated compared to just a couple of years ago. It has constructed its own AI models to enhance its in-car assistant and is creating a fully automated driving system for its passenger cars and upcoming robotaxis in collaboration with Uber. It’s even developing proprietary silicon, a 5 nm processor, alongside Arm and TSMC. Additionally, there is a $5.8 billion joint technology initiative with VW Group concentrating on electric infrastructure and software. 

Scaringe’s ambitions go beyond Rivian itself, though there’s always a connection back to it. He initiated a micromobility skunkworks project that has since evolved into a startup called Also. Furthermore, he founded an industrial robotics venture named Mind Robotics, which is now 31% owned by Rivian and considers the automaker a potential client. (Scaringe announced on stage that former Ford, Apple, and Tesla executive Doug Field has been appointed president of Mind Robotics.)

Scaringe’s vision was apparent as he discussed not only a broader joint venture with VW that includes AI and autonomy, but also the prospect of licensing Rivian’s technology to “numerous different brands.”

So, what will Rivian look like at the end of the decade? Certainly not just a manufacturer of electric trucks and SUVs. The term that kept resonating was “technology provider.” 

Consider this. Rivian is paving the way to license its electric architecture and software, along with its AI models and autonomy framework. Meanwhile, the Mind Robotics business it partially owns is constructing industrial robots intended for use in Rivian’s facilities but also to be sold to other enterprises. Additionally, there’s the robotaxi venture. A successful rollout with Uber could lead to further agreements.

And Scaringe is already envisioning how autonomy might redefine Rivian’s own vehicles.

“Where it becomes particularly intriguing is when you move beyond R2 and R3 and start to envision the design forms you aspire to create for a personally owned autonomous vehicle. It opens countless design possibilities. It can foster completely different experiences,” he articulated in an interview with TechCrunch. “I’m immensely excited about what the next decade holds for transportation because of what this allows us to imagine and contemplate.”

Deals!

Bedrock Robotics, a startup focused on autonomous driving technology for construction vehicles, is seeking to raise $600 million, Axios reported, citing sources. 

Dot Ai has finalized its acquisition of the electric vehicle car-sharing firm Envoy Technologies in a $65 million all-stock deal, according to a regulatory filing. 

Evos, a startup that crafts AI agents to streamline operations for industrial companies, including freight brokerages, secured $9 million in a pre-seed funding round co-led by Uncork Capital and Field Ventures.

Mana Battery, a startup from Colorado that is developing anode-free sodium batteries, has raised $10.15 million in a funding round led by Voyager Ventures. 

PlusAI, the autonomous trucking innovator, and Texas Ventures Acquisition III have submitted a Form S-4 registration statement with the Securities and Exchange Commission, progressing their previously announced decision to go public through a merger with a SPAC.

Volvo is enhancing its automated driving capabilities by expanding its nearly decade-long alliance with Nvidia. Neither entity disclosed the financial details of this agreement. The automaker intends to utilize a suite of Nvidia products to provide hands-free, point-to-point driving on highways and city streets. 

Waabi, a startup engaged in self-driving truck development, has secured a commercial partnership with carrier Warrior Logistics. According to the arrangement, Waabi’s self-driving trucks will transport freight between Dallas and Houston.  

Notable reads and other tidbits

Sales of electric vehicles in the United States are beginning to show positive trends.

Laplace Intelligence, one of 20 startups that participated in TechCrunch’s Startup Battlefield 200, is constructing automated red-teaming that evaluates robots and self-driving vehicles. 

MIT startup LNK Energies claims to have devised a method to convert existing heavy-duty trucks from diesel to hydrogen, significantly reducing pollution and costs while protecting operators from price fluctuations.

The Paris Auto Show highlights Chinese automobiles prominently.  

Tesla showcased its SpaceX-powered Roadster through an impressive kickflip demonstration. The Roadster employed a combination of high-powered fans and thrusters to ascend an inverted ramp, perform a spin, and land on the ground below. CEO Elon Musk asserts that the vehicle will be available for public use by 2028. However, will future owners be legally permitted to execute these maneuvers on public roads? I don’t foresee how that could be the case. 

Waymo is making inroads into Canada — specifically, Toronto — after obtaining approval to be part of an automated vehicle pilot initiative. Waymo has also made its robotaxi service available to everyone, meaning there are no longer waitlists, in San Antonio, San Diego, and Tampa.

One more thing …

TechCrunch reporter Tim De Chant recently had the opportunity to test drive the 2026Porsche Macan Turbo EV. His experience with the electric vehicle has him reconsidering the very concept of what a sports car should embody. Read the complete review.

According to a report released by the AI firm on Friday, Anthropic’s workforce composition and the nature of work being performed have evolved considerably over the last year.

Anthropic suggests that its internal work trends might be signals of “how AI may influence work organization, recruitment, and labor needs within the broader economy,” particularly since the company “has fundamentally adapted its workflows around AI.” However, even if other businesses do not adopt AI in the same comprehensive or identical manner as Anthropic, the report provides a fascinating insight into the company’s current operational framework.

Significantly, Anthropic is hiring fewer junior staff, with the share of junior-level employees in technical teams decreasing from 21% to 14% since early 2025, and the proportion of junior technical hires declining even more drastically, from 27% to 9%.

Moreover, Anthropic indicates that technical teams now comprise a reduced percentage of the total workforce, dropping from 61% to 56% — a shift attributed to the “natural expansion of business operations as the company grows,” along with “coding tools increasingly enabling non-technical staff to develop software independently.”

One employee reportedly noted an “ongoing pressure at Anthropic to expedite processes and utilize Claude for all tasks.” In an April 2026 survey, 27% of employees mentioned they had increased their productivity thanks to Claude, and 20% indicated their responsibilities had shifted “from performing actual work to managing Claudes”; reportedly, one engineer likened this experience to babysitting: “When you are babysitting, each child presents issues that equally demand your attention, and the various [Claudes] similarly do that.”

Indeed, the report implies that “human attention and judgement” are becoming a “bottleneck” for productivity, as Claude generates a “flood” of text that requires review and organization.

Predictably, the report claims that AI integration has enhanced the company’s productivity, as shown by “more code per individual, closing more sales per representative, and processing a greater number of contracts.” The firm acknowledges that “more code does not automatically equate to better code or business outcomes,” yet it has discovered no evidence of declining code quality.

According to the report, meeting time per employee has risen by 20% compared to a year prior, with a 16% increase noted for technical contributors. The company attributes this to the significance of “human trust and accountability,” as employees frequently describe “in-person coordination as essential.”

Replit, the AI coding platform favored by startup founders and developers alike, has ceased its collaboration with Anthropic, revealed the company’s CEO and co-founder Amjad Masad on Thursday during his appearance at TechCrunch Disrupt.

Masad indicated that the choice stemmed from various factors, cost being one of them. While Replit announced a collaboration with OpenAI in August, there had been no signal indicating a termination of the partnership with Anthropic. 

“I believe ChatGPT is superior now,” Masad remarked, noting that the company employs an internal benchmark to evaluate metrics like performance. “We also release a public benchmark, and consistently, Sol and Astra achieve better results in constructing vibe-coded applications from the ground up. This was a decision influenced by cost, partnership, and performance. For all these reasons, we felt they were the right partner at this time.”

Nonetheless, it appears that Anthropic has not been entirely removed from Replit, which offers a cloud-based software development environment where users can create and launch apps directly in their browsers. 

Following Disrupt, Masad clarified that users still have access to Anthropic models on the platform, but OpenAI’s ChatGPT is now the primary option. This change might lead to a notable decrease in Replit’s expenses with Anthropic, as most users tend to stick with the default setting instead of opting for a different model.

Before establishing a partnership with OpenAI, Replit depended on Anthropic’s Claude models for its platform operations. This collaboration was believed to have significantly contributed to the company’s revenue growth, and Masad remarked during the event that Replit had, until recently, been one of Anthropic’s top five clients.

“Replit continues to be model agnostic,” Masad communicated to TechCrunch over the weekend. “Our users have the option to select Anthropic’s Claude or other supported models…OpenAI is now our preferred partner for powering Replit Agent, based on performance, economics, and the robustness of our partnership.”

However, insiders acquainted with the relationship between Anthropic and Replit disputed Masad’s statements made on stage. One individual familiar with the situation, who opted to remain anonymous, asserted that the partnership remains active.

Evidence from the company’s website, blog entries, and social media updates shows that Replit was still advocating for Anthropic products just weeks ago. For example, a blog post from late September references Anthropic’s Claude Fable 5 and Claude Fable 5.1 as models operational within Replit Agent’s production framework. Simultaneously, a Replit product employee shared on X about utilizing Anthropic’s leading model Opus 5.5 and highlighting its features.

Masad also offered perspectives on the company’s growth while on stage at Disrupt, indicating that enterprise contributions are increasingly becoming a larger segment of revenue.

“We’re not disclosing ARR numbers just yet, but in terms of growth, enterprise was about 2% of revenue last year,” Masad stated. “This year, I would estimate that approximately 15% of our revenue comes from enterprise contracts, with more than 50% of our revenue attributed to B2B.”

Backstage, Masad conveyed to TechCrunch that he anticipates more than half of Replit’s revenue will originate from enterprise and government clients by 2027.

Masad also emphasized the importance of establishing robust AI safety protocols.

“Just as humans require a conducive environment for work—such as a comfortable office, good food, air conditioning, and computers—agents need a suitable habitat to thrive and perform effectively,” he noted. “This habitat must include proper orchestration, effective error recovery, strong safety and security measures, sandboxing, solid networking, and reliable identity and authorization. Once you navigate all of that, you can implement AI safely.”

This article was initially published on October 16 at 2:22 pm PT. It has been revised to provide additional context regarding Replit’s associations with Anthropic and OpenAI. The title has also been updated.