Techin' Geek

Anthropic has allocated almost one-third of its eagerly awaited IPO prospectus to risk factors, as reported by the Financial Times, which states it has examined the filing recently. The document outlines particular concerning behaviors that Anthropic claims its models have exhibited or might exhibit, such as attempts to “resist shutdown,” to “hide or alter information,” and actions “akin to blackmail,” according to Reuters.

The revelations are particularly bleak for a company whose backers are optimistic it could debut at over $2 trillion, more than twice its $965 billion valuation from May, potentially marking the largest IPO in history. It’s a peculiar situation for any firm — cautioning that its product could jeopardize humanity, while simultaneously making many of its initial investors and staff exceptionally affluent in the process.

Reuters was the first to disclose the financial specifics detailed in the prospectus on Monday, highlighting that Anthropic experienced an operating loss exceeding $8 billion in 2025 due to skyrocketing spending on computing power, while its revenue surged twelvefold to nearly $4.6 billion; however, escalating infrastructure costs in the previous year drove total operating expenses to nearly $13 billion.

Additionally, according to Reuters, Anthropic’s prospectus indicates plans to invest a staggering $518 billion in cloud computing and infrastructure in the coming years. (Anthropic has already secured computing agreements this year with Google, SpaceX, and Nscale, among others to achieve this goal.)

The FT reports that Anthropic’s metrics have escalated even more rapidly in 2026. The revenue for just the second quarter reached $11.5 billion, and the company is on course for its second consecutive quarter of operating profit on an adjusted basis.

Per the FT, the prospectus also highlighted customer concentration, with nearly a quarter of the previous year’s revenue generated from just two clients. (Details on these clients have not yet been revealed.)

The revelations, which are said to encompass “existential risks to humanity” — a first, based on a brief review of the SEC’s database — emerge as concerns regarding AI safety escalate rapidly.

CEO Dario Amodei has spent the month publicly advocating to “modulate the frontier” of AI advancement, informing the UN Security Council last week that AI poses a risk to humanity and deeming it “the most critical global security issue confronting the world today.” Competitors Sam Altman and Elon Musk have also voiced their support, marking a rare moment of unity among rivals who have usually taken the opportunity to publicly criticize each other.

Another competitor, Mark Zuckerberg, has dismissed the worries, stating to NBC News last week that he does not “believe we need some sort of industrywide coordination.”

The alerts follow a series of security breaches where AI agents have infiltrated external systems. Indeed, OpenAI revealed last week that its tools had hacked “dozens” of external sites, including government sites, one of which was the SEC’s. Earlier on Monday, the company announced it had abandoned plans to launch its latest model due to safety fears.

Peak XV Partners, a leading venture capital firm investing in sectors such as India and Southeast Asia with over $10 billion in managed assets, has raised its investment per startup through Surge, its seed-stage investment platform, while introducing a new cohort of 18 companies.

At least three companies within this cohort had already secured external funding, in some instances from Peak XV itself, prior to joining Surge.

The latest group, dubbed Surge 12, is the first to function under Peak XV’s elevated investment limit of up to $5 million per startup, an increase from the previous $3 million. The venture firm has invested more than $50 million in this batch, which has collectively garnered over $90 million in seed financing, according to Peak XV. The median investment per company has also risen, though the firm did not reveal the specific amount.

“The threshold for raising a Series A has significantly increased,” Rajan Anandan (shown above), managing director at Peak XV, stated in an interview. He noted that the firm is witnessing more capital-intensive businesses, especially in deeptech, raising larger seed rounds.

Anandan informed TechCrunch that Surge has become increasingly global with each cohort, with the current group featuring founders and companies from San Francisco to Sydney. Among the 18 startups in Surge 12, only five target the Indian market, while over half are based in India. The other 13 aim at global markets, emphasizing the distinction between where the companies are created and where they anticipate finding customers.

Since its inception in 2019, while Peak XV was operating as Sequoia Capital India and Southeast Asia, Surge has supported more than 180 startups founded by entrepreneurs from over 18 nationalities. Peak XV reports that the 10 largest firms emerging from those cohorts now generate more than $1 billion in combined annual revenue.

Anandan characterized Surge as Peak XV’s main channel for seed-stage investing, with the firm continuing to invest as companies move through subsequent funding stages. The founders it supports typically consist of repeat entrepreneurs, seasoned operators, and highly specialized technical founders, he noted, with around 50% to 60% of a typical cohort composed of individuals with operational backgrounds at established tech firms.

This cohort’s startups cover AI, robotics, space, consumer products, healthcare, music, and fintech, encompassing everything from AI safety and personal computing to autonomous robots designed for underground pipelines and satellites engineered for detecting radio-frequency signals from orbit.

The Surge 12 cohort

Alma — launched by Nischith Shadagopan M N and Vinod Ganesan — is developing a personal computing platform aimed at making computer usage faster and more cost-effective. Its founders had prior experience at Microsoft Research and were founding engineers at Sarvam AI, a startup in Bengaluru focused on AI models for Indian languages.

August AI — established by Anuruddh Mishra, an IIT-BHU graduate who began the company in 2022 following a personal medical misdiagnosis — offers a healthcare platform that integrates AI with physician-led services, reaching over 9 million users across 160 nations.

Ditto — co-founded by UC Berkeley dropouts Allen Wang and Eric Liu — serves as an AI-driven dating matchmaker within iMessage, designed to assist college students in transforming digital introductions into face-to-face interactions. (TechCrunch reported more on this last month.) The business had previously raised $9.2 million in a Peak XV-led seed funding round disclosed earlier this year.

GameStock — founded by Antoine Mistico, Easton Dana, and Vivek Indlebele Narasimha Prasad — integrates competition mechanics into financial markets, turning investing and trading into a more competitive experience. Mistico is a two-time founder and a former professional baseball player.

HiLoop — started by Jad Ghalayini, Karan Brar, and Thomas Boser — assists AI firms in adapting general-purpose open-weight models for specific uses through its post-training framework. Its founding team includes former engineers from Reducto and a PhD from Cambridge who completed his doctorate at 24.

Hoola Health — initiated by Deeksha Senguttuva — centers on healthcare for children and their families, providing consultations, vaccinations, medications, diagnostics, developmental therapies, and dental services on a unified platform. Senguttuva hails from a family involved in healthcare, having built and managed a hospital group.

Kello — launched by Mona Gandhi and Subramanya Jingade — is creating an AI-driven talent-discovery platform focused on assessing a candidate’s potential and growth rather than solely on traditional credentials. Gandhi claims she was Airbnb’s first female engineer and previously founded Upraised, while Jingade co-founded AmbitionBox.

Kindling — established by Adam Miller and Sachin Shah — is constructing what it terms a “storytelling operating system” for tech startups, utilizing AI to aid companies in developing and producing their communication and content.

Puralink — founded by Harrison Crowe-Maxwell, Shyeon Delnawaz, and Thien “Long” Tran — is creating autonomous robots capable of navigating underground pipeline systems. Crowe-Maxwell has been building robots since he was young and transformed university research into the patented drive technology supporting the startup.

Reinforce Labs — established by Anish Das Sarma — is designing tools for evaluating, red-teaming, and rectifying enterprise AI systems. Sarma previously founded a company acquired by Airbnb and went on to serve as a director at Google, where he led AI and machine-learning teams.

Riffle — co-founded by Anurag Choudhary and deo — is crafting a web-based platform for musicians to create, collaborate, and share music, minimizing the need to switch between different tools during the creative process.

Rosella — co-founded by Chris Dwyer and Sean Stuart — is establishing an AI-native commercial insurance brokerage for U.S. businesses, using AI to automate segments of the traditionally manual process of sourcing and securing business insurance. Rosella raised about $2.5 million in a pre-seed funding round led by Peak XV and Intact Private Capital earlier this year.

Tribe Money — founded by Himanshu Arora and Nikhil Shanker — offers an AI-driven personal finance platform that assists users in tracking their finances, researching investments, and making investment choices.

ULOOK — created by Adheesh Boratkar and Siddhesh Ravindra Naik — is developing autonomous satellite systems for radio-frequency sensing and spectrum intelligence, targeting global clients. Its founders have participated in over 12 satellite missions. The company had already secured roughly $2.3 million in seed funding from growX Ventures and InfoEdge Ventures prior to joining Surge.

Wingit — founded by Nikunj Kothari and Saksham Khandelwal — is constructing a beauty platform aimed at India’s burgeoning premium-consumer market. It concentrates on how consumers discover and purchase higher-end beauty products.

Three additional startups in this cohort have not yet publicly disclosed their identities or offerings. Peak XV mentioned they are engaged in education, applied AI, and medical products.

The autonomous vehicle tech firm Aurora informed investors last week of its plan to deploy over 30,000 self-driving trucks on the streets, aiming to generate $5 billion in yearly revenue by the conclusion of 2030 — an ambitious goal given that it anticipates finishing 2026 with merely 200 autonomous trucks and an $80 million revenue rate.

CFO David Maday believes that the seemingly lofty target is not as unattainable as it may seem.

“While 30,000 certainly feels substantial — and it does indeed in the field of autonomy — in relation to the overall truck market, it’s actually quite modest,” he shared with TechCrunch in a recent chat, further noting that the four leading truck manufacturers generate between 250,000 and 300,000 new trucks annually. “I don’t view it as aspirational,” he remarked, “I genuinely think we can achieve this.”

Investors have not warmly received Aurora’s vision for 2030. Shares have continued to decline following the company’s annual analyst and investor day on September 23. On Monday, shares fell by 12.42%, closing at $5.29.

However, investors have time to adjust their views, and according to Maday, the major “unlock” for Aurora is anticipated to begin in 2027, accelerating from that point onward. The company projects a leap from 200 driverless trucks by the end of 2026 to over 1,000 a year later.

Currently, Aurora runs what it describes as a transportation-as-a-service business — a proof-of-concept model that it aims to cap at around 500 trucks. It owns and operates the self-driving vehicles while charging clients, including Detmar Logistics, Hirschbach, McLane, and Werner, roughly $2 per mile, which incorporates a fuel surcharge.

This pricing aligns closely with the typical rates offered by other carriers. The significant transformation — and the real cost savings, according to Maday — is expected next year when Aurora shifts to a driver-as-a-service model. In this new framework, customers will purchase the self-driving trucks and pay Aurora a per-mile subscription fee for the autonomous technology, which is projected to be about $0.85. With this system, customers will own and maintain the trucks, while Aurora will look after the self-driving technology and its necessary hardware.

Removing the trucks from Aurora’s balance sheet is vital for scaling — and likely what investors are focusing on. The company anticipates reaching breakeven gross margins (where revenue will cover direct operational costs of running the trucks) on a run-rate basis in the first half of 2027 with approximately 500 trucks operational.

The next significant advancement is expected at the end of 2027 with Aurora’s third-generation hardware — the sensors, computers, and other equipment enabling its trucks to operate autonomously — which will be mass-produced autonomous vehicle hardware developed by its partner, Aumovio (formerly known as Continental). Aumovio is not only engineering and fabricating the hardware kit; they are also financing it for Aurora — alleviating some financial pressure off the autonomous truck company. Additionally, Aumovio will provide servicing and repairs for the kits to customers.

Simultaneously, Aurora intends to expand its operations. By 2030, the company plans to extend its reach beyond just a handful of states in the South to cover most of the continental U.S., as stated by Maday.

“By 2028, I expect that our cost structures will be exceptionally strong, which is why you see our gross margin anticipated to rise significantly …” Maday noted. “Once we reach that stage, I believe entering the ride-hailing sector will be appropriate,” he confirmed, reiterating that Aurora still aims to eventually venture into the robotaxi market.

Modal Labs, a provider of AI inference infrastructure, is close to securing a $750 million funding round led by Accel, with a valuation reaching $15.75 billion including this investment, as per a source familiar with the funding process. The specifics of the round have not been reported before, although Axios and Bloomberg have provided other information regarding the deal.

This new funding will more than triple Modal’s valuation from the $4.65 billion it achieved when it announced a $355 million fundraising just four months prior.

Modal Labs opted not to provide comments.

The agreement occurs against a backdrop of increasing demand for inference services, which involve executing an AI model that is already trained to produce outputs, especially from clients using open-source models. Other startups focused on inference are also negotiating for new investments at much higher valuations. Baseten is approaching a capital raise at a valuation of $26 billion, which would be double its worth in June, according to Bloomberg. Additionally, Fireworks and Fal, a startup offering inference for video and image generation, have been in discussions with investors regarding new rounds that would significantly elevate their valuations, as reported by The Information.

Despite rapid revenue growth for these companies, their profit margins are slim, largely due to the high costs associated with acquiring or leasing computing resources. Fireworks revealed in July that its annualized revenue reached $1 billion, a fivefold jump from the previous year. Multiple startups focused on inference are anticipated to achieve this revenue milestone by year’s end, per our source.

Founded in 2021 by CEO Erik Bernhardsson and CTO Akshat Bubna, Modal has roots in significant industry experience. Bernhardsson, a Swede, spent over 15 years developing data teams at firms such as Spotify, where he contributed to the creation of the music-streaming service’s recommendation system, and Better.com, an online mortgage lender, where he was the chief technology officer. Bubna has a background in math and computer science from MIT and was an early staff engineer at Scale AI, the data-labeling startup, prior to co-founding Modal.

Based in New York and employing around 150 people, the company enables developers to train AI models and execute other resource-intensive workloads without the need to manage their own servers. Its website features customers such as the coding company Cognition, the AI music generator Suno, the fintech firm Ramp, and the publishing platform Substack.

As of May, Modal reported to Reuters that it had exceeded $300 million in annualized revenue.

The discussions for fundraising come two months after Modal became involved in one of the AI industry’s most scrutinized security incidents. In late July, Modal revealed that a customer’s data was compromised during the same hacking campaign conducted by a rogue OpenAI agent against Hugging Face.

Modal’s Chief Technology Officer Akshat Bubna stated that the breach was linked to a vulnerability in the customer’s own code, not Modal’s systems. “We’re aware a Modal customer published an unauthenticated endpoint that allowed anyone on the internet to use their sandboxes for code execution,” Bubna noted in a statement to media at that time. “This was exploited by the rogue agent. Modal’s platform was not compromised in any way,” he added.

AMD is set to acquire World Labs, a top creator of deep learning models designed to comprehend physical reality, in a $8.2 billion agreement, the two firms announced today.

World Labs explained the rationale behind the deal in a statement, indicating that the development of AI necessitated “intensive collaboration across model research, systems, and computing.” AMD, conversely, asserts that grasping cutting-edge workloads, such as those generated at World Labs, will influence its chip production strategy.

The acquisition will lead to World Labs founder Fei-Fei Li assuming the role of executive vice president and chief scientist at AMD. AMD and World Labs established an inference optimization and training collaboration last year, maintaining a strong connection since then. Significantly, Li participated as a guest at AMD’s CES presentation earlier this year.

Li, a computer science professor at Stanford, is lauded as a trailblazer in AI, especially in computer vision, due to her role in creating the ImageNet database and the AI competitions it incited. In 2024, Li launched World Labs to create deep learning models with a deeper comprehension of the physical realm, advocating that true general intelligence necessitated a grounding in physics and the capability to interpret and reason about information beyond text.

In a communication announcing the deal, Li described the collaboration as stemming from the ambition to expand World Labs’ technological innovations beyond the laboratory. “Now that we have tangible evidence of the possibilities, we aim to do everything possible to expedite the future,” Li wrote in the communication. “Accomplishing this requires amplifying our initiatives, broadening our impact, and moving closer to the hardware.”

“World model” is still an ambiguously defined concept, covering a range of applications from language models trained to interpret visual data, to models adept at generating and maintaining a high-fidelity representation of reality. World Labs’ inaugural product, Marble, is marketed as a solution for crafting entertainment experiences, as well as for generating simulated environments for training robots.

The acquisition is expected to boost AMD’s competitiveness against long-time adversary Nvidia in the development of a specialized AI chip ecosystem. While Nvidia has already launched a range of open-weight world models like Cosmos, AMD has thus far made available only text- and video-based models to the public.

World models are regarded as crucial in efforts to implement generative AI models on robotic systems, from self-driving cars to industrial robots and versatile humanoid machines. The scarcity of practical real-world data for training general-purpose robots specifically implies that synthetic data from world models will be essential to achieving the vision presented by companies like Tesla and Figure.

The acquisition is projected to finalize before the year’s end, pending regulatory approval.

While certain retailers, such as Amazon (and Adidas, it seems!), are preventing AI agents from making transactions on their platforms, e-commerce site Shopify has opted for the opposite approach.

On Monday, the firm revealed that browser-based AI agents are now able to finalize purchases on Shopify merchant websites, expanding their functionalities beyond merely searching for items and adding them to shopping carts.

Shopify had already endorsed WebMCP for its front-end stores and carts, enabling browser-based AI agents to navigate a Shopify vendor’s stock, look for items, and incorporate them into a cart. The newly added WebMCP support for checkout, which encompasses Shop Pay, allows these agents to interpret the checkout interface, modify it, and finalize the transaction with the purchaser’s consent, without depending on screenshots or web scraping, according to the company.

This enhancement brings forth three new tools — get_checkout, update_checkout, and complete_checkout — which permit agents to examine a checkout, adjust elements like the customer’s delivery address or shipping choices, and subsequently finalize the order after receiving the buyer’s permission.

The feature is currently being deployed to all qualifying Shopify merchants, as indicated by Gil Greenberg, a product manager focusing on agentic commerce at Shopify, in a post on X.

Shopping with an agent shouldn’t feel like watching paint dry. 🥱
Today, we’re launching WebMCP support for checkout, including Shop Pay, for all eligible Shopify merchants.

Learn how it works, where UCP fits, and what the data shows: https://t.co/U96VtJaeyc

— Gil (@gilgNYC) September 28, 2026

Shopify already operates a managed Model Context Protocol (MCP) server, enabling agents to function server-to-server. The proposed standard WebMCP, on the other hand, is tailored for agents working within the buyer’s browser. Both utilize Shopify’s Universal Commerce Protocol (UCP), providing a unified method for searching for and finding products, creating carts, and completing checkouts.

Leading AI agents such as Muse and Instinct have already established direct collaborations with Shopify for agentic commerce. The partnership with Instinct was announced today.

“If your agent is operating in the buyer’s browser, utilize WebMCP tools available on storefront and checkout to effectively execute order placement, instead of navigating HTML designed for humans,” Greenberg noted on X. “These WebMCP tools offer structured and efficient APIs, intentionally crafted — via UCP — to guarantee precise commerce facts, necessary disclosures, and handoff requirements.”

Tesla is once again delaying the unveiling of its revamped, second-gen Roadster, this time due to a prediction of severe weather.

Initially set for October 1 in Waco, Texas, the event has been rescheduled to October 15, according to Tesla.

“We’ve been monitoring the weather closely with local meteorologists, but in light of the predicted severe conditions & considering this event must be held outdoors, we’ve made the tough decision to postpone,” the company stated in a post on X.

The event likely requires an outdoor setting because Tesla aims to incorporate cold gas thrusters from SpaceX intended to enable the Roadster to achieve flight in some manner.

This reveal has faced multiple delays in the past, with Musk pushing it back for months leading up to this date. At one point, Musk even suggested holding the event on April Fools’ day this year. He mentioned during Tesla’s annual meeting in 2025 — the same meeting where he received a $1 trillion pay package — that hosting the event on April Fools’ day would provide him “some deniability” should it be postponed again.

“Like, I could claim I was just joking,” he commented at that time.

Tesla first presented its concept for a second-generation Roadster in 2017 during an event where it launched the Semi, the company’s electric big rig. The redesigned roadster was supposed to be the first supercar designed entirely by the company, as the original Roadster, which launched the company in the early 2010s, was predominantly based on the Lotus Elise.

At that time, Tesla assured that the new Roadster would be ready in just a few years. It gathered $50,000 deposits from would-be customers for a car expected to have a base price of around $200,000. Some individuals even paid the company $250,000 to secure one of 1,000 “Founders Series” versions of the supercar.

The project stalled for years until Musk reportedly assigned the Tesla team to redesign the second-gen Roadster and committed to the concept of using SpaceX thrusters.

The prospect of a sensational flying car hasn’t quelled critics of Musk and Tesla. Last year, OpenAI CEO Sam Altman posted on X that “7.5 years has felt like a long time to wait” for his own new Roadster, although Musk claimed Altman had received a refund.

Similar to the Roadster’s situation at this event, the shipping date for the vehicle remains uncertain. Musk himself has mentioned that he anticipates it will take a year or more before the new Roadster enters production after its eventual reveal.