a16z speedrun SR006 companies

Acceler8 is a San Francisco-based HR tech startup founded in 2025 that builds an AI-powered workforce intelligence and planning platform. It targets a core operational challenge for large enterprises: as AI adoption compresses organizational change that once took years into a matter of quarters, executives can no longer rely on annual workforce planning cycles to answer questions like which roles to automate, where to redeploy people, and who is best suited to step into critical positions. Traditional approaches — spreadsheets, consultants, and static HR systems — are slow, expensive, and often obsolete by the time they produce answers. Acceler8 appears to replace that episodic, manual process with a continuous, agent-driven system that gives leaders on-demand visibility into how their workforce should evolve as humans and AI agents increasingly work side by side. The company was co-founded by a second-time founder who previously scaled a venture-backed company to 400 people, and a former Chief People Officer and PwC workforce restructuring director with experience across 35 markets and 30 companies.

Advocate is a New York-based startup founded in 2025 that provides an AI-powered voice assistant designed specifically for older adults and their families. Seniors can dial a single phone number to request everyday help — such as transportation, grocery delivery, home repairs, or medical appointments — and Advocate handles each request from start to finish. Family members receive text updates when tasks are completed, and the platform proactively reaches out to seniors as well, so changes in routine or potential concerns don't go unnoticed. All activity is logged to a shared record accessible to the whole family, reducing the coordination burden that typically falls on one individual caregiver. The team of four brings together backgrounds in management consulting, cloud AI, and professional senior care.

Alike is a San Francisco-based startup, founded in 2025, that builds coordination infrastructure for teams running multiple AI agents simultaneously. As autonomous agents become capable of producing work at machine speed, the company argues that existing collaboration tools — designed around human communication cadences — create bottlenecks and errors when applied to agent workflows. Alike provides a shared coordination layer that lets agents exchange context, follow organizational rules, and maintain a live picture of ongoing work without human-mediated handoffs. The platform sits beneath tools teams already use, such as Claude Code, Codex, Cursor, Slack, Linear, and GitHub, rather than replacing them. The team of four, with backgrounds spanning San Francisco and UK institutions, appears to be targeting early-adopter engineering and product teams that are already running agent-heavy workflows.

Amdahl, founded in 2025 and based in San Francisco, builds an AI-native context layer designed to make enterprise go-to-market data usable by software agents. Rather than letting sales, customer success, and support teams work from fragmented or poorly structured data, Amdahl ingests and organizes that information into a shared, queryable corpus that AI agents can draw on to research deals, draft outbound messages, prepare for calls, and identify what drives revenue. The platform exposes a single API through which multiple specialized agents — for outbound, research, pre-call preparation, and more — can retrieve cited, deal-grounded context and share learnings with one another, so that insights one agent surfaces become available across the entire fleet. The five-person team appears to have roots at companies such as Databricks and Coinbase and has attracted early commercial interest from well-known consumer and fintech companies.

Antihero Studios is a Barcelona-based mobile game studio founded in 2024 by a team of roughly ten developers who previously held lead roles on titles such as Clash Royale, Brawl Stars, and Candy Crush. The studio's first title, MISFITZ, appears to be the first extraction game built natively for mobile: up to 16 players drop into a shared map, form ad-hoc alliances with strangers, collect loot, and then decide whether to extract together or betray one another at the gate. Early testing has drawn over 83,500 pre-alpha playtesters, with an average daily session length of roughly 50 minutes. Beyond the game itself, the team is working to build Antihero Studios as a recognizable player-facing brand, supported by a creator program that already counts more than 800 enrolled creators, with the goal of compounding organic reach across future titles.

August is a Tel Aviv-based fintech company, founded in 2025, that builds AI-powered software for commercial banking operations. The platform targets core banking workflows such as underwriting, loan servicing, and risk management, replacing or augmenting manual processes with domain-specific AI agents. As the system is used, it appears to build a proprietary context layer that grows more capable over time. August was founded by a three-person team of veterans from Israel's Unit 8200 intelligence corps, who previously developed AI systems in a military context and are now applying that expertise to the financial sector. The company appears to have already deployed its platform at a major bank with over $170 billion in assets.

Auto is a Los Angeles-based startup, founded in 2025, that has built an AI-powered camera application for iOS. The app allows users to photograph everyday objects and situations — a meal, an outfit, a plant, a restaurant bill — and automatically surfaces a relevant "Frame," which is a small, single-purpose mini-app that takes a specific action on the image. For example, a photo of lunch might trigger a calorie-tracking Frame, while a photo of an outfit might trigger a style-rating Frame. Users can also build and share their own Frames, creating a community-extensible layer on top of the core camera experience. The three-person founding team comes from backgrounds in camera and creator products at Snapchat and appears to bring significant prior startup experience to the project.

Belong is a New York-based fintech and creator-economy startup, founded in 2025, that builds artist-owned membership programs backed by real financial infrastructure. The platform connects fan identity to financial rails, allowing fans to be recognized and rewarded for their long-term participation and spending. Rather than treating fandom as purely a social or attention-driven phenomenon, Belong attempts to make it measurable and financially meaningful — for both artists and their most dedicated supporters. The company was co-founded by Nick Holmsten, formerly Spotify's Global Head of Music, and Ash Pournouri, who managed and developed Avicii into a globally recognized artist.

Bilrost is an Oakland-based AI-native infrastructure company founded in 2025 that targets the operational bottlenecks inside commercial lending. The platform converts the documents behind a loan deal — offering memoranda, tax records, rent rolls, financials, and borrower history — into a structured context graph that supports the entire credit lifecycle, from initial intake through underwriting, servicing, and ongoing portfolio monitoring. Rather than treating document processing as the end goal, Bilrost focuses on making extracted information accurate enough to act on and traceable enough to satisfy risk and compliance teams. The company reports having processed over 10,000 deals and is actively piloting its platform with commercial lenders. With a team of five, it appears to be in an early but traction-bearing stage.

Bota is a Mountain View-based startup, founded in 2026, that provides infrastructure for capturing and processing real-world, offline conversation data for AI agents. The company offers a hardware-plus-API platform that allows developers to add offline meeting and phone-call recording capabilities to their products—using either existing devices or Bota's white-labeled wearable hardware. Once conversations are captured, Bota handles all hardware logistics (manufacturing, shipping, support, and returns) and delivers recordings, transcripts, and metadata through APIs that plug into transcription, summarization, search, or other AI-driven workflows. Bota appears to position itself as the foundational layer connecting AI agents to real-world, in-person context—a gap that online-only data pipelines currently leave unfilled. The four-person team brings prior experience in hardware development and cloud infrastructure, and the company reportedly grew to over $700K in contracted ARR within its first five weeks of selling.

Cascade is a San Francisco-based startup, founded in 2025, that builds AI-powered software for architecture, engineering, and construction (AEC) firms. The platform helps these firms identify relevant project opportunities and compete for them more effectively. On the discovery side, Cascade continuously scans thousands of government portals—at the state, local, and federal level—to surface request-for-proposal (RFP) listings that match a firm's track record and capabilities, scoring each opportunity so teams can make go/no-go decisions quickly rather than spending days reviewing portals manually. On the win side, the platform drafts proposals, analyzes project specifications, and flags scope inconsistencies, compressing a process that typically takes weeks into a matter of days. Future capabilities appear to include pre-bid intelligence drawn from budget announcements and government meeting minutes, as well as private-market signals such as land purchases and permit filings. The two-person founding team brings backgrounds in machine learning at Google Search, fintech scale-up, and applied AI in public-sector contexts.

Cedar is a San Francisco-based startup, founded in 2025, building an AI-native professional network with a focus on recruiting. The core insight behind the product is that the strongest candidates rarely apply through job postings — they are typically reached through trusted referrals. However, referral programs tend to break down as organizations grow, not because employees lose their networks, but because manually activating those relationships at scale becomes unmanageable. Cedar addresses this by deploying AI agents that conduct short structured calls with employees to map their trusted professional relationships, surface relevant candidates, and continuously enrich a proprietary talent graph. The network effect compounds over time: each new customer and each additional call makes the underlying graph more valuable. Cedar appears to position itself as an invitation-only, high-signal alternative to incumbent professional networks, serving companies that prioritize quality of hire over volume of applicants. The company was co-founded by Greg Gunn and Beier Cai, who previously scaled Hootsuite from a small team to over 1,500 employees and from zero to $100M ARR, giving them firsthand experience with both the power and the fragility of referral-based hiring at scale.

Clair Health, founded in 2025 and based in Mountain View, California, is developing what it describes as the first wearable device capable of continuously monitoring female hormones — including estrogen, progesterone, LH, and PdG — without blood draws. The product takes the form of a jewelry-inspired wrist wearable that captures over 130 biomarkers through a multi-modal sensor array, which feeds into a proprietary female biology model to generate personalized hormonal insights. Those insights appear aimed at helping women navigate a range of health areas, including fertility planning, skin health, mood, athletic performance, and perimenopause. The company is a small team of seven and appears to be in the pre-shipment phase, with founding member devices expected to ship in late 2026.

Coalition Systems is a San Francisco-based startup, founded in 2025, building a secure operating system for national security coordination. The company addresses a structural gap in defense software: military units, intelligence services, government agencies, and critical infrastructure operators currently rely on consumer-grade tools — such as messaging apps and spreadsheets — that were never designed for operational security or cross-organizational coordination. Coalition OS appears to provide a purpose-built platform for communication, collaboration, and decision support across these interconnected nodes, designed to function under field conditions and on field-appropriate devices. The founding team brings together expertise in AI engineering, critical infrastructure, and U.S. Department of Defense intelligence, and the company has attracted advisors with senior military and defense intelligence backgrounds.

Concorda is a New York-based legal technology startup, founded in 2025, that is building an AI-native operating system designed specifically for litigation. Rather than layering AI features on top of existing legal workflows, Concorda appears to embed AI directly into the end-to-end process of running a litigation matter — from drafting complaints and managing discovery to coordinating positions across multiple parties. The company was founded by a small team of lawyer-engineers with backgrounds spanning Harvard computer science, UChicago Law, and Northwestern Law, giving the team direct experience with the practical limitations of current legal tools. Concorda targets practice areas including IP litigation, plaintiff's litigation, commercial litigation, and in-house legal departments.

Crebit is a San Francisco-based fintech startup, founded in 2025, that addresses a core friction point in stablecoin-based foreign exchange: the short expiry window of blockchain exchange rate quotes. Because on-chain quotes typically expire within minutes, businesses have historically been unable to commit to stablecoin FX for predictable, forward-looking payment obligations. Crebit built infrastructure that allows enterprises to lock in blockchain exchange rates for periods of up to 180 days, bringing a level of rate certainty to stablecoin FX that previously only existed in traditional currency markets. The company appears to be gaining traction with enterprise customers and is processing meaningful monthly payment volume, with signed clients representing substantial projected annual FX flow. Crebit was co-founded by two engineers with academic roots at Stanford and MIT, and professional backgrounds spanning major technology and government organizations.

Emanate is a San Francisco-based AI software company, founded in 2026, that builds autonomous revenue operations tooling specifically for industrial materials businesses. Its platform handles the full commercial workflow—processing inbound inquiries, generating quotes, identifying account expansion opportunities, prospecting new customers, and informing pricing decisions—without requiring constant human intervention. The company targets a segment of the economy that has historically been underserved by modern sales technology: manufacturers and distributors of physical materials that rely on complex, high-volume transactional processes to grow revenue.

Grove Tax is a San Francisco-based fintech company, founded in 2025, that builds AI-powered workflow automation software for tax preparation firms. The platform covers the entire lifecycle of a tax return through a structured six-stage pipeline: reading prior-year returns to identify what has changed, generating a personalized document checklist for each client, running an adaptive AI-driven questionnaire in place of static intake forms, organizing and labeling collected documents into smart workpapers, drafting the return automatically for preparer review, and finally delivering a branded summary with e-sign and payment in one place. The tool appears to address a well-documented capacity problem in the tax industry, where a large share of practitioners spend the majority of their time on administrative tasks rather than actual tax work, and where the supply of qualified preparers has been shrinking steadily. Early design-partner firms using Grove have reportedly processed returns roughly three times faster than before.

Hammock is a Brooklyn-based fintech and AI startup, founded in 2025, that helps consumers get more value out of their Health Savings Accounts (HSAs) and Flexible Spending Accounts (FSAs). The platform uses an AI agent to scan a user's purchases and identify eligible healthcare and wellness expenses that may have been overlooked. Through a network of partnered physicians, Hammock can generate Letters of Medical Necessity, which allows everyday wellness spending—such as gym memberships, massage therapy, and recovery technology—to qualify for HSA/FSA reimbursement. The result is a largely automated process that reduces the paperwork and complexity typically associated with these tax-advantaged accounts.

Heavi is a San Francisco-based startup founded in 2025 that is building an AI-powered workforce for the heavy-duty vehicle repair and service industry. The company targets a market where a growing shortage of experienced mechanics — driven by veteran technicians retiring and taking decades of specialized knowledge with them — leads to longer downtime, parts sourcing delays, and significant revenue losses for fleet operators and repair shops. Heavi's platform appears to act as an AI agent that handles parts identification, supplier sourcing, pricing comparison, and order coordination on behalf of shop technicians, drawing on OEM schematics, aftermarket catalogs, and live vendor outreach. The company reports early traction with five design partners, contracted revenue, meaningful reductions in employee time spent on parts-related tasks, and improved shop margins.

Hotbox is a San Francisco-based AI platform founded in 2025 that helps brands selling high-demand, limited-availability products convert social media engagement into actual purchases. Rather than letting purchase intent get buried in Instagram comments, story replies, and DMs, the platform surfaces the highest-intent followers, matches them against a seller's ideal customer profile using enrichment data, and then autonomously moves those prospects through a checkout experience directly inside their DMs. The company appears focused on verticals where scarcity and fan excitement drive buying behavior, including sports jerseys, trading cards, collectibles, tickets, and limited-edition brand drops.

Idilio is a Bogotá-based mobile streaming platform founded in 2025 that specializes in short-form, serialized drama produced specifically for Spanish- and Portuguese-speaking audiences in Latin America and beyond. The app delivers original "microdramas" — vertical, one-to-three-minute episodes written by telenovela veterans and shot in a format native to how people already consume content on their phones. With over 1.3 million downloads, 29-plus original series, and reach across 50 countries, Idilio appears to be the leading microdrama platform targeting the Latin American market. Beyond content distribution, the company is building AI-assisted production infrastructure and proprietary audience data tools designed to scale original storytelling output efficiently.

Kaaro is a San Francisco-based startup building AI agents designed to automate the operational work that keeps railroads running day to day. Tasks such as paperwork, data entry, incident reporting, billing, and coordination workflows that currently travel through phone calls, emails, and institutional memory are the core targets. The two-person founding team brings direct industry familiarity — both founders are second-generation railroaders — as well as hands-on engineering experience scaling an AI voice product at an early-stage startup. Kaaro appears to be in the early commercial stage, already engaged with CPKC, Canada's largest railroad, in late-stage conversations with CN, and actively building alongside Gulf & Ohio Railways, a group of four short-line operators.

Loops AI is a San Francisco-based startup, founded in 2025, that provides an intelligence layer designed to help e-commerce brands make their product catalogs and commerce operations legible to AI-powered systems. The platform centers on structured, AI-readable product data and deploys agents across use cases such as AI search, virtual try-on, omnichannel selling, and product data enrichment. By interpreting shopper intent at multiple touchpoints, Loops AI appears to help brands increase conversion, reduce product returns, and maintain consistent engagement across channels.

Miraka is a New York-based cardiac care company founded in 2025 that combines continuous remote monitoring with clinician-led intervention to support heart patients between hospital visits. Using data from wearables and electronic health records, Miraka's AI models identify emerging risk signals and trigger timely clinical responses before a patient's condition escalates to an emergency. The platform appears to be staffed by real clinicians who manage personalized care plans—covering everything from exercise therapy to medication adjustments—delivered entirely from the patient's home. Miraka reports enrolling the vast majority of eligible patients referred by partner hospitals, far above typical industry enrollment rates, and has achieved a substantial reduction in 30-day readmissions among its surgical patient population.

Mirror Mirror AI is a San Francisco-based startup, founded in 2024, that operates a two-sided marketplace connecting fashion models with brands seeking AI-generated campaign imagery. Models on the platform proactively license their likeness under agreed terms, and brands can browse a curated roster, select a model, and produce lookbooks, e-commerce shots, or full campaign visuals in a matter of hours rather than weeks. Every image generated is tied to a consent and compensation agreement, meaning the model earns a fee each time a brand licenses content featuring their likeness. The platform appears to already include agency-represented models with high-profile editorial credits. The company was co-founded by Yusan Lin, a fashion model with a computer science PhD background, positioning the product at the crossroads of the fashion industry's appetite for faster content production and the growing advocacy among models for fair compensation when their image is used in AI-generated work.

Modaic is a San Francisco-based startup building infrastructure that makes AI-driven decisions more reliable in production environments. The company focuses on judgment-heavy tasks such as content classification, moderation, and automated evaluation — areas where teams have historically been forced to choose between costly full human review and brittle automated pipelines. Modaic's approach involves assigning a confidence score to each AI decision, selectively routing only low-confidence cases to human reviewers, and feeding that human input back into the system to refine its underlying instructions over time. The net effect is a decision-automation layer that becomes progressively more accurate while reducing the volume of manual oversight required to sustain it.

Modern Industrials is a New York-based startup, founded in 2025, that builds custom AI systems designed to modernize back-office and sales operations for distributors in the building materials industry — a sector estimated at over $450 billion. Its initial product automates labor-intensive tasks such as quoting, order entry, and business analytics, turning what has traditionally been a manual, error-prone back office into a faster and more accurate operation. The four-person founding team brings together engineering experience from machine learning and AI research roles alongside management consulting backgrounds. Longer term, the company appears to be working toward becoming a comprehensive operating platform for physical goods distribution more broadly.

Oasis is a New York-based startup building a collaborative workspace designed for teams that work alongside AI agents. Rather than requiring deep technical expertise to configure and manage multi-agent systems, Oasis provides a shared environment — organized around rooms, documents, and agent fleets — where both humans and AI agents can communicate, edit files, and complete tasks together. A core feature is persistent, team-owned memory: the platform retains context from every decision and task, so agents remain effective even when models are swapped or upgraded. Founded in 2026 by former Palantir Field Deployment Engineers with hands-on experience shipping agent systems, the five-person team appears to be targeting the emerging category of human-agent collaboration tools within the broader future-of-work space.

Oasiz is a San Francisco-based startup founded in 2025 that is building a social distribution platform for AI-generated software, starting with games. The core premise is that as AI tools make software creation dramatically easier, software itself is becoming a form of content — yet existing distribution channels were built for a different era. Oasiz aims to be the destination where AI-native creators publish, share, and grow an audience around their creations, much like short-video platforms reshaped how video creators find viewers. Within weeks of launching, the platform had accumulated over 9,000 hours of user engagement and surpassed 300,000 plays. The company has also secured a partnership with Atari to bring classic gaming IP into this new AI-built, socially distributed format. The four-person team draws on backgrounds from Stanford, Berkeley, Google, Tesla, Nintendo, Amazon, and Superhuman, and the founders previously reached the number-one spot on the App Store's social charts with a prior product.

Omi Health is a New York-based startup founded in 2025 that describes itself as the world's first dog longevity platform. The company offers at-home blood and saliva testing that measures 80+ biomarkers, detects over 1,000 conditions in dogs before symptoms appear, and uses AI to generate personalized health protocols for each pet. The core premise is that preventative, data-driven healthcare—already popular among health-conscious human consumers—should extend to pets, giving owners the tools to catch disease early and potentially extend their dogs' healthy lifespans. Developed by two co-founders who built the product for their own dogs, Omi appears to be targeting a generational shift in how younger pet owners think about animal wellness. Since its March 2025 launch, the platform has attracted over 4,000 waitlisted pet parents and operates on an invitation-only early-access basis.

Panorama is a San Francisco-based AI platform, founded in 2024, that helps enterprises understand and improve how their teams actually work with AI tools. It does this by analyzing existing workplace data sources — such as Slack messages, emails, and documents — to map out the informal workflows that teams have already developed. From that analysis, it surfaces personalized recommendations for AI adoption, delivered through a feed-style interface designed to reduce friction and encourage organic uptake across an organization. The four-person founding team brings together experience in large-scale technology adoption at major consumer fintech and ride-sharing platforms, alongside deep academic and applied machine learning expertise.

PartyHat is a very early-stage startup founded in 2025 and based in San Francisco, California. With a two-person team, the company appears to be in the initial stages of building its product or service. Based on the name "PartyHat," the company likely operates in the events, social experiences, or entertainment technology space, though the specific focus remains unclear given the absence of publicly available information.

PayPath is a New York-based fintech company, founded in 2025, that builds AI-powered infrastructure for the debt management industry. The global debt market has historically relied on manual processes, spreadsheets, and aging software systems; PayPath aims to replace that stack with agentic AI that automates the full debt lifecycle—from client enrollment and document collection to ongoing account management and resolution workflows. The platform allows human agents to stay involved where needed while delegating routine tasks to AI. With a small team of three, the company has reportedly reached $2M in annual recurring revenue within its first year and has over half a billion dollars in assets being managed across more than a dozen enterprise customers.

PicPet is a San Francisco-based consumer social app, founded in 2025, that ties everyday photo sharing directly to the care of a virtual pet. Users keep their in-app pet alive and thriving by sending photos to friends, making the act of staying in touch a gameplay mechanic rather than a passive feed. The product sits at the intersection of social messaging and casual gaming, and it appears to have gained meaningful early traction, with over 240,000 daily active users and a 90-day bounded retention rate above 45%—metrics that suggest strong habitual engagement for a product launched so recently. The company operates under the parent name O3O Labs and has a team of six.

Piper is a San Francisco-based construction technology company, founded in 2025, that builds an AI-powered operating system for preconstruction teams at general contractors. The platform ingests and connects project documents — drawings, specifications, addenda, subcontractor quotes, estimates, and schedules — into a single, continuously updated project understanding. It layers that live project context with a firm's own historical cost data, templates, and standards to automatically generate scope documents, bid leveling sheets, estimates, and project reviews, with every output traceable back to its source document. The goal is to shift preconstruction teams away from manual information assembly and toward higher-value decision-making work, giving senior staff broader visibility across more projects than they could personally review on their own.

Pluvo is a San Francisco-based company, founded in 2025, that builds an AI-powered operating platform designed specifically for corporate finance teams. Rather than replacing individual tools, it connects a company's existing financial systems into a single governed data layer, then deploys AI agents to carry out recurring tasks such as variance analysis, monthly close reconciliations, forecasting, and board-level reporting. Every output the platform produces can be traced back to its underlying source data, calculations, and actions, giving the humans responsible for those numbers a clear audit trail they can verify and stand behind. The platform is designed to work with multiple AI model providers, so teams are not dependent on any single vendor as the technology evolves. With a team of 14, Pluvo appears to be targeting the growing demand among finance functions for automation that is reliable and explainable rather than opaque.

Prior Foundry is a San Francisco-based startup building a decision intelligence platform aimed at government agencies and public policy teams. The platform combines data integration, scenario modeling, simulation, and stakeholder input to give policy teams a continuous, evidence-based view of how their programs are performing — from initial rollout through long-term outcome measurement. Rather than relying on periodic reports or one-off analyses, governments using the platform can monitor implementation in real time, identify where programs are falling short, and model alternative approaches before committing to changes. Over time, the system accumulates institutional context and evidence, making cross-program comparisons and policy learning more practical.

Quanto is a San Francisco-based startup founded in 2024 that builds AI-powered workflow automation tools specifically for accounting firms. The platform covers the full client lifecycle — from identifying and qualifying prospective clients, to booking discovery calls, to onboarding new clients into QuickBooks and cleaning up their books to meet a firm's internal standards. Designed by CPAs with accountants in mind, Quanto aims to help small and mid-sized accounting practices grow their client base and improve margins without needing to hire additional staff. The company appears to have gained early commercial traction, with a reported contracted ARR and a growing sales pipeline.

Quinn is a New York-based AI startup founded in 2025 that builds training and operational readiness software for frontline workers — a global population estimated at 2.7 billion people in trades and similar hands-on industries. The company's core premise is that this workforce has historically relied on expensive, human-delivered training programs and largely lacks access to modern, scalable learning infrastructure. Quinn appears to use AI agents to deliver that infrastructure as software, bringing the kind of structured training tooling that desk-based workers take for granted to workers in fields such as construction, manufacturing, and skilled trades. The team of ten has grown the product quickly since founding, reaching over 65 customers and crossing $1 million in annual recurring revenue within roughly five months of launching.

Quo Labs, operating under the product name "Sam," is a San Francisco-based startup founded in 2025 that builds voice-first companion devices designed for older adults. The product combines a conversational AI assistant with daily wellness check-ins, cognitive monitoring through games embedded in natural conversation, medication and appointment reminders, and a controlled calling and messaging system that limits contact to pre-approved family members. The underlying goal is to help seniors maintain independence at home while giving their families a continuous, structured view of their loved one's mental and physical wellbeing. With an estimated 1.4 billion seniors worldwide, the company is positioning itself as foundational infrastructure for the eldercare space rather than a single-use gadget.

SafeWorld is a Palo Alto-based startup, founded in 2025, that builds a safety testing and evaluation platform for robotic systems that operate in close proximity to people. The platform is designed to surface rare but high-consequence failure scenarios, giving robotics teams a structured way to measure and understand the risks their systems pose before and during real-world deployment. The company's founding team combines academic expertise in robot safety research with hands-on experience building and scaling AI companies, positioning SafeWorld at the intersection of physical AI, cybersecurity, and AI infrastructure.

Sentra is a San Francisco-based enterprise software company, founded in 2025, that builds a managed organizational memory layer for companies. It continuously captures activity from communication and collaboration tools — such as meetings, chat threads, email, code repositories, and deal data — and organizes that information into a structured, permissioned context store. The goal is to give both employees and AI agents a coherent, real-time understanding of what is happening across the organization, including who made decisions, why those decisions shifted, and what commitments are still open. By connecting information that would otherwise sit in isolated tools, Sentra appears to reduce coordination overhead and improve the quality of AI-driven workflows that depend on accurate organizational context.

Simula is a San Francisco-based adtech startup, founded in 2025, that builds advertising infrastructure purpose-built for AI-powered applications. Rather than porting conventional banner or interstitial formats into AI products — where they tend to disrupt the user experience — Simula designs ad units that blend naturally into AI-driven interfaces. These include interactive formats such as AI brand ambassadors and games played alongside AI companions. The company appears to be integrating its technology across a growing set of AI apps and is targeting a user base in the tens of millions of daily active users. The five-person team brings backgrounds spanning app development at scale and roles in the advertising industry.

Sirius is a San Francisco-based startup founded in 2025 that builds AI-powered retention software for subscription businesses. The core problem it addresses is early subscriber drop-off, a pattern where a large share of customers cancel within the first few months of signing up. To tackle this, Sirius has developed a proprietary AI model designed to predict and prevent churn by managing customer interactions at scale. The platform appears to already be in production, handling millions of interactions daily on behalf of consumer companies with significant market scale.

Smart Bricks is a San Francisco-based applied AI laboratory, founded in 2024, that builds autonomous reasoning systems designed to modernize real estate investing. The platform deploys a suite of AI agents that handle distinct stages of the investment lifecycle: scanning markets for opportunities, conducting institutional-grade underwriting across a large set of variables, coordinating due diligence and transaction execution, and continuously monitoring portfolio performance. The core premise is that real estate — despite being the world's largest asset class — remains fragmented and opaque, and that purpose-built AI models can turn raw, unstructured property market data into structured, decision-ready intelligence. With a team of roughly 20 people drawing on backgrounds spanning institutional finance and frontier AI research, Smart Bricks appears to target the inefficiencies that have historically kept real estate returns below those of other major asset classes.

Snag is a New York-based marketplace for short-term and month-to-month rentals, targeting younger renters who currently navigate fragmented, informal channels such as social media groups and classified sites to find sublets. Founded in 2025 by repeat consumer founders Selin Sonmez and Niko Georgantas, the four-person team has built a platform that sits between short-stay vacation rentals and long-term apartment listings — a segment that has historically lacked structure and trust. Snag's core technology appears to use AI to automatically convert unstructured social posts into clean, standardized listings, and to match renters with suitable options in near real time. The platform has reportedly accumulated around 3,000 listings in New York City and processed approximately $6.5 million in monthly rental requests, growing roughly 40% month over month.

Snapp AI is a San Francisco-based sports technology company, founded in 2025, that has built an AI-powered sports analyst accessible directly through a messaging interface. Rather than requiring fans, bettors, and fantasy players to jump between multiple sites and apps, Snapp consolidates real-time data across more than 50 leagues — covering everything from the NFL and NBA to Formula 1, cricket, and esports — into a single conversational AI. The platform connects to live scores, injury reports, betting odds, player statistics, and schedules, then delivers personalized insights based on what each user actually follows. It appears to be the work of the founder of Caviar, applying a similar on-demand convenience model to sports information and analysis.

Sparta is a San Francisco-based infrastructure company, founded in 2025 by Berkeley computer science alumni, that builds an intelligent data delivery layer for distributed computing environments. Instead of relying on fixed transfer settings, Sparta uses a real-time reinforcement learning engine that continuously adjusts to live network conditions, optimizing how large volumes of data move across clouds, clusters, and distributed systems. The platform appears to target the growing inefficiencies caused by modern AI workloads, high-throughput streaming, and data-intensive applications, where slow or unpredictable data movement leads to wasted compute resources and higher operating costs. By slotting in on top of existing storage and transfer tooling without requiring infrastructure rewrites, Sparta aims to make data delivery a non-issue for organizations operating at scale.

Straia is a San Francisco-based startup, founded in 2025, that builds an AI-powered data and operations platform for colleges and universities. The platform connects fragmented institutional data sources — including student information systems, learning management systems, and CRM tools — into a unified environment, then deploys AI agents that can analyze that data and surface actionable insights. Campus leaders across functions such as enrollment management, academic affairs, and executive administration can query the platform in plain language to get answers about enrollment trends, student retention, course outcomes, financial performance, and more. Straia appears to be positioning itself as a modern replacement for the legacy software infrastructure that currently underlies most higher education institutions.

SUN is an AI-powered audio platform founded in 2025 and based in Palo Alto, California, that lets users generate fully personalized audio content — including podcasts, audiobooks, and structured listening experiences — on virtually any topic. Unlike traditional audio, which is fixed and one-size-fits-all, SUN's content is dynamic: users can adjust parameters such as length, voice, and language, and can ask questions mid-playback and receive answers in real time. The platform also connects with apps a user already relies on and continuously adapts its recommendations to that person's evolving interests, making it an ambient, ever-updating learning and listening companion. The company appears to sit at the intersection of consumer media, edtech, and generative AI, targeting people who want to learn or stay informed through audio but find existing offerings too rigid or generic.

Syncere is a Palo Alto-based consumer robotics startup founded in 2025 that builds home robots designed to blend seamlessly into everyday living spaces. Its debut product, Lume, is a floor lamp that conceals a pair of robotic arms; when a chore needs doing, the arms deploy to handle tasks such as folding laundry, then retract so the device returns to looking like an ordinary lamp. The team of eight draws on robotics research backgrounds from Stanford, UC Berkeley, and the University of Toronto, as well as hands-on hardware and design experience from companies including Tesla, Nuro, AMD, and Apple. Lume appears to be approaching its initial commercial launch, with manufacturing and shipping targeted for summer 2025, backed by hundreds of pre-orders and substantial waitlist demand driven largely by organic online attention.

Taxnova is a San Francisco-based fintech and AI company, founded in 2025, that automates the preparation of R&D tax credit and capital expenditure documentation for technology companies. Rather than waiting until year-end to reconstruct qualifying engineering activity, Taxnova connects read-only to the tools engineering teams already use — such as project trackers, code repositories, internal wikis, and messaging platforms — and continuously extracts and organizes evidence of qualifying work throughout the year. The result is a structured, source-linked claim package that a company's existing tax advisor can review and submit, with traceability designed to hold up under scrutiny from tax authorities in the US, UK, France, and Canada. The platform appears to surface qualifying projects that might otherwise be overlooked, and it eliminates the need for time-consuming engineer interviews or last-minute documentation sprints.

ThirdBrain Labs is a San Francisco-based AI infrastructure startup founded in 2025 that helps organizations build, own, and continuously refine domain-specific AI models. Rather than relying on general-purpose, closed AI APIs, teams can encode their proprietary knowledge directly into fine-tuned models they fully control. The platform handles the full post-training lifecycle—from capturing expert decisions in real time to deploying task-optimized models—without requiring large labeled datasets or extensive manual data-labeling pipelines. The result is a model portfolio that a team owns outright, including the weights and underlying data, and that improves continuously as new domain signals are captured in production.

Variant, founded in 2025 and based in Tel Aviv, Israel, builds real-time decisioning and content generation infrastructure that allows websites to automatically tailor what each visitor sees — including copy, imagery, product listings, and navigation flows — without requiring engineering or design work from the merchant's team. Rather than running traditional A/B tests on a fixed schedule, Variant's system continuously analyzes visitor behavior, identifies high-impact opportunities, generates brand-consistent experiments, and encodes the results into a persistent memory layer so that the platform grows more effective over time. The company appears to target e-commerce and other conversion-driven websites that want personalized, dynamically generated experiences for both human visitors and automated agents. Its six-person founding team draws on backgrounds in Israeli defense intelligence and special forces.

Vereda is a Brazilian startup founded in 2025 and based in Rio Verde, Brazil, that is building a direct purchasing channel between agricultural input manufacturers and farmers. The company targets a fundamental structural problem in Brazilian agriculture: a fragmented, multi-layered distribution chain that inflates the cost of inputs such as fertilizers and agrochemicals, compressing farm margins. By shortening the path between industry and farm, Vereda aims to give producers access to manufacturer-level pricing — reportedly up to 30% cheaper than traditional channels — while giving suppliers a more direct and cost-efficient route to reach dispersed rural demand across the country.

Verra, operating under the product name Meridian, builds an AI-powered platform designed specifically for management consulting firms. The core problem it addresses is that consulting firms accumulate vast amounts of institutional knowledge — past engagements, decks, models, and research — that typically sits in scattered, unstructured files and walks out the door when employees leave. Meridian acts as a centralized, searchable repository of a firm's entire body of work, and layers AI agents on top of it to help teams generate new deliverables directly from prior work rather than starting from scratch each time. The system enforces firm-specific templates, surfaces relevant precedents in seconds, and is deployed as a single-tenant instance inside the client's own cloud environment to satisfy enterprise security requirements. Founded in 2025 and based in San Francisco, the two-person team appears to have backgrounds at organizations focused on AI and national-security-grade software.

ZeroDrift is a New York-based cybersecurity and AI infrastructure company, founded in 2026, that built what it calls an enforcement runtime for AI systems. As organizations increasingly rely on AI agents to communicate and act on their behalf, those agents operate across regulated industries where strict rules govern what can and cannot be said or done. ZeroDrift sits between AI models and their outputs, evaluating every message or action in real time and either passing, rewriting, blocking, or escalating it before it reaches its destination. The platform comes pre-loaded with a broad set of regulatory frameworks and allows organizations to layer their own internal policies on top, generating audit-ready records for every enforcement decision. The company appears to be targeting heavily regulated sectors such as financial services, healthcare, and insurance, where the cost of a non-compliant AI output is particularly high. With a team of 12 that draws on backgrounds from organizations including Google DeepMind, Microsoft, and Goldman Sachs, ZeroDrift is positioning its runtime as foundational infrastructure for any company deploying AI at scale.