AI Companies for Investors to Watch

I’ve been waiting for this: Cerebras’ quiet IPO re-filing last week all but confirms they’ll be the first semiconductor company to go public this year! But I'll bet they won’t be the only one. The market is ready for a pure-play challenger to Nvidia’s dominance in AI compute, and they’re the clearest contender. Between their January partnership with OpenAI and a recent $1.1B raise, the decade-long mission they’ve been on has never felt more validated. While some debate whether an IPO is the ‘right’ outcome for a semiconductor startup, I see it differently. An IPO doesn’t make a company successful, it proves they already are, and IPO momentum is starting to build in the AI sector. I believe four other startups that have been diligently building their technology, teams, and ecosystems for years are also coming up soon. Here’s who I’m also watching to ring the bell this year: Lightmatter: The ‘Photonics Frontrunner’ If there’s one that I’d place my chips into the middle for, it’s Lightmatter. Many photonic startups have a strategic lever or two. Lightmatter has several: core IP, top-tier talent, key supply chain relationships, and a clear multi-year roadmap. They’ve been refining photonic interconnect technology for over a decade, and the market has massive potential (evidenced by Nvidia’s $4B investments in Coherent and Lumentum yesterday). This is the kind of profile public markets reward. Ayar Labs: The ‘AI Ecosystem Enablers' Ayar Labs just became the most institutionally validated optical interconnect startup enabling next-gen AI infrastructure when they announced a $500M Series E this morning. Led by heavyweights like ARK Invest, Sequoia Global, AMD, MediaTek, AlChip and Nvidia. If Lightmatter is attacking photonics at the processor level, Ayar is embedding optics into the infrastructure backbone itself. Different angles, same bottleneck, and public markets have loved companies solving systemic constraints. Axelera AI: The ‘Inference Accelerator’ Training has been AI’s biggest story, but adoption is exploding in inference and Axelera has been positioning for this moment. They’ve proven they can build from concept to deployment quickly, and with strong backing from European VCs, they have the capital, the team, and the customer traction to go the distance. When they layer in enterprise wins that will come with their Europa chip, I believe that signals broad market readiness/IPO. Tenstorrent: The ‘Strategic Wildcard’ Tenstorrent is on my radar for a different reason than the others. They’ve always had strong leadership and potent AI compute technology. What was missing was cohesion, a way to bring those elements together in a unified direction with a software stack to tie it together. Now that they seem to have those figured out, an IPO could mean strategic access to capital without further private dilution. What do you agree/disagree with above? And is there anyone I’m missing? #artificialintelligence #startups #semiconductorindustry

I’ve been impressed watching companies like Cursor, Lovable, ElevenLabs, Bolt, and Midjourney scale quickly. It’s a reminder of how AI tools can help small teams scale revenue fast—quietly proving what’s possible. These examples (below), leveraging AI and engaging ecosystems across X, LinkedIn, and other platforms, need less capital to scale. These companies offer a glimpse into the future of how software companies grow. It will be interesting to see how the VC industry adjusts to companies requiring less talent, and less capital to scale. Here is the summary: Cursor soared to $100 million in ARR in two years with 20-30 staff and $72 million in funding, revolutionising tools for developers. Midjourney, boasting $200 million to $300 million in ARR with a lean crew of 11-40, has taken the lead in AI-generated imagery, possibly with minimal VC input (up to $50 million speculated). ElevenLabs reached $100 million ARR with 50 employees and $80 million in funding, transforming voice synthesis for content creators. Lovable hit $10 million to $17 million ARR in mere months with 15 people and $6 million in seed funding. Bolt scaled to $20 million ARR in just two months with 15 staff and $10 million to $12 million in VC, likely shaking up commerce or payments.

Remember the investment atmosphere of the late ‘90s? The dotcom boom? PCs flew off the shelves, online access got cheap, browsers became mainstream. But the clearest signal? Investors doubled down. What have we seen in the last week? Similar perfect storm trends are alive in the #AI world; #infrastructure, #advertising, #voice, #security, #radiology, #schools, #dentists, #lawyers, #robots, #orchards, #sales, #seniors Here is a list of only some of the VC investments in AI in just the last week:* • DataBank, makes infrastructure for data centers, raised $250M • StackAdapt, a programmatic advertising firm, raised $235M • ElevenLabs, makes AI voice software, raised $180M • UVeye, uses AI to inspect cars, raised $150M • Semgrep, an AI-powered app security startup, raised $100M • Rad AI, develops AI software for radiology, raised $60M • Quibim, makes AI models for medical imaging, raised $50M • MagicSchool AI, generative AI software for schools, raised $45M • SafelyYou, AI software for senior living facilities, raised $43M • VideaHealth, develops AI software for dentists, raised $40M • Conifers.ai, working on AI cybersecurity, raised $25M • SuperOps, makes AI tools for IT teams, raised $25M • Paxton, develops AI software for lawyers, raised $22M • Jump, helps financial advisors utilize conversations, raised $20M • Ivo, an AI-powered contract review startup, raised $16M • Bonsai Robotics, makes robots to manage orchards, raised $15M • Unwrap.ai, AI software to help understand customers, raised $12M • qeen.ai, AI agents for e-commerce, raised $10M • Palona AI, AI agents for customer engagement, raised $10M • Little Otter, an AI-powered family mental health startup, raised $9.5M • Aligned, makes AI software for sales teams, raised $8M *This data and this photo comes from Stephanie Palazzolo's great AI Agenda newsletter at The Information #discoverthefuture

8 AI companies moving faster than 99.9% of private companies. All 8 saw a 300+ point jump in Mosaic. Mosaic is CB Insights’ success probability score that factors in financials, commercial momentum, industry health, and management team strength. And ahem, it's 4.7x more predictive of startup success than being funded by top-decile VCs. You probably haven't heard of most of companies...yet. Firecrawl (+392): Already embedded into the AI development stack, Firecrawl reached 350K developers using their AI-optimized web scraping API and formed partnerships with LangChain and Weaviate. Their success has now attracted a fresh $14.5M Series A with participation from Shopify's CEO and Zapier. Leo AI (+361): Specialization wins and Leo AI is proving it with 20K+ active engineers at Scania, HP, Siemens, and Mobileye. Now with a $5M seed, they can scale distribution of their domain-specific mechanical engineering AI that performs at 96% accuracy (2x generic tools). Extend AI (+344): Profitability at Series A is rare. Extend AI reached multi-million ARR and cash-flow positive status counting Square, Brex, Checkr, Flatiron Health, and Fortune 500 enterprises as top customers. With a $17M Series A and the launch of a self-serve platform for faster onboarding, they’re looking to Extend their lead in the document processing space. Sola (+340): Sola secured production deployments at Fortune 100 enterprises and AmLaw 100 law firms within 2 years, grew headcount 300%, and advanced commercial maturity from validating to deploying. Raised $17.5M from a16z. Samaya AI (+331): Winning Morgan Stanley as a customer and 100% MoM growth brought in a headline $43.5M from NEA with participation from Eric Schmidt and Yann LeCun. With thousands of analysts already using the platform, their recent launch of Causal World Models for autonomous economic modeling is one to watch. PlayerZero (+327): PlayerZero is delivering results for enterprise customers like Zuora (80% drop in support escalations, 90% reduction in investigation time), recently added major telecom and manufacturing customers, launched CodeSim for AI-generated code testing, and doubled headcount in 6 months. $15M series A. Invisible (+314): Invisible doubled revenue in 2024, ranked #2 fastest growing AI company on Inc. 5000, hired ex-McKinsey QuantumBlack CEO and former VMware CTO, doubled its engineering org, and secured Microsoft and SAIC as customers. Upscale (+304): Upscale assembled a founding team from Palo Alto Networks, Innovium, and Cavium and is tackling the $20B+ AI networking infrastructure market with open-standard alternatives to vendor lock-in. They also attracted a $100M seed from Mayfield, Maverick Capital, and Qualcomm Ventures. Predictive intelligence spots momentum months before it's obvious. Explore the next rocketships in our free GenAI tracker →  https://lnkd.in/ew9zDUdR

I've been covering AI for nearly four years now. The conversation used to be: Who's going to build the next OpenAI? I'm not really hearing that anymore. The smartest founders I'm watching aren't trying to build general-purpose foundation models. They're using these powerful models to solve very specific problems like medical diagnostics, compliance automation, document parsing, code review. For my latest at Inc. Magazine., I talked to Tim Tully at Menlo Ventures and Kulveer Taggar at Phosphor Capital about which early-stage AI companies are poised to break out this year. The full list: - OpenEvidence, founded by Daniel Nadler (they're a pretty big deal already) - Delve, founded by Karun Kaushik and Selin Kocalar - Listen Labs, founded by Florian Juengermann and Alfred Wahlforss - cubic (YC X25), founded by Paul Sanglé-Ferrière and Allis Yao - Axiom, founded by Carina Hong - Reducto, founded by Adit Abraham and Raunak Chowdhuri - AMI Labs, founded by Yann LeCun (with Alex LeBrun as CEO) - Eureka Labs, founded by Andrej Karpathy What these companies tell us about where AI is actually heading, and why investors are betting on them now, is in the piece. Gift link below 👇

Maybe AI isn’t a bubble? It’s the new infrastructure. A slide shown at OpenAI’s Dev Day 2025 recognized developers whose products had processed over one trillion tokens through OpenAI’s API. Those names were later mapped (on Reddit) to their companies, creating what’s now circulating online as a “leaked list of OpenAI's top 30 customers” list. The top 30 companies using over one trillion tokens each: 1. Duolingo – Education / EdTech  2. OpenRouter – AI Infrastructure  3. Indeed – HR & Recruitment  4. Salesforce – Enterprise SaaS / CRM  5. CodeRabbit – Developer Tools  6. iSolutionsAI – AI Automation & Consulting  7. Outtake – Creative / Video AI  8. Tiger Analytics – Data & AI  9. Ramp – Finance Automation / Fintech  10. Abridge – Healthcare / MedTech 11. Sider AI – Developer Tools  12. Warpdev – Developer Tools  13. Shopify – E-commerce / Retail Tech  14. Notion – Productivity / Collaboration  15. WHOOP – Health / Wearables  16. HubSpot – Marketing / CRM  17. JetBrains – Developer Tools  18. Delphi – Data Analysis / Decision Support  19. Decagon – Healthcare / AI Communication  20. Rox – Workflow Automation  21. T-Mobile – Telecom  22. Zendesk – Customer Service / SaaS  23. Harvey – LegalTech  24. Read AI – Meetings & Productivity  25. Canva – Design / Generative Creativity  26. Cognition (Devin) – AI Coding Agent  27. Datadog – Cloud / DevOps  28. Perplexity – AI Search  29. Mercado Libre – E-commerce / Fintech (LatAm)  30. Genspark AI – Education / AI This list is an incredible mix of private unicorns, public companies, and established enterprises. It shows that leading companies have already operationalized AI at scale, not as a side project but as part of their core infrastructure. What are the revenue implications? At the current enterprise pricing, that equates to about 5 million dollars in compute per company on this list. That’s hundreds of millions of dollars in recurring AI spend. The long tail of smaller customers likely contributes another $1.5 to $4 billion in annual API revenue. This also means the majority of OpenAI’s total usage and income now comes from widespread, recurring enterprise adoption rather than a few large accounts. Yes, AI valuations are high, but for a reason. This data shows that AI is already embedded into the core of the economy, showing up as real, recurring revenue across every sector.

OpenAI, Anthropic, and SpaceX are all preparing for potential IPOs this year. If you don’t follow tech markets, that might not sound like news. But all three going public in the same window, at combined valuations in the trillions, represents something bigger than just stock listings. For context, Saudi aramco’s roughly 29 billion IPO in 2019 was the largest in history. For the first time since the original internet boom, a new wave of genuinely transformative technology companies is becoming accessible to regular public‑market investors instead of staying locked up in private hands. OpenAI has been around since 2015, funded by venture capital and massive strategic investment from Microsoft and more recently Softbank. Most people couldn’t invest even if they wanted to. Going public changes that. But there’s a more interesting shift happening. Private companies can operate with limited disclosure, burn cash for years, and postpone hard questions about profitability. Public companies face quarterly scrutiny, have to explain their strategies to analysts, and eventually need to show they can make money. That discipline is a forcing function. Consider that SpaceX has Starlink subscription revenue and government contracts. Anthropic focuses on enterprises willing to pay premium prices for reliable, safety‑oriented AI systems. OpenAI has consumer‑level scale. All three need capital for different reasons, but they’re all choosing to tap public markets in the same era. The timing matters. Figma’s blockbuster IPO last summer was massively oversubscribed and widely seen as opening the floodgates for large tech listings. So, the questions are: 1/ Are these companies mature enough to handle public markets? 2/ What happens when the companies building AI infrastructure have to operate with transparency and prove their business models work? 3/ What approaches will create value beyond the hype? Ultimately my hope is we get better products, clearer strategies, and broader public participation in the upside. The age of AI as a private club is ending. What comes next should be more interesting. _______ Sources: The Wall Street Journal: https://lnkd.in/gQhUvJhc The New York Times: https://lnkd.in/g8fGpMn3 Barron's: https://lnkd.in/gP4s-bWj? LinkedIn News: https://lnkd.in/gm7ESvWD _____ Image created with Google Nanobanana Pro