LavaLab

1.1K posts

LavaLab

@usclavalab

Pioneering the student → builder → founder pipeline @usc

Los Angeles, CA Katılım Nisan 2013

241 Takip Edilen1.2K Takipçiler

LavaLab@usclavalab·16h

Traditional ideation is broken in the AI era. In the last 3 years we’ve seen the transition from: SaaS → LLM wrappers → AI-native, vertically integrated apps How do you decide what to build? Our curriculum directors (@EthanPalosh, @allisoncyhuang, Atticus Kim) mapped the space: 4 types of AI companies: 1. Labs (foundation models) 2. Pickaxes (data + tooling) 3. Vertical apps 4. Infrastructure (agentic future) For student-founders, the opportunity lies in areas 2-4. Key lessons: > Build for where models are going, not where they are > Implementation isn’t a moat > Rethink pricing (outcomes > seats) Get ahead of the shift.

English

105

LavaLab@usclavalab·1d

What if leaders could test the effects of their decisions before making them? This is the question our team at Vantage is trying to answer. Their approach: simulating potential world events through large-scale, long-running agentic systems that take on the roles of real-world entities. They’re looking to apply this to healthcare, one of the most complex and high-stakes industries out there. There are already strong signals this space is gaining traction, with companies like @isara_labs raising $94M and YC-backed @societies_io securing $5.3M in seed-stage funding, plus 3 more raises in the last two months. At LavaLab, we build with the belief that models and underlying technology will continue to improve rapidly and that this progress should compound the value of our cohort’s products. That’s why we’re especially excited about Vantage: they’re tackling a cutting-edge concept with massive potential, and we have strong conviction in their ability to execute. 🚀

The Wall Street Journal@WSJ

Exclusive: OpenAI is backing a new AI startup that aims to build software allowing so-called AI “agents” to communicate and solve complex problems in industries such as finance and biotech on.wsj.com/4bTvwKd

English

279

LavaLab@usclavalab·2d

Enterprise technology is shifting from people to agents. @lightspeedvp and @SycamoreLabs are actively building the systems behind the switch. This Tuesday we’re hosting their CFO of venture business, Catherine Chen, to talk through her thoughts on founders, how to stand out as a company, and the current moment. Really excited to bring this perspective directly to our founders.

Lightspeed@lightspeedvp

Great companies are built from great products. Unforgettable companies are forged from long-standing relationships. Excited to share that we are leading the Seed round for @SycamoreLabs, founded by @SriViswan, alongside @coatuemgmt. After 15 years of history, we are excited to back Sri at this pivotal time. Enterprise software is overdue for a change, and Sycamore is building an AI-powered system to transform traditional SaaS tools. If the last era of enterprise optimized for systems, this one optimizes for intelligence. We are looking forward to this new chapter.

English

258

LavaLab@usclavalab·2d

🚨 Another $42M raised for LavaLum 🚨 Huge congratulations to @AustinSpiegel and the entire @SiftStack team! 🎉 In the coming weeks, Austin, a LavaLab alumnus, former Demo Day judge, and active community member, will be joining us to speak with our student founders about his journey and methodology. We couldn’t be more excited, and once again what an incredible milestone for Austin and the team at Sift!

Sift@SiftStack

Today we're announcing a $42M Series B backed by @stepstonegroup , @GVteam (Google Ventures), @Riotventures, @fikavc and @CIV. The AI era for physical systems starts with the data infrastructure underneath it. Hardware teams are building spacecraft, propulsion systems, autonomous vehicles, and defense platforms. They're generating terabytes of high-frequency telemetry per test. But the tools they have were built for software logs, not for sensor data across vehicles running on independent clocks. We built Sift from first principles: every data type, across 11+ hardware-native formats, lands in a structured schema that's readable by engineers and by models. Historical data stays queryable for the life of the program without breaking a budget. 70 employees today. Doubling with this round. All from our new headquarters in Marina del Rey. We're hiring.

English

655

LavaLab@usclavalab·5d

Fundamental to the startup process is getting feedback from customers early and often. That means iterating quickly on MVPs and getting them in front of design partners and potential users whenever possible. Today, the quality bar for these MVPs is higher than ever. Modern tools make deploying a functional and beautiful product the baseline, so this is exactly what we expect from our teams. Tight feedback loops between customer and product is how early stage teams win. Highlights from cohort MVP demos:

English

158

LavaLab retweetledi

Mark McGuire@markmcguire123·6d

Many early-stage founders don’t think about branding until it’s too late. Working with startups at LavaLab, it’s clear: good design and good branding have outsized returns. A few patterns we saw across last semester’s cohort: > The strongest brands were unique to their domain, but always simple and easy to interpret > Teams with a clear audience had way sharper messaging and visuals > Early investments in brand paid off in credibility with design partners and early users Branding is about building trust, clarity, and making it obvious why you matter. Some of my favorite examples from last semester’s cohort ↓

English

260

LavaLab@usclavalab·25 Mar

Before Terafab, Optimus, OpenAI, or Starlink, @elonmusk sat down with our student founders. Here are the lessons that still apply 12 years later: > First-principles thinking > Build where people think it’s impossible > Define huge goals and rally around them Today’s industries and technologies look very different from 2014, but this mental model remains just as important. At LavaLab, we’ve spent over a decade building the ecosystem that makes this brand of impossible ambition normal for the next generation of founders. First principles → conviction → execution. Simple, not easy.

LavaLab@usclavalab

Incredibly honoured to have @elonmusk speak at our dinner last night at @USC.

English

771

LavaLab@usclavalab·24 Mar

If you’re a college student trying to break into startups, here are your teachers: > @theresidency → hands-on cofounding + demo environments > @zfellows → community of top 1% young builders > @fdotinc → hands-on sprint programs in SF > @ycombinator Startup School → free + elite startup curriculum > @vercinyc → creative + tech collective in NYC (founded by LavaLab alums) Bonus: USC student? → @usclavalab USC’s premier student startup accelerator Learn from the best. Find your cofounders.

English

251

LavaLab@usclavalab·24 Mar

Strong points on the shift from seats to usage and the agentic economy. At LavaLab, we’ve always optimized for the reality of growth first. Small, fast-moving teams win, so we teach founders to stay scrappy, leverage every tool, launch, and iterate. The rest follows.

David George@DavidGeorge83

x.com/i/article/2036…

English

693

LavaLab retweetledi

Mark McGuire@markmcguire123·18 Mar

“we niched down … [and] started to notice patterns” Is what LavaLum @colegawin_ said helped him find Lemma’s first customer. They started too broad. Too many industries. Too many use cases. Narrow the lens and get your first client. uselemma.ai

English

237

LavaLab@usclavalab·13 Mar

More here: linkedin.com/posts/usclaval…

English

125

LavaLab@usclavalab·13 Mar

Takeaways from Altos Ventures ($10B AUM, early Roblox & Quizlet): > Extreme focus beats charisma > Show your treasure map, not just your TAM > “Boring” industries hide the biggest opportunities @cathysshin calls this being the hedgehog. Stay obsessed, stay focused, bet long.

English

336

LavaLab@usclavalab·12 Mar

At weekly LavaLocks, teams share traction reports. That traction starts with conversations. Right now, our founders are deep in user interviews across: > AI data curation & protection > Equestrian health wearables > Predictive inventory management Strong start for S'26🚀🌋

English

282

LavaLab@usclavalab·11 Mar

Pictured: the teams, their mentors, and their posters.

English

291

LavaLab@usclavalab·11 Mar

🚨BREAKING: Lavalab S’26 is underway. Each semester, we select 28 of USC’s top student founders. 2 weeks ago, 7 teams were formed: 2 devs, 1 designer, 1 PM each. Now teams have already: > Performed 40+ user interviews > Explored dozens of industries > Pitched 3 different ideas