Google AI Futures Fund team discusses AIFF's India market entry with Accel's Prayank Swaroop

Excited for K-Dense Inc. to partner with Google to provide accurate & comprehensive scientific data analysis to the world at scale! 1 of only 5 startups are selected out of 4,000+ applicants for the AI Futures Fund × Accel Atoms program. Pictured (L to R): Jonathan (Google), Arthur (Google), Dave (Biostate), Prayank (Accel), and David (Google).

𝗚𝗼𝗼𝗴𝗹𝗲 𝗷𝘂𝘀𝘁 𝗽𝘂𝗹𝗹𝗲𝗱 𝗼𝗳𝗳 𝗮 𝗺𝗮𝘀𝘁𝗲𝗿𝗰𝗹𝗮𝘀𝘀 𝗶𝗻 𝘁𝗮𝗹𝗲𝗻𝘁 𝗮𝗰𝗾𝘂𝗶𝘀𝗶𝘁𝗶𝗼𝗻—𝘄𝗶𝘁𝗵𝗼𝘂𝘁 𝘁𝗵𝗲 𝗮𝗻𝘁𝗶𝘁𝗿𝘂𝘀𝘁 𝗵𝗲𝗮𝗱𝗮𝗰𝗵𝗲. 🧠 🚀 Google didn't just sign a deal; they brought the 𝘀𝗼𝘂𝗹 of Character.ai back home. By hiring Noam Shazeer and Daniel De Freitas, they’ve secured the architects of the modern LLM era. 💡 Here’s my take: This isn’t an acquisition; it’s a 𝘀𝘁𝗿𝗮𝘁𝗲𝗴𝗶𝗰 𝗲𝘅𝘁𝗿𝗮𝗰𝘁𝗶𝗼𝗻. 1️⃣ 𝗧𝗵𝗲 𝗡𝗲𝘄 𝗠&𝗔 𝗣𝗹𝗮𝘆𝗯𝗼𝗼𝗸: Just like Microsoft did with Inflection, Google is using licensing to bypass the FTC. It’s genius, and it’s effective. 2️⃣ 𝗧𝗮𝗹𝗲𝗻𝘁 > 𝗧𝗲𝗰𝗵: Shazeer literally co-authored *𝗔𝘁𝘁𝗲𝗻𝘁𝗶𝗼𝗻 𝗶𝘀 𝗔𝗹𝗹 𝗬𝗼𝘂 𝗡𝗲𝗲𝗱*. In the AI arms race, the researchers are the ultimate prize. 3️⃣ 𝗧𝗵𝗲 𝗚𝗿𝗮𝘃𝗶𝘁𝘆 𝗼𝗳 𝗕𝗶𝗴 𝗧𝗲𝗰𝗵: Consumer AI is becoming a money pit. Even the biggest startups are seeking the safety of hyperscaler pockets. I believe the era of the independent AI unicorn is shifting toward Big Tech satellite. The talent war is over, and the giants won. 📈 How do you compete for talent when your rivals have infinite capital? Is the independent startup dream dying in the age of LLMs? Let’s talk below. 👇 #Google #CharacterAI #AITalent #ArtificialIntelligence #TechStrategy

BREAKING: SolveAI has announced a WHOPPING $45m Series A just 8 months after launching! The company, founded in the UK by former Palantir engineer Steve Basher, lets any employee build fully-compliant enterprise software without touching a line of code. It raised a previously unannounced $5m pre-seed round led by Accel (Cecilia Wang), and has just announced raising a $45m Series A led by Tom Hulme at GV (Google Ventures). The company is targeting a slightly different market to other vibe-coding platforms by exclusively targeting enterprise customers and helping them deploy secure, functional code. This is VERY COOL and I would loved the opportunity to use this back in the day. Will be watching this one VERY CLOSELY. Congrats all 👏 Interested in European Tech? Sign up to my newsletter to stay up to date: https://lnkd.in/d54jdWXP

🤔 OpenAI's IPO is coming — but investors aren't exactly lining up 👇 According to The Information, OpenAI is at least six months from going public. Despite raising $110B at a $730B valuation weeks ago, the investor community is skeptical. 💰 Much of OpenAI's funding is circular. Amazon invests — tied to AWS usage. Nvidia invests — tied to chip purchases. These aren't bets on fundamentals. They're supply chain deals disguised as investments. 📉 OpenAI is projected to lose $14B in 2026. Revenue grows fast, but spending grows faster. 🔮 The WeWork parallel is real: a company propped up by investors who are also its biggest customers. If it fails, the supply chain goes down with it. ⚖️ I don't think OpenAI will implode. But the Q4 2026 IPO window — alongside Anthropic and SpaceX + the open source competitors from China - will force a reckoning. Sustainable business or hype machine? Read the article: https://lnkd.in/er2ZAxz3. Follow for more Agentic AI Insights: Job van den Berg | ai.nl - Agentic AI Insights | The Automation Group | eBrain.ai | Proxies.

OpenAI has just closed an unprecedented $110 billion funding round from major investors including Amazon, SoftBank and Nvidia — one of the largest private financings in history. This deal values OpenAI at about $730 billion before the investment, and potentially around $840 billion post-money when including the capital raised. According to recent valuations circulating widely online, these are among the most valuable privately held companies in the world today: SpaceX – ~$1.25 trillion, often cited as the most valuable private company. OpenAI – $730 billion+ after the $110 billion round. ByteDance – ~$480 billion. Anthropic – $380 billion from recent funding. Stripe – around $140 billion+ after a recent share sale event. Databricks – around $134 billion. Ant Group – $79 billion valuation cited. Revolut – ~$75 billion. Binance – around $62 billion. Canva – around $42 billion.

"Philippine developers are really world-class when it comes to building IT systems." That's not us talking. That's the CEO of a Singaporean AI startup that partners with Google, NVIDIA, and Microsoft - describing her experience working with our team. We built the production platform for Nudgyt, a company that combines behavioral science with AI. They bring the intelligence. We gave it a home that scales. We wrote about what cross-border tech partnerships in Southeast Asia actually look like when both sides show up as equals. https://lnkd.in/gQjEdSnj

Why haven’t Google or Amazon already built this? Our Founder, Head of Innovation, Finbar O’Hanlon, has recorded a new 10‑minute episode that goes straight to a question we hear often about ION: If Virtual Video is so powerful, why didn’t the big platforms invent it first? In “Why hasn’t Google already built that? Finbar shares four real stories that all point to the same conclusion: the most important breakthroughs rarely come from inside the dominant players. They come from people who are willing to question assumptions everyone else treats as fixed. He then connects that pattern to ION. For decades, video has been treated as a file to be stored, delivered and played. ION started from a different premise: what if video is actually a structure that can be separated from its content and assembled dynamically in real time? That question led to Virtual Video, and to the IP portfolio and new control‑layer patents we’re now commercialising as infrastructure for AI‑native, agentic content. ION is focused on the part of the stack that remains unsolved at scale: the assembly layer that makes programmable, governed video possible. If you want a clear view of how we think about innovation and why ION exists where it does in the value chain, this is worth watching. #ION #VirtualVideo #DisruptiveInnovation #AIInfrastructure #AgenticContent

I don't claim any visionary power, but it's not difficult to see one result of the AI frenzy. 1. The trigger will be the bankruptcy of OpenAI. That's now inevitable, and given the pull back of investors, their absurd burn rate, the absence of senior management with deep operations and sales experience, and their ridiculous valuation, it will occur near term. Could they be saved? Absolutely, but it would require a complete restructuring of the firm and their objective. The problem is the restructuring isn't politically or culturally possible inside that organization. This is not a company in distress — it’s a company in terminal acceleration toward insolvency. 2. Investors will be much more reluctant to pour in funding to every other AI firm. It isn't difficult to understand there's an old pattern that applies here too: capital flight from the sector once the flagship shows signs of collapse. 3. Startups and early - mid stage firms that are not profitable and rely on continued investing will vanish. OpenAi's burn is driven by compute cost, train, and infrastructure. These costs are identical across the board. 4. Two major firms will survive: Microsoft and Google. Both have deep pockets and multiple revenue streams and can absorb even a $100B loss without foundational shock. 5. Anthropic can survive, but only near term. Long term survival would be viable if Google and Microsoft propped them up... but why would they do that?

Three people who think about AI at the highest level came together in a room with our founders at Founder Connect. Prayank Swaroop, Partner at Accel, Josh Woodward, VP of Google Labs and Google Gemini, and Prateek Jain, Distinguished Scientist at Google DeepMind, joined us for a conversation on what it actually takes to build right now. Here are a few things that stood out: • Your first version will probably be something no one wants. Ship it anyway, within 50 to 100 days. That's when the real information starts. • Small teams, blurred roles, incentives tied to learning. This is how Google Labs moves. It's worth reverse-engineering. • Voice, fluency, and context assembly are still unsolved. That's not a warning, that's a whitespace. • Count the minutes you spend in front of your customer each day. Whatever that number is, it should be higher. Watch the full conversation from the link in the comments. AI Futures Fund · Jon Silber · Shekhar Kirani · Arthur Soroken · David Benjamin · Google for Developers

there's a ceiling on how good your product decisions can be. and it has nothing to do with how good you are. i dual-hatted as a product + engineering lead at scaleups for years. lived the scramble of finding the right signals when it matters and acting on it. meanwhile Brian Ngo led engineering teams at Slack and Palantir, deeply embedded in product decisions. we both ended up at Google (in the Maps team!). came from different paths and landed on the same conclusion: big tech doesn’t always have better product talent. but their teams are definitely better equipped. we had teams of PMMs, researchers, analysts constantly feeding us signals. the infra made good decisions the default. coming from startups, this resource gap was hard to ignore. and when LLMs came along, we saw the first real shot at closing this gap. give every team the kind of signal coverage that used to require a whole org. we left Google to build ThriveAI for the product teams we used to be on. now clocking in at 200k+ signals analyzed daily.