About
Independent Non-Executive Director (INED) and venture capital investor with 25+ years of…
Activity
5K followers
-
Alfred Chu reposted thisAlfred Chu reposted thisThe latest Vobile Rights Brief is here, tracking evolving rights and content protection developments across media and entertainment. This month’s brief looks at how Disney and TikTok are creating a licensed framework for fan-made content, giving select creators access to approved studio assets from Disney, Pixar, Marvel, Star Wars and other brands. It also examines how HBO Max is using AI and human editorial review to turn library clips into a personalized short-form discovery experience. The brief also explores new AI transparency requirements in Europe, including machine-readable marking and disclosure requirements for certain AI-generated and manipulated content under the EU AI Act. At Vobile, we help rights holders identify, monetize, protect, and manage copyrighted video and audio across platforms including YouTube, Meta, and TikTok. Read the August Rights Brief below.
-
Alfred Chu reposted thisAlfred Chu reposted thisToday, we announced that we crossed $7B in revenue run-rate, growing over 80% year over year. We also shared: 📊 $100M+ revenue run-rate for Lakebase 📈 $1.5B+ revenue run-rate for Lakehouse, growing over 100% YoY 💰 Continued positive adjusted free cash flow And we raised $5B in our latest fundraise. Thank you to Coatue, Blackstone, MGX, T. Rowe Price, and Sixth Street Growth for leading this round. We’ll use this capital to invest in: • Lakebase, our serverless Postgres database built for AI agents • Genie, our AI coworkers that actually understand your business data • Unity AI Gateway, our multi-AI governance solution that helps control costs Victor Dey shares more in Forbes: https://lnkd.in/gcdhCYvWDatabricks Hits $190 Billion Valuation As CEO Ali Ghodsi Claims AGI Has Already ArrivedDatabricks Hits $190 Billion Valuation As CEO Ali Ghodsi Claims AGI Has Already Arrived
-
Alfred Chu reposted thisAlfred Chu reposted thisMost debates about AI token economics ask: “Is AI worth the price?” Some point to companies desperately cutting AI costs and conclude the spending is unsustainable. Others argue that buyers will pay almost anything for frontier intelligence. Both can be right. The mistake is treating AI demand as one market. The question behind every AI dollar is: Is intelligence actually the bottleneck on what the buyer is trying to produce? If no, you’re in an efficiency market. AI is reducing the cost of an existing output, and once a model is good enough, more intelligence adds little value. The buyer optimizes for the cheapest adequate intelligence: routing, caching, lower-cost inference, open source. If yes, you’re in an expansion market. Better intelligence creates more output: more alpha, better drug candidates, more software, new products. Here, the value created can dwarf the token bill. This leads to a counterintuitive conclusion: Most enterprise AI will run on commodity models, but frontier intelligence will still retain enormous pricing power. That sounds like bad news for closed-source labs, and in part it is. A large share of enterprise AI spend today sits in efficiency workloads that once required frontier models. As open models become “good enough,” those workloads migrate to cheaper alternatives and the premium disappears. But frontier pricing power survives wherever intelligence remains the bottleneck. Closed labs need to keep moving into those markets faster than open source commoditizes the markets behind them. There, even a small capability edge can be worth millions or billions. Even frontier demand has two different economics. In zero-sum markets like trading, AI spend ratchets upward as competitors match each other, even if industry output stays flat. In positive-sum markets like drug discovery, spend compounds because better intelligence expands the economic pie itself. So “Is AI worth the price?” is the wrong question. Ask instead: Is intelligence still the bottleneck? If no → buy the cheapest model that clears the bar. If yes → frontier intelligence can remain cheap at almost any price. We unpack the full framework in The Two Token Economics of AI: https://lnkd.in/ghJAd_WA
-
Alfred Chu reposted thisAlfred Chu reposted thisI sell an AI platform to the largest retailers on the planet. I can tell you, we are at the tip of the iceberg on AI consumption. Many companies still have barely integrated AI into a majority of their processes. Yet still... Databricks is growing 80% YoY at about $7B in ARR. Even if we decelerate to 60% over the next five years, at the end of that five years we would be at $72.35B which I believe is a distinct possibility. To put that in perspective. Databricks would be larger than the following Fortune 100 Companies: Cisco Systems -- $57B Nike -- $46B Coca Cola -- $47B Tyson -- $54.4B Oracle -- $67.4B Salesforce -- $41.5B Capital One -- $69.5B IBM -- $67.5B Delta Airlines -- $63B Uber -- $52B (FWIW... this list is unfair because it assumes 0% growth for any of the above companies which will obviously not be the case) This is not a knock on these companies whatsoever. These are the best companies and brands on the face of the planet. They've built incomprehensibly incredible businesses that provide massive value to their customers and shareholders. But I think we've seen a lot of talk in the market about "valuations are ridiculous" and "have we overpriced future growth" of these AI companies. The question is... Would you rather own stock in an AI company that is $7B in revenue and growing 80% YoY. Or any other company that is growing 10-15% YoY and doing $30B in revenue? You'd have to ask yourself. "Is this growth sustainable" I believe we are at the very, very beginning. What an awesome time to be in software.
-
Alfred Chu reposted thisThe FIFA World Cup showed the incredible scale of sports fandom across social video. Check out our latest insights on how fans watched, shared and engaged with World Cup content across social platforms.Alfred Chu reposted thisThe World Cup wasn’t just the biggest tournament in social media history. It marked a turning point in how fans experience the game. Vobile’s latest analysis found: ▶️ 3.9M social videos 👁️🗨️ 388B social views 👍 17.6B engagements 📱 73% of views came from videos under one minute Creator-led storytelling and viral player moments were key to engagement, as was the inaugural halftime show. The time around, the tournament showed how sports fandom is becoming increasingly social-first, short-form, and global. Check out the full findings below. If you're interested in learning how Vobile data can inform your strategy? Let’s connect!
-
Alfred Chu reposted thisAlfred Chu reposted this🚀 Databricks has reportedly signed a term sheet to raise at a $𝟭𝟴𝟴𝗕𝗻 𝘃𝗮𝗹𝘂𝗮𝘁𝗶𝗼𝗻 led by Coatue. Back in 2013, Andreessen Horowitz led Databricks’ $𝟭𝟰𝗠𝗻 𝗦𝗲𝗿𝗶𝗲𝘀 𝗔 𝗮𝘁 𝗿𝗼𝘂𝗴𝗵𝗹𝘆 $𝟬.𝟮𝟲 𝗽𝗲𝗿 𝘀𝗵𝗮𝗿𝗲. That single investment has now compounded by ~𝟭𝟬𝟬𝟬𝘅 (after dilution) before even considering a16z’s follow-on investments. — 🤯 Let that sink in. 🧠 This is exactly what Marc Andreessen talked about at the last A16Z summit. He + Ben Horowitz still spend most of their time investing at the earliest stages. Not because it’s glamorous, but because that’s where the asymmetry lives. 💰A $14Mn cheque into a company that very few believed could become one of the defining technology businesses of its generation can ultimately be worth $10Bn+ (with their total position with follow ons being worth much more). 💡 But the cheque was never the whole story. When Ali Ghodsi reportedly described Databricks as a potential $10Bn company to a recruit, Ben Horowitz challenged the assumption entirely. “𝘞𝘩𝘺 𝘢𝘳𝘦 𝘺𝘰𝘶 𝘵𝘩𝘪𝘯𝘬𝘪𝘯𝘨 𝘴𝘰 𝘴𝘮𝘢𝘭𝘭?” Then, in 2019, when Databricks was reportedly considering selling for around $4Bn, Ben told Ali they were “severely underselling the opportunity.” Later, when Databricks was struggling to gain traction inside Microsoft, Ben personally introduced Ali to Satya Nadella. That relationship became foundational to Databricks’ Azure partnership. — The very best venture investors don’t just provide capital. They expand ambition, challenge assumptions & open doors. And sometimes, they fundamentally change the trajectory of a company. 📈 Many people outside venture often push back: “But you’ll just get diluted along the way.” Of course you will. Every great company raises capital. But that’s exactly the point. These 900x+ returns are after years of dilution, not before. When a company compounds from under $50Mn to $188Bn, & perhaps one day even $1 Trillion the value created overwhelms the reduction in ownership. 🚀 That’s what makes venture so different. 𝗩𝗲𝗻𝘁𝘂𝗿𝗲 𝗶𝘀𝗻’𝘁 𝗷𝘂𝘀𝘁 𝗮𝗻 𝗮𝘀𝘀𝗲𝘁 𝗰𝗹𝗮𝘀𝘀. 𝗜𝘁’𝘀 𝗮𝗻 𝗮𝗰𝗰𝗲𝘀𝘀 𝗰𝗹𝗮𝘀𝘀. 🖇️ in comments to Coatue’s rationale. 🔔 PS - if you enjoyed this, ♻️ consider sharing with your network and follow me, Akhil Paul, for more. #startups #venturecapital #investing #tech
-
Alfred Chu reposted thisAlfred Chu reposted thisDatabricks just set a new mark: $188 billion. 📈 A Coatue Management-led ~$3B round values the data + AI company ~40% above its February valuation (a $54B step-up in about 7 months). The trajectory: ~$38B in 2021 → $188B today. Roughly 5x in under five years. But the number that matters isn't the check, but what it signals: 🔹 At ~1.6% dilution on ~$6.9B of revenue, this isn't a company that needs cash. It's a re-pricing and a liquidity event. 🔹 Mega-rounds like this are letting mature companies stay private and defer the IPO on their own timetable. 🔹 Private capital keeps concentrating at the very top of AI. In Q1, five companies took roughly three-quarters of all US venture dollars. Databricks was one of them. Databricks may be the clearest example yet of a real-revenue business doing what the AI model labs do on narrative alone. The question for 2027: will public markets pay 27x revenue when the IPO window finally opens? 👇 (Round is a signed term sheet, expected to close later this summer) PitchBook Morningstar #PrivateMarkets #VentureCapital #AI #Databricks #IPO #AIBQ
-
Alfred Chu shared thisCXMT overtook Moutai as China’s largest company by market cap.. The difference is one stores memory, and the other erases memory.. 🖥️ 😂🍸Alfred Chu shared thisChina's memory chipmaker CXMT made a blockbuster debut on Shanghai's STAR Market on Monday. CNBC's Elaine Yu explains all you need to know about CXMT, or Changxin Technology Group, which is now the most valuable China-listed company. Read more here: cnb.cx/44PR0o4China memory chipmaker CXMT skyrockets 500% in blockbuster Shanghai debutChina memory chipmaker CXMT skyrockets 500% in blockbuster Shanghai debut
-
Alfred Chu reposted thisAlfred Chu reposted thisMyPOV: This AI demand is bigger than today's semicon market. $1.4T by 2030! AMD #AMDAdvancingAI
Alfred Chu commented on a post
1w
Databricks' growth is accelerating at scale.
-
Alfred Chu liked thisAlfred Chu liked thisMyPOV: still the best at full self driving. TeslaFull Self-Driving Is the Best Thing About the New Tesla Model YFull Self-Driving Is the Best Thing About the New Tesla Model Y
-
Alfred Chu liked thisAlfred Chu liked thisHad a great time at Ray Summit 2026. The Marriott Marquis is one of my favorite venues -- it was also home to ISSCC for many years
-
Alfred Chu liked thisAlfred Chu liked thisLiveX is now live at Moscone Center! The first 30 minutes at a big event are usually the hardest. Registration lines, an unfamiliar building, too many sessions to choose from. We're partnering with the team at Moscone Center to change that: AI concierges on the floor that greet attendees, answer questions in any language, and make finding your way a little more enjoyable. One simple goal together: make the convention center experience AI-first, and a lot more fun. Special thanks to Leonie, Suzanne, and Dan for the great support in making this happen. True partners every step of the way. Here's are a few photos of the devices from the floor 📸 If you're at Moscone, come say hi.
-
Alfred Chu liked thisAlfred Chu liked thisIn this tutorial, Jules Damji (Databricks) walks through a real problem for agent developers: they already use multiple coding harnesses, but switching them makes it hard to share context, carry policies and guardrails, collaborate without pasting into Slack, and see whether the agents did the right thing. Omnigent sits atop those harnesses as a session-scoped meta-harness, so context and policies persist, and OpenTelemetry traces go to MLflow. In the demo, Polly orchestrates a coding task, assigns the coding work to Codex, and routes review and a PR to Claude. In MLflow, you can inspect the tool chain, token counts, duration, and which agent ran which span. #Omnigent #MLflow #AIAgents
-
Alfred Chu liked thisAlfred Chu liked thisAI was supposed to destroy call center jobs. But Philippine BPO employment grew 4% last year even as AI adoption accelerated. AI creates a Jevons paradox for human labor. The logic is that employment only falls if productivity rises faster than demand. If AI lets each worker handle 2x as much work, but the lower cost causes companies to consume 3x as much of the service, you actually need more workers, not fewer. AI lowers the cost of customer support, back-office work, research, accounting, recruiting, etc. Companies respond by offering more support, serving smaller customers, outsourcing more workflows, and doing work that previously wasn't economical at all. So labor required per task can collapse while total employment still rises because the number of economically viable tasks expands even faster. This doesn't mean AI won't replace these jobs eventually. But the transition could look very different where automation may expand labor-intensive industries before it shrinks them.
-
Alfred Chu liked thisAlfred Chu liked thisSovereign AI will become a board-level issue for every enterprise over the next 12 months. The question won’t be who has the best model. It will be who controls the context, workflows, permissions, and decisions around it. That’s where the real moat will sit. I’m looking forward to discussing this at the Agentic Harness Summit, September 8–10: how enterprises can deploy powerful AI agents without giving up control of their IP, data, or business logic. The next AI advantage won’t come from the model alone. It will come from the harness around it. Registration details: September 8–10. https://lnkd.in/g8epvKFu #ainativegtm #aiarchitecture Hard Skill Exchange Lyn Zhang Audra Proctor Gen #SovereignAI
-
Alfred Chu liked thisAlfred Chu liked thisMyPOV: @OpenAI claims its new chips outperforms NVIDIA Blackwell - with Ashley Webster on Fox Business Network @Varneyco Constellation Research, Inc. https://lnkd.in/gfm4GFJhOpenAI claims its new chip outperforms NVIDIA's Blackwell | Fox Business VideoOpenAI claims its new chip outperforms NVIDIA's Blackwell | Fox Business Video
Experience & Education
Languages
-
Chinese
-
Other similar profiles
Explore more posts
-
Ryosuke Kimura
Sila, the silicon-anode battery company, just raised $300M to expand its Moses Lake, Washington plant, aiming to scale from roughly 2 GWh to as much as 250 GWh of anode material over five years. If it lands, that would be one of the world's largest anode facilities. The detail I keep coming back to is not the chemistry. It is the manufacturing. Plenty of battery startups win in the lab and then die in the valley between a working cell and a gigascale line that actually yields. The hard part of deep-tech hardware is rarely the first prototype; it is the thousandth identical one at cost. Sila moving into a second build-out phase is a bet that it can cross that gap. There is a second signal worth sitting with. The demand driver has quietly shifted. This capacity is justified less by electric vehicles now and more by AI, defense, electronics, and space, framed around domestic supply and technology sovereignty. The customer for advanced batteries is changing. For anyone building physical deep-tech: what is the real bottleneck between your best lab result and a line that yields at cost, and who on your team has actually crossed it before? https://lnkd.in/gEujYBx8 #DeepTech #Batteries #AdvancedManufacturing #EnergyStorage #SupplyChain #VentureCapital #LifetimeVentures
1 Comment -
Jake (JH) P.
Excited to finally share what we’ve been building over the past year. Today we’re announcing the first investment from ARKITEKT EQUITY. Over the past year, we have been closely observing the structural shifts occurring within semiconductor ecosystems. As manufacturing systems become more complex, the infrastructure enabling those systems becomes increasingly strategic. These environments demand operational clarity, coordination, and long-term alignment. ARKITEKT was built around a simple idea: durable value is created where frontier technology meets pragmatic execution. Our investment in BBTech reflects this conviction. This is just the beginning of what we aim to build at ARKITEKT EQUITY. Stay tuned! https://lnkd.in/g_pQiBZY
-
Eric Bentov
Batteries are shifting from components to catalysts across EVs, defense, and space. In her latest Forbes article, Arieli Group Managing Partner Lisya Bahar Manoah outlines how advances in materials and manufacturing are shifting cost, performance, and supply-chain dynamics, and where long-term, infrastructure-level bets can create durable value. I’m proud of our firm’s role in this ecosystem and the opportunities we’re building toward. 𝗥𝗲𝗮𝗱 𝘁𝗵𝗲 𝗳𝘂𝗹𝗹 𝗮𝗿𝘁𝗶𝗰𝗹𝗲 𝗵𝗲𝗿𝗲: https://lnkd.in/dGhfpA_b #BatteryTech #DeepTech #StrategicCapital #ArieliGroup #ForbesFinanceCouncil
1 Comment -
Anjli Jain
**Navigating Regulatory Landscapes in Tech** Tesla's recent avoidance of a 30-day suspension underscores the complex relationship between innovation and regulation in the tech industry. As the company refines its Autopilot features, it emphasizes the need for agility and compliance in a rapidly evolving landscape. At ElevenX Capital, we recognize the importance of understanding regulatory challenges, which can impact investment strategies in the tech sector significantly. How can investors better assess the regulatory risks of emerging technologies? #investing #innovation #venturecapital #entrepreneurship
Explore top content on LinkedIn
Find curated posts and insights for relevant topics all in one place.
View top content