About
Successful, highly strategic marketing executive with broad experience from start-ups to…
Articles by Jonathan
Activity
6K followers
-
Jonathan Symonds shared thisThis week includes some thoughts on the US/China Summit and what it means for #sovereignAI and what was also an interesting week in the #UK. There is a lot to process these days....
-
Jonathan Symonds shared thisSeeing this market develop in real time is pretty cool. Agree that securitization follows standardization.Jonathan Symonds shared thisThe capital required for AI Infrastructure will extend well into the trillions. Financing that buildout will require broader and deeper capital markets with better tools for managing risk. Securitization requires standardization. Standardization requires transparency and that in turn delivers fungibility. Those are the building blocks of what we are seeing in the rapidly coalescing AI Compute market. Unsurprisingly, NVIDIA has led the way. Its DSX AI Factory Blueprint provides a common design framework, while Exemplar validates cluster performance. NVIDIA's contributions are coupled with innovative work being done by companies like Silicon Data, Ornn and OneChronos and you have the components of an AI Compute market, which lets players in every segment manage risk more effectively. I share my thoughts on how this is shaping up and what operators, customers and capital providers ultimately benefit. Link in the first comment.
-
Jonathan Symonds shared thisPlenty on the #SovereignAI docket this week. The US/China Summit leads the way, with the EU failing on the chip front and Canada upping its game. Good research (who uses what model) and lots of Bits and Bytes.
-
Jonathan Symonds reposted thisThe financial architecture to fund NVIDIA AI Infrastructure is taking shape in real time. OneChronos is building a marketplace to invest in NVIDIA AI factories as an asset, independent of the offtakers. Efficient large-scale capital formation is around the corner. Something that is possible because NVIDIA AI factories are productive, durable, fungible and standarized. Capital partners such as Apollo Global Management, Inc., Bank of America Merrill Lynch, BlackRock, Blackstone, Blue Owl Capital, Brookfield, The Carlyle Group, Citi, Goldman Sachs, J.P. Morgan, KKR, Macquarie Group, PIMCO, UBS will increasingly be able to work with NVIDIA Cloud Partners to underwrite and finance NVIDIA AI Infrastructure at scale. Excited to see more - Kelly Littlepage and OneChronos.Jonathan Symonds reposted thisIt was great to be on the NYSE floor to talk to Marley Kayden about why compute needs its own financial market. AI hardware isn't fungible, but the economic claims on it can be. We talk about why the better approach is to securitize the cash flows generated by individual compute assets, bringing together offtakers, project builders, and lenders through combinatorial auctions. Crucially, our approach gives investors tools to price debt based on the financial fungibility and investability of the AI factory, not the borrower or the offtakers. That's the key unlock. Helping scale AI compute as safely, sustainably, and rapidly as possible, for the benefit of society, is the generational opportunity we're excited to be working on at OneChronos. Thank you again to Marley and Schwab Network for having me on. https://lnkd.in/gwKpBE8t #OneChronos #Compute #AI #GPUsOneChronos CEO on Three Players Shaping the GPU Compute Market | Schwab NetworkOneChronos CEO on Three Players Shaping the GPU Compute Market | Schwab Network
-
Jonathan Symonds shared thisTons of interesting stuff in the #SovereignAI weekly for September 16....this week we cover the existential (or not) threat of AI and how the US and China see it (differently), NVIDIA giving a sovereign designation to Australia and Europe taking an increasingly hard line. Plus some good bits and bytes.
-
Jonathan Symonds shared thisReally interesting take. The ability to encode repeatable first principles into the build process is the model for faster, more economic and more operationally resilient data centers. Well worth the time.Jonathan Symonds shared thisThe heroic AI data center builds make headlines: 122-day sprints, expedited supply chains, 24/7 executive oversight. They also don't scale. Heroics don't compound. First principles do. If you’re building one campus a year, heroics might pull you across the finish line. But the leading companies in the space aren't building one site anymore; they’re building five, ten, or more simultaneously. You cannot hero-build a portfolio across multiple customers and multiple geographies. The core issue in AI infrastructure today is making sure your first principles don’t confuse replication with repeatability: • 𝗥𝗲𝗽𝗹𝗶𝗰𝗮𝘁𝗶𝗼𝗻 (blindly copy-pasting a static blueprint) breaks the second it hits local reality: constrained grid interconnects, climate shifts, water limits, or local permitting. • 𝗙𝗶𝗿𝘀𝘁-𝗣𝗿𝗶𝗻𝗰𝗶𝗽𝗹𝗲𝘀 𝗥𝗲𝗽𝗲𝗮𝘁𝗮𝗯𝗶𝗹𝗶𝘁𝘆 accepts that every site is different, but strictly governs how far those local differences are allowed to travel down the engineering stack. Start with what the compute requires from power, cooling and controls. Work outward to understand how each site can meet those requirements. Then establish interfaces that keep a local difference from forcing changes throughout the factory. A different utility voltage requires an appropriate electrical solution. It shouldn’t automatically reopen the compute topology. A different way of rejecting heat shouldn’t mean reinventing the cooling loop at the rack. Getting those boundaries right is where the hard engineering earns its value across multiple builds. The next team inherits qualified configurations, tested interfaces and evidence that the system works. It can concentrate on the conditions that actually require a new answer. That is the substance behind Radiant’s ambition to reuse roughly 80% of an AI factory’s architecture. The percentage will vary but the ambition imposes a discipline on deciding which engineering decisions remain valid and which need to be revisited. This reality has created a massive strategic vulnerability. Compute architectures now move on an annual cadence, while physical infrastructure runs on multi-year timelines. If you redesign your core facility architecture for every single build, you risk shipping a multi-billion-dollar facility that is structurally impaired at RFS. Manesh Patel PhD CEng MIET, Radiant’s Director of Data Center Development, frames this as a product-architecture problem and details how to draw that boundary and why repeatability depends on understanding what makes each site different. Link in the first comment.
-
Jonathan Symonds shared thisBig week. We go into the Hugging Face acquisition, the Mistral raise, the coming storm in Malaysia and post four! new pieces of research on #sovereignAI. Super interesting time....
-
Jonathan Symonds shared thisGreat content this month including a summary of the state of #sovereignAI which one theme behind NVIDIA's $13B Hugging Face acquisition today.Jonathan Symonds shared thisWelcome to the September edition of the Radiant Throughline. This month we cover "behind-the-meter" power, how performance gets designed in before groundbreak, and why identity is now core infrastructure. Read the full newsletter to find out more:
-
Jonathan Symonds shared thisInteresting week in the world of #SovereignAI. Lot's of chatter at the #G20, SemiAnalysis drops a deep dive into South Korea's AI ambitions, the contest model takes hold and a Chinese open-weight model trained on Chinese accelerators climbs the OpenRouter leaderboard.
-
Jonathan Symonds liked thisJonathan Symonds liked thisA cluster's installed GPU count doesn't tell you how many are available for your next workload. For cluster administrators, fleet management decisions require a clear picture of availability, ownership, topology, and health. Radiant FlightDeck brings that picture together in one comprehensive view of the fleet, with the controls to allocate, recover and put every GPU to work. João Coelho, our VP of Product, breaks it down in our latest blog. Link in comments.
-
Jonathan Symonds liked thisThis is such a great read! Subscribe for some really insightful content.
-
Jonathan Symonds liked thisJonathan Symonds liked thisI am honored to serve as The Hartford’s next CEO. As a result of Christopher Swift’s tremendous vision and leadership, we are exceptionally well-positioned for the future with financial strength, a disciplined strategy and an unwavering focus on excellence. The Hartford has been a trusted partner in some of America’s most defining moments – insuring Abraham Lincoln’s home in Illinois, the construction of the Golden Gate bridge, and providing accident insurance for Babe Ruth. These examples remind us of our enduring commitment to progress, innovation, resilience and leadership. Today, we are carrying our legacy forward, serving millions of businesses of all sizes and tens of millions of individuals and their families. I am excited and inspired to lead The Hartford and see myself as a steward of the company, preserving our strengths and positioning us for success for generations to come. News release: http://ms.spr.ly/6044aY3Op
-
Jonathan Symonds liked thisJonathan Symonds liked thisRadiant is heading to NVIDIA GTC next month in Berlin 🇩🇪 If you'll be there, let's connect & discuss how Radiant's foundations can support your organization's AI Infrastructure capacity plans: Capital + Powered Land + Compute + Software You can book time with our team here to learn about the magic: https://lnkd.in/gHBGqHTN Send me a DM for an invite to our Tuesday night private event! Looking forward to bringing together clients and partners for a great evening overlooking Berlin. #Radiant #PoweredIntelligence #NvidiaGTC
-
Jonathan Symonds liked thisJonathan Symonds liked thisAged 38, Connor Teskey became the youngest ever CEO of Brookfield Asset Management. From junior analyst to the C-suite, from Excel spreadsheets to boardroom mergers - and his playbook for leading a $1 Trillion empire. New S&P Leaders conversation with Connor coming Thursday! Follow Joseph Cass to get it on your feed first.
-
Jonathan Symonds liked thisJonathan Symonds liked thisHonoured to have delivered the inaugural lecture on Agentic AI & Graph Engineering on the first ever Vibe Coding course at UCL Computer Science alongside Philip Treleaven! I still have very fond memories as a CS undergraduate at UCL and amazed how AI has changed software engineering and learning in general. Some of my reflections based on some really good questions with students: 1.) the pace of innovation is crazy (I had to rewrite this lecture 3 times over the summer !) so changes how students stay on top of new breakthroughs 2.) changing career landscape (we will all need to learn how to leverage AI tools and become managers of hybrid AI-human teams, but more importantly higher value output will become a currency in the workplace) 3.) we need more compute and tokens to drive innovation in the next generation of talent (the next AI nobel prize winner/unicorn founder was in that lecture hall but how do we arm him/her with AI compute and tokens to experiment and innovate?) 4.) Live demos (the old adage of never giving a live demo changes with claude code ☺️).
-
Jonathan Symonds reacted on thisJonathan Symonds reacted on thisThe capital required for AI Infrastructure will extend well into the trillions. Financing that buildout will require broader and deeper capital markets with better tools for managing risk. Securitization requires standardization. Standardization requires transparency and that in turn delivers fungibility. Those are the building blocks of what we are seeing in the rapidly coalescing AI Compute market. Unsurprisingly, NVIDIA has led the way. Its DSX AI Factory Blueprint provides a common design framework, while Exemplar validates cluster performance. NVIDIA's contributions are coupled with innovative work being done by companies like Silicon Data, Ornn and OneChronos and you have the components of an AI Compute market, which lets players in every segment manage risk more effectively. I share my thoughts on how this is shaping up and what operators, customers and capital providers ultimately benefit. Link in the first comment.
Experience & Education
Projects
-
Blue Ocean Strategies for Revenue Marketers
-
I spoke with Steven MacDonald of Content Strategies about what it takes to create successful thought leadership content, how to carve out uncontested space with content and why your content is the voice of your brand in the modern B2B buying journey.
Given MinIO doesn't have any a commissioned sales team, marketing is responsible for filling the pipeline - literally. We do that through content. The modern B2B buyer wants to self-educate and content (in the form of the website…I spoke with Steven MacDonald of Content Strategies about what it takes to create successful thought leadership content, how to carve out uncontested space with content and why your content is the voice of your brand in the modern B2B buying journey.
Given MinIO doesn't have any a commissioned sales team, marketing is responsible for filling the pipeline - literally. We do that through content. The modern B2B buyer wants to self-educate and content (in the form of the website, calculators, blogs, podcasts, documentation, videos) is how they move along their journey.
Feel free to watch the entire episode or google it for the 4 min summary.
Recommendations received
7 people have recommended Jonathan
Join now to viewOther similar profiles
Explore more posts
Explore top content on LinkedIn
Find curated posts and insights for relevant topics all in one place.
View top contentOthers named Jonathan Symonds
37 others named Jonathan Symonds are on LinkedIn
See others named Jonathan Symonds