About
As a recovering engineer (license to code was revoked in 2008, currently under appeal), I…
Activity
7K followers
-
Graham Brooks shared thisHuge congratulations to Manish Sood and the entire Reltio team on their acquisition by SAP. This is a well-deserved outcome for a team that has spent the better part of a decade tackling one of the hardest and most important problems in the enterprise and reflects not only exceptional execution, but also the power of a long-held thesis. Back in 2014, we developed a strong conviction around one of the hardest problems in the enterprise: getting to complete, clean, and trusted data. While much of the market focused on applications and analytics, we believed those layers would ultimately be constrained by the quality of the underlying data. Simply put, garbage in, garbage out or worse, decisions made on incomplete data. That belief led us to co-lead Reltio’s Series A with Crosslink and back Manish and team as they set out to solve data mastering in a fundamentally new, cloud-native way. While Reltio grew a great data infrastructure business, their Context Intelligence solution-has become the critical data foundation needed to build AI agents in the enterprise. Whether it is a sales rep or an AI agent making a decision, having full, real-time context matters. Knowing a customer has an open service issue before pushing an upsell is a simple example, but behind it sits a complex challenge: aggregating, cleaning, and unifying fragmented data across systems, technologies, and schemas. Reltio addressed this head on with an AI-ready platform that delivers real-time global access, AI-driven data quality, and a graph-native architecture that enables true contextual understanding. Manish and team have executed incredibly well, earning the trust of many of the world’s largest enterprises and growing the business 100x since our initial investment. SAP is the ideal partner to take this even further. As one of the most important enterprise platforms globally, SAP sits at the center of mission-critical workflows and now, with Reltio, at the center of trusted data for AI-driven decision making. Excited to see what comes next for this combination and grateful to have been part of the journey from the beginning. Liam Donohue Anshuman Kanwar James Redfern Don Bulmer Connie Puglia https://lnkd.in/e-KpVwnGA new chapter for Reltio: SAP to acquire Reltio to help make enterprise data AI-ready - ReltioA new chapter for Reltio: SAP to acquire Reltio to help make enterprise data AI-ready - Reltio
-
Graham Brooks shared thisCongratulations to the Keebo team on welcoming Eric Shoemaker as CEO. Eric has successfully scaled enterprise infrastructure before, including at .406 portfolio companies Vericode and CloudHealth, where he was the first GTM hire and built that team from the ground up. Now he returns to the .406 Ventures portfolio for a third time, stepping in to lead Keebo as performance and cloud efficiency become mission-critical. We’re thrilled to be back at it together. Read more here: Keebo Appoints Eric Shoemaker as Chief Executive OfficerKeebo Appoints Eric Shoemaker as Chief Executive OfficerKeebo Appoints Eric Shoemaker as Chief Executive Officer
-
Graham Brooks posted thisThree weeks ago, I wrote about the importance of a unified national AI framework to ensure the U.S. remains competitive. Last week’s announcement of the executive order on AI preemption is a meaningful step in that direction. It signals clear intent to avoid a state-by-state regulatory landscape that would otherwise make it harder for U.S. companies to build, train, and deploy state-of-the-art AI systems. That said, it’s also worth being candid about the limitations of this approach. An executive order is, by nature, an imperfect vehicle. While there is logic to enforcing this AI preemption via an executive order because AI regulation will impact interstate commerce, there are real questions about how effective it will be in practice. In particular, it is inherently reactive. States must first pass laws, and only then does the federal government respond through litigation or by reducing the state’s funding. Several states have already indicated that they will challenge the legality of this EO. Resolving the conflict between state laws and the EO will take time with high uncertainty, which will be inherently difficult for start-ups looking to move fast with clear regulations. Ideally, the existence of the EO acts as a deterrent, but it is not the same as durable legislative clarity. All that said, given the current political reality, this was the best near-term option available. States’ rights versus a unified AI policy remains a hot-button issue that does not map cleanly to either side of the aisle, and with midterms approaching, there has been little appetite for anyone to take the lead on comprehensive federal legislation. With AI preemption language cut from the defense bill and no obvious legislative vehicles on the horizon through the end of 2026, a congressional solution may still be years away, and the likelihood of that solution will be dependent on the midterms. In that context, the executive order provides a necessary, if temporary, bridge. For founders building at the frontier, clarity matters. And while this is not the final destination, it is an hopefully a step toward creating the conditions U.S. innovation needs to continue leading globally. #AI #AIRegulation #AIExecutiveOrder
-
Graham Brooks shared thisWas great to join all of the AI Innovators at NeurIPS 2025, an incredible gathering of minds across academia, applied research, and the infrastructure builders shaping the next wave of AI. Sharing Kevin C. Wang's impressions and adding mine. After years of attending NeurIPS, I again left genuinely energized by the quality of conversations and by how many teams are now thinking deeply about deployment, scalability, and real-world impact. But I also came away reflecting on a broader structural shift. Historically, academia sat at the frontier — and NeurIPS was where they and the leading labs showcased their most ambitious work. This year, there were noticeably fewer papers and fewer attendees from the major frontier labs. In my discussions, this raised another a fundamental question for our industry: as AI innovation becomes increasingly dependent on massive compute and data — assets that frontier labs uniquely hold — what will be the evolving role of academia in driving breakthroughs, openness, and diversity of ideas? If labs continue to internalize cutting-edge work and publish less, how might that shape the pace and trajectory of innovation going forward? #NeurIPS #AI #InnovationGraham Brooks shared thisBack from NeurIPS 2025, and the shift in tone this year was unmistakable. The conversation has finally moved from “make models bigger” to “make AI systems controllable.” A few things really stood out: • Formal methods moving from theory toward practical tools for constraining model and agent behavior • Long-context reasoning is shifting from bigger windows to usable long-term memory and planning • Multi-modality becoming the assumed baseline • Safety work turning into real evaluation and monitoring tools We also hosted our annual AI Innovators & Builders Dinner, bringing together faculty and researchers from UC San Diego, The University of Texas at Austin, Massachusetts Institute of Technology, University of Pennsylvania, University of California, Berkeley, Caltech, and Google DeepMind. Across the table, the theme was the same: the real frontier now is control, evaluation, and efficiency, not just scaling. The conversations at NeurIPS made clear that the next phase of AI will be defined by rigor, accountability, and smart systems design, and this is exactly the work we have been leaning into for years. #NeurIPS2025 #NeurIPSanDiego .406 Ventures Graham Brooks
-
Graham Brooks posted thisThere’s been a lot of chatter this week about the White House’s (leaked) draft executive order aimed at preempting state-level AI laws. While nothing has been finalized (as least as of this morning), it’s an important moment in the broader conversation about how the U.S. will regulate AI and whether we’ll remain competitive on the global stage. Europe’s experience with GDPR is a cautionary tale: well-intentioned regulation that ultimately hampered the region’s ability to innovate in AI. Meanwhile, China continues to advance rapidly without the political and regulatory friction we face here. If we want to lead, we need a coherent federal framework and not 50 different versions of what “acceptable AI” looks like, especially when it comes to how data can be used for training and fine-tuning. That’s the real risk today. A patchwork of state regulations, especially around data use, could severely constrain our ability to build and deploy state-of-the-art models. And the political history here matters: when federal preemption was introduced as part of the Big Beautiful Bill, it was struck down 99–1 in the Senate, but that wasn’t a referendum on the merits of a national standard. It was procedural. The mechanism was wrong, not the idea. An executive order isn’t perfect either, but it does kick-start federal alignment while Congress explores slipping AI language into the year-end defense bill, likely the most viable legislative vehicle on the table. Even in early drafts cited in the media, the direction is clear: federal agencies would be directed to challenge state laws that may interfere with interstate commerce or conflict with emerging federal standards. That’s already creating tension, especially among Republicans torn between pro-innovation instincts and long-standing commitments to states’ rights, hence the “big tech bailout” critiques. But for those of us building and backing the next generation of AI companies, one thing is obvious: establishing a unified national framework is the single most important regulatory step the U.S. can take to compete globally. It’s also a big part of why I joined the National Venture Capital Association Board, to advocate for innovators and help ensure the U.S. maintains leadership in the AI technologies we pioneered. The sooner we get clarity, the more room our founders will have to innovate and the more competitive the U.S. ecosystem becomes as a result. #AI #AIRegulation #AIPolicy #ExecutiveOrder
-
Graham Brooks reposted thisGraham Brooks reposted thisIndico Data today announced a strategic investment from Aviva, the corporate venture capital fund for Aviva plc, one of the UK’s largest insurers. The investment reinforces Indico’s growing leadership in the London Market and our expanding adoption among global property and casualty carriers. This investment builds on Indico’s growing footprint among top global carriers, who leverage its Agentic AI platform to automate underwriting, claims, and operations processes that depend on unstructured data. https://lnkd.in/gx2ubtnN
-
Graham Brooks shared thisProud to back Mona Rakibe, Maxim Lukichev, and Telmai as they launch the #DataReliabilityAgent. At the frontier of AI, trust at the data layer is no longer optional—it’s a prerequisite for scale. The difference between AI and BI is how the bad data shows itself. With BI, you get a chart that looks wrong, signaling you that you have a data problem. With AI you get increased hallucinations and bad decisions, increasing the outliers where AI doesn't work by an order of magnitude. With AI, it is hard to tell if it is a data problem or something wrong with the model or prompt. For AI to deliver on it's promise, it needs accurate data. Check out the video below to see how Telmai's data reliability agents is ensuring that your AI is built on a stable data foundation.Graham Brooks shared thisWe’ve rebuilt Telmai for the new world of AI 🌍 Over the past year, I found myself on what sometimes felt like an insane pursuit of clarity: Where does data reliability truly stand in the age of autonomous AI agents? As these systems shift from analyzing to acting, unreliable data doesn’t just mean poor insights; it means poor decisions made at machine speed. That realization pushed us to rethink everything : our product, our approach, and our vision. Today, we’re excited to share the early preview of Telmai’s Data Reliability Agent 💃 🎉 🥂 This milestone belongs to our incredible team, always learning, adapting, and delivering at speed while staying true to our culture & values. Deep gratitude to our customers who light the way for us, our partners who march with us towards this change, our investors who believe in the vision, and the broader community that has been our guiding star on this journey 💚 TL;DR --Telmai’s Data Reliability Agent brings: ✅ Continuous validation at ingestion (structured, semi-structured & unstructured) right at the landing i.e Data Lakes(Iceberg, Delta & Hudi) ✅ MCP-compliant signals flowing in real time to Humans & Agents, with 7 specialized subagents (Help, Validation, Diagnosis, Configuration, Insights, and more) ✅ Natural language chat interfaces so both business and technical users can interact with data trust directly Watch the launch video: https://lnkd.in/g9GH4KsQ We’d love your thoughts, and if you’d like early access, reach out to us. Read the press release(comment) #AgenticAI #DataReliability #Lakehouse #AIReady #dataobservability #datamonitoring #OpenArchitecture
-
Graham Brooks shared thisLast week I had the chance to participate in two excellent gatherings of data and AI leaders: the first was a Chief Data & AI Officer dinner in Chicago hosted by ICONIQ and Russell Reynolds Associates; the second was further west at #connectedstack in San Francisco, which brought together founders, executives, and investors shaping the future of enterprise AI, including CDOs from the .406 Data + AI Executive Council. Both were energizing opportunities to engage with peers and share perspectives on where this market is headed. Across both rooms, several themes consistently surfaced. There’s a clear shift toward C-level focus on defining ROI for AI investments, moving away from “AI at all costs” toward measurable value creation. Yet despite that focus, enthusiasm for the biggest models remains strong, even as most organizations have yet to seriously explore small language models and open-source approaches; this is something we expect to change as US-based open-source models start to accelerate again. AI security was another key theme, with broad recognition of the new attack vectors this technology introduces, an area where .406 continues to lean in given our deep network at the intersection of AI and cybersecurity. Finally, conversations often turned to the recent MIT paper and its suggestion that cracks may be forming in the #AIhype cycle. My perspective: ROI from AI is very real but capturing it requires a more thoughtful approach to integrating it effectively into existing business processes and work patterns than many expected. In the current hype, we are seeing too many ‘copycat’ businesses with minimal differentiation, leaving buyers with an inability to make data driven purchase decisions. A slowing in the AI hype would be healthy and help buyers focus on high-ROI innovation and allowing the most impactful solutions to rise above the noise. At .406 Ventures, we’re committed to backing the entrepreneurs building those high-ROI solutions — the ones reimagining how enterprises apply AI in ways that are secure, scalable, and transformative. #DataAIdinners #ConnectedStack #AI #Data #AIsecurity #DataLeadership #EnterpriseAI #AIROI Tommy Dwyer, Andrew Bauer, Michelle Gansle, J.D. Williams, Ph.D., Michael Oppenheim, Naveed Asem, Louise Keely, Naveen Singla, Mark Peng, Russell Garner, Danny Greene Jon Callaghan, True Ventures
-
Graham Brooks shared thisThanks to all of our #Founders, #LPs and the many ecosystem partners for helping us find and build industry defining companies. So excited to partner with all of you and most importantly the full team at .406.Graham Brooks shared thisWe’re honored to be ranked #60 on TIME’s list of America’s Top Venture Capital Firms of 2025. Big thanks to the founders who trust us with their boldest ideas, the LPs who back our work, and the .406 team who brings it every day. This recognition is ours together. We’re lucky to do what we do, and even luckier to do it with such incredible people. Thanks to TIME for the spotlight. https://lnkd.in/eCbENF3B Liam Donohue, Graham Brooks, Payal Agrawal Divakaran, Greg Dracon, Trip Hofer, Kathryn Taylor Reddy, Kevin C. Wang, Rebecca Redfield, Marin Lang, Austin Kwoun, Joanna Skoler Gilman, So June Min, Joe SantaBarbara, Esther Dominguez, Kelci Horan
Graham Brooks replied to a comment
1d
James Redfern thanks so much for shepherding Reltio through to this amazing exit. Has been awesome collaborating with you.
Graham Brooks replied to a comment
1d
Connie Puglia Thanks so much! Been awesome working with you and am in awe of the team you have built at Reltio.
Graham Brooks replied to a comment
1d
Manish, thanks again for enabling us to be on this incredible journey with you. Your vision from day 1 was spot on and so excited for you and the team!
-
Graham Brooks liked thisGraham Brooks liked thisI'm building a tech solution to one of Europe's biggest unsolved problems. Come talk to me about joining early — I'll be at the Kiuas x byFounders x Null Fellows event at Maria 01 Thursday, April 9 at 16:30. Currently looking for hungry GTM people, engineers who want real ownership, and early-stage generalists who want to grow fast. Link to join in the comments.
Experience & Education
-
.406 Ventures
View Graham’s full experience
See their title, tenure and more.
or
Licenses & Certifications
Patents
-
System and method for musical sonification of data
Issued US 7138575
See patentA musical sonification system and method sonifies data, such as real-time financial market data, to produce an audio signal output including a musical rendering of the data. A sonification engine converts the data into sound parameters based on configuration data and a sonification and mapping scheme. A sound generator generates the audio output signal from the sound parameters.
-
Infant Feeding System
Filed US 20080179311
See patentAn apparatus for heating a liquid is described. The apparatus includes a housing. A liquid reservoir is contained within the housing. A fluid circuit conveys liquid from the reservoir to a dispenser. A thermal energy storage unit is contained within the housing and disposed to be in thermal contact with the fluid circuit. The thermal energy storage unit is constructed and arranged to heat liquid as it passes through the fluid circuit. The thermal energy storage unit can include a phase change…
An apparatus for heating a liquid is described. The apparatus includes a housing. A liquid reservoir is contained within the housing. A fluid circuit conveys liquid from the reservoir to a dispenser. A thermal energy storage unit is contained within the housing and disposed to be in thermal contact with the fluid circuit. The thermal energy storage unit is constructed and arranged to heat liquid as it passes through the fluid circuit. The thermal energy storage unit can include a phase change material or a combination of a phase change material and a high thermal conductivity material.
Other similar profiles
Explore more posts
Explore top content on LinkedIn
Find curated posts and insights for relevant topics all in one place.
View top content