Feb 27, 2026 · 49m · neon-show
AI Needs to Know Why You took THAT decision | Ashu Garg, Investor at Foundation Capital
0:00 / 0:00
▶ Watch on YouTube →
gold bands on the timeline = statements, start to end. Hover to read, click to jump. CC turns on captions
In this in-depth interview on The Neon Show, venture capitalist Ashu Garg of Foundation Capital unpacks the revolutionary paradigm of context graphs as the persistent institutional memory layer enabling autonomous enterprise AI agents. He outlines how capturing decision traces will disrupt legacy SaaS systems of record, shift enterprise value toward workflow orchestration, and unlock a multi-trillion-dollar market opportunity.
How this conversation actually went
Every chapter scored 0–10 on four independent dynamics. Hover any point for the reasoning behind the score. How this is scored →
Siddhartha as informed peer 4.0 Guest teaching 5.4 Guest disagreement 1.5 Siddhartha pushing back 1.6
speaking balance: gold is Siddhartha, purple is the guest (3 minute bins)
Sharpest disagreement ▶ 28:52 Guest firmly rejects the backseat storytelling premise
Ashu directly and forcefully disagrees with Siddharth's proposition that storytelling has taken a backseat to raw technical capability, arguing that storytelling remains the bedrock of human enterprise.
Hardest push from Siddhartha ▶ 14:04 Host challenges startup access to core enterprise moatsSiddharth pushes back against the viability of third-party context graphs by questioning why any rational enterprise would hand over its proprietary institutional moat to an outside vendor.
Biggest teaching moment ▶ 35:26 Guest deconstructs the winner-takes-all AI narrativeAshu systematically dismantles the common narrative that AI foundation models will consolidate into a single monopoly, citing the multi-model parity shift between OpenAI, Gemini, and Anthropic.
Siddhartha holds their own ▶ 20:26 Host articulates real-world agent management from portfolioSiddharth demonstrates deep practical domain expertise by describing how portfolio startup Buddy implemented UI-less AI teammates that actively manage human sales workflows.
the scores for every segment, with the reasoning behind each
| Chapter | Topic | Siddhartha as informed peer | Guest teaching | Guest disagreement | Siddhartha pushing back | Why |
|---|---|---|---|---|---|---|
| The Technical Anatomy of Enterprise Decision Traces | 3 | 6 | 1 | 1 | Ashu breaks down the foundational mechanics of decision traces and how they aggregate into context graphs. Siddharth acts as an active listener and prompts Ashu on why the thesis resonated so strongly. | |
| Engineering Context Graphs and Capturing Institutional Memory | 5 | 5 | 1 | 2 | Siddharth offers the concrete analogy of Amazon leadership principles to frame institutional memory. Ashu explains how human interpretation of principles forms the decision traces needed to train agents. | |
| The Continuum of Automation in Enterprise Dealmaking | 4 | 6 | 1 | 2 | Siddharth poses practical questions about how agents capture tacit interpersonal context outside digital chat logs. Ashu explains the value-earning staircase where automation wins customer data access incrementally. | |
| Digital Twins and Individual Institutional Memory | 3 | 6 | 0 | 1 | Ashu illustrates digital twins through his portfolio company Vivun, explaining how personal state and institutional memory are captured. Siddharth asks about reconciliation deltas and the evolution from query tools to decision engines. | |
| Securing Context Graphs and Managing Access Governance | 4 | 5 | 2 | 3 | Siddharth challenges Ashu on why an enterprise would share its core competitive moat with an external startup. Ashu reframes the concern, emphasizing identity controls, pairwise context, and PII-stripping infrastructure like Skyflow. | |
| Disrupting Systems of Record and Workforce Roles | 4 | 5 | 2 | 1 | Ashu outlines how systems of record will lose value capture to context graph platforms, citing Tessera's massive SI contract displacement. Siddharth follows along and explores the shifting dynamic between agents and human workers. | |
| Systems of Agents and Underpinning Data Infrastructure | 6 | 4 | 0 | 1 | Siddharth demonstrates domain knowledge by sharing his portfolio company Buddy's AI teammate model and querying underlying infrastructure like Databricks and Snowflake. Ashu details architectural trade-offs across Postgres, Neo4j, and SLMs. | |
| Market Landscape and Real-World Digital Twin Testing | 4 | 4 | 1 | 2 | Siddharth inquires whether vertical or horizontal context graph platforms will win out. Ashu explains that multiple winners will emerge across both domains, sharing his personal experience beta-testing his own digital twin. | |
| Agent Swarms and the Observability Layer | 5 | 5 | 0 | 1 | Siddharth brings up Arise, highlighting its strategic position in the observability infrastructure layer. Ashu elaborates on the necessity of model-based evaluation police to monitor agent swarms and manage hallucinations. | |
| The Dual Imperative: Deep Tech and Storytelling | 3 | 7 | 5 | 2 | Siddharth suggests that storytelling has taken a backseat to deep technical capability in the agentic era. Ashu flatly rejects the premise, explaining that storytelling remains the core differentiator across customers, investors, and talent. | |
| Executive Playbook for Enterprise Context Graph Adoption | 3 | 6 | 1 | 1 | Ashu delivers an actionable executive playbook for CIOs, CFOs, and VPs of Sales on how to leverage context graphs. Siddharth prompts him on enterprise readiness and adoption starting points. | |
| Macro Venture Theses: AI, Blockchain, and Model Economics | 3 | 7 | 5 | 2 | When Siddharth asks if AI is becoming a winner-takes-all market, Ashu directly disagrees by pointing to recent competitive resurges from Gemini and Anthropic. He outlines Foundation Capital's dual macro-bet on AI and blockchain. | |
| Solving Memory Systems and Enterprise Graph Topology | 4 | 6 | 2 | 2 | Siddharth probes whether persistent memory must be solved before context graphs can flourish. Ashu explains that memory is already multi-faceted and that enterprises will maintain federated topologies of multiple intersecting graphs. | |
| Incumbent Strategies and Human-in-the-Loop Governance | 4 | 5 | 2 | 2 | Siddharth suggests startups have an innate advantage because incumbents must rewrite their entire stack. Ashu provides a nuanced counter, explaining why incumbents like Salesforce and ServiceNow retain distribution power while facing architectural trade-offs. | |
| Compounding Decision Traces and Future Horizon | 4 | 6 | 1 | 2 | Siddharth brings up industry skepticism from figures like Dharmesh and Arvind regarding agent reliability timelines. Ashu argues that decision traces compound iteratively, making current primitive implementations rapidly improve. | |
| The Orchestration Layer as the Ultimate Value Capture | 5 | 4 | 0 | 1 | Ashu explains why orchestrating workflows is the only way to infer the 'why' behind decisions. Siddharth caps the discussion with a concrete portfolio example (Pit Crew) building orchestration for wealth advisors. |
Statements from this episode (32)
Ashu Garg: Context graphs serve as an organization's institutional memory
“Think of a context graph as the institutional memory for an organization. At its simplest level, it is an aggregation of the knowledge of why decisions were made, how they were made, who made them. And the why often has to be inferred from the how and the who.…”
Garg: Systems of record capture outcomes rather than decision traces
“Historically, organizations do not have these decision races in a form that can be processed. They have the outcome. So systems of record capture the output of these decision traces.”
Garg: Winning AI startups will build moats through context graphs
“Our belief is that the most successful AI startups of the future will be those that are able to build a moat by building a unique context graph.”
Ashu Garg: Foundation Capital backs Tessera, Olive, and Reggie for context graphs
“There is Tessera who's doing it for ERP migrations and automations. There is Olive that is doing that for sales and Reggie for demand generation. So we have many examples of these, this in our portfolio, and there are many examples outside our portfolio to be …”
Garg: Fully autonomous agent transactions require context graphs
“To get to that, and that might not make sense for everything, but there are many situations where that makes sense. To get to that nirvana, we need context graphs.”
Garg: Context graph winners must deliver value to earn data access
“We believe that the best, the companies that win in the era of context graphs will find a way to deliver enough value that they earn the right to build a context graph.”
Ashu Garg: Context graphs cut finance reconciliation exceptions by up to 80%
“So look, obviously this varies from customer to customer, but the common pattern we're seeing is we're seeing a 30 to 80% reduction in exceptions.”
Garg: Foundation Capital backed Viven, founded by Eightfold's co-founders
“I have a company called Vivid, started by another Ashu guard and Varun Kacholia. So Varun and Ashu, who are also the founders of Eightfold, had this idea that We need to capture the institutional memory of a human as a digital twin.”
Ashu Garg: Enterprise users connect AI to emails, not Zoom calls
“Most people give it access to email. They give it access to their Google Docs, or SharePoint, or Box, Dropbox folders. Most people don't want to start out by giving access to all Zoom calls. Or phones.”
Garg: Digital twins can answer 20% to 80% of workplace questions
“So depending on the role, depending on the person, maybe it's 20%, maybe it's 80%, but somewhere between 20 and 80% of the questions you have are answerable by a digital twin in that case, because that twin has built an institutional memory of you, the individ…”
Garg: Context in AI is a function of pairwise relationships
“Because context is a function of pairwise relationships. When I ask you a question how things are going, your wife asks you the question how things are going, and your CFO asks you the question how things are going, you have very different answers to the same …”
Ashu Garg: Public SaaS stocks collapsed 30% to 40% in six weeks
“You know, in the last six weeks since we talked about context graphs SaaS stocks have collapsed 30 to 40%.”
Garg: Systems of record will capture a declining share of tech value
“Existing systems of record are not the natural owners of the context graph, and so they will capture a declining share of the value that's created by technology.”
Ashu Garg: AI startup Tessera scaled from zero to $25M in bookings
“Tessera won a twenty-five million dollar contract from a Fortune 50 corporation, and they went from zero to twenty-five million dollars in bookings.”
Garg: I was a seed investor in Databricks
“I was very lucky to be a seed investor in Databricks specifically”
Ashu Garg: Over 100 winners will capture the $1T context graph market
“So I think there will be the trillion dollar plus opportunity will get captured by a hundred plus winners in context graphs.”
Garg: Foundation Capital backed healthcare context graph startup Tenor
“And we're investing in both, you know, for example, my partner Joanne has a very successful context graph company in Tenor, which is going after healthcare.”
Garg: Superhuman AI requires breaking aggregate tasks into multi-agent systems
“So I think you can have superhuman performance. For specific classes of tasks, but it requires us to break down the aggregate task into smaller tasks, have agents to do smaller tasks.”
Foundation Capital: AI investing focuses on agent systems and context graphs
“So we are either investing in two things. We're either investing in systems of agents, or we're investing in enabling infrastructure. The common theme across these is the context graph.”
Garg: Databricks took three to four years to reach product-market fit
“I mean, Databricks started in 2013, but it took three, four years, and a lot of going around in circles for them to get to the point where they had product market fit.”
Ashu Garg: Context graph adoption will dictate S&P 500 winners by 2030
“I believe that four years from now, The winners and losers on the SNP or the Nifty in India, the winners and losers will be separated by their ability to embrace context graphs.”
Ashu Garg: CIOs migrating ERPs without context graphs are failing their duties
“If you are the CIO organization in a Fortune hundred company, and you are thinking about an SAP or an Oracle migration, well, call one 800 Tessera, because if you are looking to drive a major, you know, system of record migration without exploring What, you kn…”
Ashu Garg: Foundation Capital was the first investor in Solana
“We were the first investors in Solana.”
Garg: AI Model Layer Is Not a Winner-Takes-All Market
“So this is by no means, even the model layer is not a winner-takes-all market.”
Garg: Every organization will operate multiple context graphs, not one
“First and foremost, every organization will have multiple Context Graphs. There is no one world model or one world context graph in any organization.”
Garg: Every Global 2000 enterprise will build in-house context graphs
“And every global 2000 company will build many context graphs in house.”
Garg: Context graphs are built on multi-system processes, not single-vendor data
“Context graphs are not bound by system constraints. You're not going to build a context graph on top of your Salesforce data. You're going to build a context graph on top of your sales process. And every business process touches dozens of systems in a large co…”
Ashu Garg is actively buying Salesforce, Snowflake, and ServiceNow stock
“And so I am bullish that they will also capture some share of this value. And so I, I'm a buyer of Salesforce and I'm a buyer of Salesforce stock. I'm a buyer of Snowflake stock and I'm a buyer of ServiceNow stock.”
Garg: Context graphs will be 10 to 100x better in three years
“The context graphs of three years of, you know, from now would be 10 to a hundred X better than today's context graphs, because the enabling technology will make them better. And the decision traces you capture will make them better.”
Garg: Enterprise decision traces compound to enrich context graphs
“See, key to the context graph idea is the notion that decision traces, you know, compound. As you capture decision traces, you earn the right to capture more decision traces. And that compounding of decision, of capturing decision traces is what makes for a ri…”
Garg: Decision traces can only be captured in the workflow orchestration layer
“Decision traces can only be captured if you're in the orchestration or you're in the flow of those decisions.”
Garg: Database and analytics vendors cannot capture enterprise decision traces
“A database vendor can't capture them because no, they're not in any one database. An analytics vendor cannot capture them because it's, they capture the outcome, not the why and the how. And the only people who capture the how decisions are made are the orches…”