Podcast episode

Rethinking Legacy Data Infrastructure with Eon Co-Founders Ofir Ehrlich and Gonen Stein

by Ken Rona ·August 28, 2026 ·No Priors

agents build-vs-buy data-brokers open-weights security

TL;DR

A founder-focused episode on No Priors with Eon's co-founders discussing enterprise data infrastructure in the AI era — covering data as a competitive moat, AI agent security risks, and the breakdown of legacy ETL pipelines. Relevant for AI infrastructure and enterprise AI operators; not a frontier-lab or model-release episode.

What was covered

Google's acquisition of Spirit Airlines' data out of bankruptcy for ~$10M — framed as a signal that AI labs and tech majors are now actively bidding on enterprise datasets from distressed companies, with the rumored competing bidder being Mercor (referred to as "Merkur" in transcript). The thesis: real-world enterprise data is scarce and valuable for agent training because synthetic data doesn't capture organizational complexity.
Enterprise data as a competitive moat — founders argue that models and compute have near-zero switching costs, making proprietary historical data the only durable differentiator. AI labs are reportedly approaching hedge funds and other data-rich institutions to acquire datasets.
The "locked data" problem in enterprises — data is scattered across business units with conflicting incentives, legacy systems, and unclear ownership. Eon's pitch is mapping, classifying, and ingesting this scattered data into a unified foundation that connects to AI/LLM workflows without compromising security or production uptime.
AI agents as a new ransomware-equivalent threat vector — agents with legitimate system permissions can drop database tables or exfiltrate data at machine speed. Eon detects this using the same irregular write-pattern and entropy-change methodology used for human ransomware attacks, but notes the velocity is far higher.
Non-human identity (NHI) proliferation — agents spawning agents inside enterprises create chains of responsibility that are nearly impossible for human IT/security teams to audit. CISOs are caught between board pressure to enable AI and existential fear of data leakage.
Cloud era vs. AI era infrastructure shift — founders (who previously sold CloudEndure to AWS, where it became Application Migration Service) argue the AI transition is the cloud transition "on steroids," but compressing timelines from years to months and generating far more data noise that old ETL tools (Fivetran, dbt) weren't built to handle. Databricks cited as the closest incumbent adapting to the new paradigm.
Forward-deployment engineering as a new GTM motion — large legacy enterprises unable to self-adopt AI are now bringing in startup engineers in a Palantir-style deployment model; companies like Cognition cited as using product-led growth followed by forward deployment into banks.

Notable claims & predictions

Ofir Ehrlich: "Models, compute — everything has relatively almost zero switching costs. The most valuable thing a company has is actually their data." Implication: data ownership, not model choice, determines long-term AI advantage.
Ofir Ehrlich: "Six months ago no one would even discuss [AI agents as an internal threat] with me. A few months ago, pretty much every leader I meet tells me either they are afraid of that happening or it personally happened to them." — Signals rapid escalation of agentic security incidents in enterprise environments.
Gonen Stein: "It creates a complete set of actors inside the organization, not bound by the rules of the organization and not necessarily running within the premises of the organization, but handling sensitive data." — Describing non-technical employees building agents via no-code tools (e.g., Lovable) that chain to external networks and internal data stores simultaneously.
Ofir Ehrlich on Harvey's legal dataset release: "You don't find too many good data sets that don't look like really synthetic data that can actually be used to really look like the real world." — Identifies real-world enterprise data scarcity as a hard constraint on agent training quality.
Ofir Ehrlich: "We're not in the time of token maxing anymore — now trying to go to actually getting value from every token because it becomes more and more expensive." — Signals inference cost pressure shifting enterprise AI strategy from exploration to ROI discipline.

Why this matters for AI operators

Data acquisition is becoming a structured M&A/distressed-asset play. The Spirit Airlines example, combined with reports of labs approaching hedge funds, suggests enterprises sitting on large historical datasets should expect inbound interest — and should assess the value and accessibility of their own data before competitors do.
Agentic security is no longer theoretical. Founders report that AI agents with legitimate permissions causing data loss or corruption is already happening frequently enough that it's now a top-two CISO concern alongside NHI management. Operators deploying autonomous agents in enterprise environments need detection and recovery infrastructure calibrated for machine-speed threats, not human-speed ones.
Legacy ETL/data-engineering stacks are not fit for agentic workflows. Tools like Fivetran and dbt were designed for purpose-built, human-directed pipelines. The volume, variety, and agent-generated noise in modern enterprise data requires new cataloging, semantic classification, and access-control layers before LLMs can reliably operate on it — a gap that represents both a risk and a vendor opportunity.
Inference cost discipline is tightening enterprise AI strategy. The shift from "maximize tokens/context" to "extract value per token" reflects real budget pressure as enterprises move from pilots to production. AI infrastructure vendors and applied-AI teams should expect procurement scrutiny to intensify around per-query costs and ROI demonstration.

Full analysis

Eon co-founders Ofir Ehrlich and Gonen Stein went on No Priors with Elad Gil to argue that in the AI era, your data is the only moat left. Models and compute have near-zero switching costs. Historical enterprise data does not. The proof points: Google buying Spirit Airlines' data out of bankruptcy for ~$10M, labs approaching hedge funds for financial data, and AI agents becoming an insider ransomware threat that moves at machine speed.

This is a briefing, not a decision. So the question for a dev manager shipping AI into production: does any of this change what your team builds next quarter, or is it a vendor pitch wearing a trend?

Reversibility: Mostly Type 2. Trying a data-cataloging layer or an agent-access-control tool is reversible. The one Type 1 lurking here is agent permissions architecture. Get that wrong and you are cleaning up after a machine-speed incident, not rolling back a config.

Forcing function: None hard-dated. But the security claim ("six months ago nobody discussed this, now every leader has a story") is the part with a clock on it, if it is true.

The Skeptic

Two founders selling data infrastructure tell you data is the only moat. Of course they do. That is the pitch. The $10M Spirit deal is real and interesting, but one distressed-asset buy does not make a market. Airline pricing and ops data is genuinely rich. Most enterprise data is a swamp of duplicated CRM exports and Slack logs nobody would pay for. The "every leader has an agent horror story" claim is exactly the kind of unfalsifiable urgency a vendor manufactures. Where are the incident reports? The named breaches? For a PM: they are telling you the house is on fire, and they sell extinguishers.

The Researcher

The durable claim here is synthetic data hitting a ceiling for agent training. Ehrlich's line on Harvey's legal dataset ("you don't find too many good data sets that don't look like really synthetic data") points at something real: agents need traces of actual organizational decisions, messy handoffs, edge cases, not clean generated examples. That is why labs are bidding on operational data specifically. But note what is NOT claimed: nobody says a model trained on Spirit's data beat one that wasn't. The moat thesis is asserted, not measured. For a PM: the labs think real-world process data is scarce and worth money, and they are voting with acquisitions.

The Compute Pragmatist

Ehrlich buried the most actionable line in the episode: "We're not in the time of token maxing anymore." Translation for your team: the era of stuffing giant contexts and letting the agent wander is over because every token costs money and enterprises now count. This is the shift from pilot to P&L. Your inference bill is now a procurement line item somebody defends in a QBR. That reprices architecture decisions you already made. Long, chatty agent loops that looked clever in the demo get value-engineered down. Retrieval that fetches only what is needed beats dumping the whole knowledge base into context.

The Builder

Forget the moat philosophy. What ships Tuesday? The agent-permissions problem is the one real thing here. An agent with legitimate credentials that can drop a table is a permissions design failure you can fix now. Scope agent identities tight, separate read from write, log every non-human action, put a human gate on destructive operations. You do not need Eon to do this. You need the discipline your DBAs already have, applied to service accounts that now think for themselves. The "agents spawning agents" chain is the genuinely hard part: once a no-code tool built by marketing calls three external agents, your audit trail is gone.

The Open-Source Advocate

The whole moat argument quietly hands the win to open weights. If models commoditize and data is the differentiator, then a fine-tuned Llama or Mistral on your proprietary data beats renting a frontier model you share with every competitor. That is the strategic read the founders gesture at but do not spell out: Elad Gil's aside about open-source models fine-tuned against proprietary datasets is the actual playbook. You do not buy the moat. You already own it, sitting on backup tapes. The question is whether your data is clean enough to train on, which is, conveniently, what Eon sells.

Where they disagree

The Skeptic and the Researcher split on the moat. The Researcher sees labs paying real money for real-world process data, which is evidence. The Skeptic notes nobody has shown that data produced a better model, so the price could be labs hoarding out of FOMO, not proven value.

The Builder and the vendor pitch split on the security fix. The founders frame agent threats as a new category needing new detection infrastructure. The Builder says most of it is old-fashioned least-privilege access control that your team can implement without buying anything.

And the Compute Pragmatist quietly undercuts the whole "deploy AI immediately" board panic: if every token now has to justify itself, the forward-deployed, ship-it-fast motion collides with the ROI discipline the same founders admit has arrived.

What this actually hinges on

Is real-world enterprise data a durable moat, or a temporary scarcity labs are overpaying to fix? If it is durable, the open-source-plus-your-data play is the correct long-term architecture and you should start treating your data pipeline as a strategic asset now. If it is temporary, this is a land grab that gets repriced the moment synthetic data closes the gap.

For your team, the low-regret moves are Type 2 and don't depend on resolving that: tighten agent permissions and non-human identity logging this quarter, and start measuring cost-per-task on your agent workflows before procurement does it for you.

Prediction: By the time OpenAI, Anthropic, or Google ship their next frontier model class in the first half of 2027, at least one major lab will have publicly disclosed a paid acquisition of a real-world enterprise operational dataset (beyond web scrapes and licensed publisher/media content) explicitly for agent or model training.

Confidence: Medium. The Spirit deal and hedge-fund approaches show the pattern is already forming.

Why: Google already paid ~$10M for Spirit's data out of bankruptcy and a competing lab reportedly bid against them, which means the behavior is real and contested, not one-off. The stated mechanism is that agent training needs traces of actual organizational decisions that synthetic data cannot fake, and labs are running short of fresh web text. The opposite outcome (labs staying quiet about such buys) is plausible because these deals are legally messy and competitively sensitive, but the same competitive pressure that made Google outbid a rival also creates incentive to signal to other data-holders that there is a buyer, so at least one disclosure is the likelier path.

Revisit by 2027-06-30: We're right if a major lab confirms a paid deal for proprietary operational/transactional enterprise data for training. We're wrong if the only disclosed data deals through that window remain web content, publisher licensing, or media archives.

One more thing worth saying plainly. The most useful line in the episode wasn't about moats. It was Ehrlich admitting the token-maxing era is over. That is the shift that actually lands on your team's roadmap, and it arrived while everyone was still arguing about who owns the data.

Listen to the episode →