VC Survey 2026
“AI is making output per person the real scaling metric; headcount growth is now often a lagging indicator.”
Lee Ben-Gal and Haim Bachar, Investors at lool ventures, joined CTech for its 2026 VC Survey.
“In software-heavy companies, AI is making output per person the real scaling metric; headcount growth is now often a lagging indicator (or even a sign you haven’t redesigned the workflow),” said Lee Ben-Gal and Haim Bachar, Investors at lool ventures.
Asked if AI has broken the correlation between headcount growth and company success, they responded: “In Israel you can already see the decoupling: high-tech employment has been basically flat as a share of the workforce (about 11.5%), while high-tech exports hit record levels and 2026 is expected to be an all-time record year for exits/M&A value."
Following the turbulence of recent years and the stabilization of 2025, the Israeli tech ecosystem is entering a new era: The Next Leap. The investors joined CTech to share insights for its VC Survey 2026, which invites prominent investors to discuss the topics, trends, and “leaps” expected in the year ahead.
You can read the entire interview below.
Fund ID
Fund Name: lool ventures
Total Assets Under Management: $200M
Partners/Managers: Avichay Nissenbaum and Yaniv Golan
Notable Portfolio Companies: Eleos Health, NoTraffic, Beewise, Ai.Work, ControlMonkey, ZeroPort
Notable Exits: Atero -> Crusoe; Voca -> Snap; Mabaya -> Criteo; Wibbitz -> Vimeo; Shopial -> Magento
The Valuation Leap: Moving past the market correction, what is the single most critical metric (e.g., EBITDA, NRR) that will drive premium valuations in 2026?
In 2026, the premium valuation metric is the Rule of X - a growth‑weighted Rule of 40. It’s a single score that reflects what the market is actually paying for now: efficient velocity, where growth matters roughly 2–3× more than free cash flow margin as long as the growth is healthy.
The Agentic Leap: As we transition from 'Copilots' to autonomous 'Agents,' which specific vertical will be the first to fully trust AI with independent decision-making?
If you look for the first vertical to “fully trust” autonomous agents, it will be customer service/contact centers - specifically post‑purchase resolution. But “fully trust” won’t mean open‑ended autonomy; it will mean delegated authority inside tight guardrails: clear policies, spend/credit limits, identity checks, audit logs, and easy rollback/escalation.
This is the best proving ground because it matches the conditions where trust forms fastest: extremely high decision frequency, crystal‑clear success metrics (resolution time, cost per ticket, CSAT, containment), and actions that are usually reversible. In practice, an agent can independently approve refunds up to a defined cap, ship a replacement, apply credits, reset access, update a subscription, and close the ticket - while routing edge cases (fraud signals, policy exceptions, high‑value disputes) to a human.
It’s also where agentic ROI is easiest to prove and govern. Gartner is already warning that many “agentic AI” initiatives will be canceled if business value and risk controls aren’t real - so starting in customer support is the most pragmatic path to build trust, demonstrate savings, and expand autonomy from there.
The Sovereign Leap: Have the geopolitical lessons of recent years pushed Israeli startups to build independent, 'sovereign' tech stacks to reduce reliance on global platforms?
Not broadly. Most Israeli startups still build on the big hyperscalers because speed, managed services and global distribution win- and in AI they’re also increasingly dependent on a small set of foundation model providers like OpenAI, Anthropic, and Google. What has changed is the mindset: founders and boards now treat both cloud and model APIs as strategic dependencies that need mitigation. So you see more portability and control - owning the sensitive layers (data, identity, encryption keys, logs), reducing hard lock-in, and building credible fallbacks: multi-region/multi-cloud readiness, and in AI specifically the ability to switch model providers, route across multiple models, or run self-hosted/open-weight models for the most sensitive workloads. The goal isn’t full independence - it’s operational resilience if access, policy, pricing, or procurement constraints change.
The Contrarian Leap: What is one sector or trend currently ignored by the herd that you believe represents the most undervalued opportunity for the coming year?
The most undervalued opportunity is queue-busting grid tech - software and power-electronics that let utilities connect new generation and new loads faster and increase hosting capacity on existing distribution networks, because interconnection backlogs and AI-driven demand growth are turning the grid into the real bottleneck.
Finally, what are 2-3 startups that, in your opinion, are likely to make a leap forward in 2026?
ZeroPort (lool ventures portfolio company) - ZeroPort is rethinking secure remote access with hardware-based, non‑TCP/IP connectivity that removes the typical VPN/remote-access attack surface - so critical networks can stay isolated while still being accessible. The fact that it’s already deployed in high-sensitivity critical infrastructure environments is a strong signal, and as third‑party access risk stays at the top of the CISO agenda, this “security by architecture” approach should see accelerating pull in 2026.
Avon AI (lool ventures portfolio company) - As enterprises move from AI demos to agents running in production, the bottleneck becomes trust, control, and governance (not models). Avon AI is building the management layer for AI agents - giving organizations visibility and control in a way that’s designed for regulated, mission‑critical environments and for business teams (not only engineers). That puts it in the path of a major 2026 shift: from experimentation to accountable, scalable deployments.
ai.work (lool ventures portfolio company) - ai.work (founded by WalkMe veterans) is building “AI Workers” that automate internal service workflows across functions by integrating into systems like ServiceNow, Slack, Salesforce, and Microsoft 365. With the market clearly moving from copilots to end‑to‑end autonomous execution - and with ai.work already exiting stealth with a $10M seed round - it’s well positioned for a step-change year as enterprises allocate real budgets to measurable automation outcomes.
