Lazy Digest #8: The Drought Broke in Three Days. Four Companies Took 84% of It.
For three straight issues this digest opened its funding section with a single number and some version of “mid-August is thin, don’t read too much into it.” Issue #5: three rounds, $116 million. Issue #6: one round, $7.5 million. Issue #7: one round that counted, and a note that “August is slow” was wearing thin as an explanation.
Then Israeli kids went back to school on September 1st and the ecosystem announced $1.4 billion across fourteen rounds in six days — more than twice what all of August produced.
So the seasonal explanation was right. It was August.
What the money did once it arrived is the more interesting part. Four companies took $1.18 billion, or 84% of the week. Salesforce turned up as an investor in three separate deals. And not one of those four companies is headquartered in Israel — which is precisely the thread issue #7 closed on, arriving a week later with four fresh examples attached.
Lazy Digest #8 — Israeli tech, one week, researched with AI and filtered by a human who reads the primary sources. Estimates are labeled as estimates. Let’s go.
The Rumor That Finally Landed
We’ve carried this since issue #5, when Globes reported Wonderful was seeking $500 million at a $5 billion valuation. Issue #6 said no update. Issue #7 said it was the last time it would get its own line unless something moved.
It moved. Tuesday, Wonderful announced a $550 million Series C at a $5 billion valuation — above the rumored raise, dead on the rumored price. Insight Partners led for the third consecutive round, Salesforce invested for the first time, and Index Ventures, IVP, Vine Ventures, 9Yards and Bessemer all followed.
The ladder is worth laying out, because the compounding is easy to miss: $34 million seed in July 2025, $100 million Series A in November, $150 million Series B at $2 billion in March, $550 million at $5 billion now. Founded early 2025. Roughly $834 million raised in fourteen months. Headcount went from 350 in March to 650, across 35-plus markets — though in March the company told CTech it planned to reach about 900 by year-end, so hiring is running behind the stated plan even as valuation runs ahead of it.
And when issue #5 covered the rumor, the press-reported revenue was roughly $7 million at the end of last year, up from about $1 million in August 2025. CTech now reports an annual run-rate near $70 million, with the company expecting to pass $100 million by year-end. Company-provided and unaudited — but if roughly right, revenue went up tenfold while the valuation went up two and a half times, which is the rare case where the multiple actually compressed through a round.
What Wonderful sells has changed shape too. It started as a customer-service agent platform for non-English markets and now calls itself an “AI operating system” — an open, model-agnostic layer coordinating agents and workflows across an enterprise. The mechanism underneath is the part I’d study: forward-deployed engineering pods that co-build with each customer, get one use case into production, then hand the capability over. That’s a services motion in software clothes, and it’s both why revenue can grow tenfold in a year and why $5 billion is a genuine bet rather than an obvious one. Does deployment expertise become repeatable margin at scale, or does headcount have to scale with revenue? Insight has now written three checks on the first answer.
Upwind Doubled Again, and Salesforce Was in Three Rooms
Upwind raised $300 million at a $3.8 billion valuation, co-led by Bessemer and TCV, with Craft Ventures, Salesforce Ventures, Greylock, Cyberstarts, Leaders Fund, Alta Park and Steph Curry’s Penny Jar. Bloomberg broke it Wednesday; the formal announcement came Friday. Total raised now exceeds $680 million.
The ladder again, because it’s the whole story: $800–900 million in late 2024 → $1.5 billion on a $250 million Series B in January → $1.6 billion on a Salesforce Ventures follow-on → $3.8 billion now. Roughly $2.3 billion of added valuation in eight months, for a company founded in 2022 by Amiram Shachar and his Spot.io co-founders (Spot.io sold to NetApp for $450 million in 2020). It sells runtime-first cloud security: map the cloud, refresh roughly every 30 seconds using Linux eBPF for kernel-level telemetry, then use live context to tell teams which of their thousands of alerts actually matter. Reported ARR was under $20 million at the end of 2025 and is press-estimated at $50–60 million now.
The demand driver isn’t subtle: Google’s $32 billion Wiz acquisition left a Wiz-shaped hole, and plenty of enterprises would rather not buy cloud security from a hyperscaler.
Put this next to issue #6’s At-Bay section — a cyber-insurance unicorn selling for 43 cents on its 2021 dollar, its CEO on the record saying “the value should be close to the cash the company creates.” Two weeks later Upwind is priced at something like 60x forward revenue. Both are true about the same market in the same month. The 2021 reset never became a rule; it became a filter. Companies growing into an obvious, urgent, well-funded category still get 2021 pricing. Everyone else gets Rotem Iram’s math.
Salesforce also led the week’s third-biggest round: $166 million into HiBob, with Farallon participating and Salesforce providing most of the money, at about $3.2 billion — up from roughly $2.66 billion in the 2023 round Farallon led. Founded in Tel Aviv in 2015 by CEO Ronni Zehavi and CTO Israel David, headquartered in London now, total raised past $700 million. Zehavi says HiBob does more than $400 million in sales, is aiming at $1 billion, and will spend part of this on acquisitions. The pitch is a repositioning — from HR platform to “organizational intelligence layer,” the trusted record of who works here and what they’re allowed to touch. His line on the AI-eats-SaaS thesis: “Do I see a company vibe-coding an HCM to replace a system like us? No way. It’s very complex and the data is very sensitive.” No IPO before 2027.
Three Salesforce appearances in one week — HiBob, Upwind, Wonderful — from a company that has spent two years watching AI agents circle the seat-license business it invented. If you’re guessing who buys Israeli enterprise infrastructure next, start here.
The Other Ten Rounds
The big four took 84%. The rest took 16% and covered more ground than any week this digest has recorded.
Lyte — $165M Series C at $1.6 billion, led by Maverick Silicon with Fidelity, Atreides, Key1 and Ora Global. Total $272 million, up from the $107 million it carried out of stealth in January — a triple in eight months. It builds the perception layer for physical AI: custom silicon plus 4D vision and inertial sensing on one synchronized timeline, so a robot measures motion directly instead of reconstructing it in software. CEO Alexander Shpunt co-founded PrimeSense, sold to Apple in 2013; chairman is Avigdor Willenz (Galileo→Marvell, Habana→Intel). Be precise, though: Lyte was founded in Silicon Valley and sits in Sunnyvale. The Israeli content is lineage, not an office.
AIR Security — $50M out of stealth, across two seed rounds weeks apart: $10 million led by Sequoia, then $40 million led by Greenoaks, with Swish, Netz and an angel list including Wiz’s Yinon Costica, Cognition’s Zach Frankel and former US deputy national security adviser for cyber Anne Neuberger. Founded February 2026 by CEO Yair Saban and CTO Niv Hoffman, both ex-8200; 40 people in Israel, company domiciled in New York. It builds an inline firewall for AI agents — vetting every skill, plugin and MCP server before it touches anything. Its own research found more than 17,800 public AI add-ons, across 6.7 million installs, pulling instructions from untrusted sources.
Sit that against issue #7’s Minimus section. Minimus raised a $51 million seed in 2022, one of the largest in Israeli history, and by 2026 it wasn’t enough — it shut down and gave $10 million back. AIR raised $50 million within six months of incorporation, pre-revenue. Same nominal number, four years apart, and the second is an entry ticket rather than a war chest.
AITAN — $41M, co-led by Deep33 and Dell Technologies Capital, total ~$54 million. Formerly eyesAtop, it builds an OEM-agnostic command-and-control OS for multi-drone operations, with more than 500,000 operational flight hours claimed since October 7th. Co-CEOs Udi Oster and Daniel Almog are ex-8200 who founded Tapingo, sold to Grubhub for $150 million. The company is named for Oster’s brother Eitan, killed in Lebanon in 2024. Continuity note: under its old name this is one of the two Israeli startups that beat Ondas for the command-and-control layer of the Defense Ministry’s Digital Bat program, covered in issue #5.
Guardio — $40M at a $1.1 billion valuation. Led by existing investors ION Crossover, Union Tech, Vintage, Cerca and Emerge, with Assaf Rappaport joining personally. Founded 2018 by Amos Peled, Michael Vainshtein and Daniel Sirota. The company-provided numbers: over a million paying customers, $150 million ARR, four straight years of triple-digit growth — against $167 million raised in total. Roughly a dollar of ARR for every dollar ever raised, which at unicorn scale is close to unheard of here. Peled on why the round was small: “There was no need to dilute us with a larger raise.”
Huskeys — $27M Series A led by Blackstone, with Zscaler Ventures, Okta Ventures, Merlin and SV Angel, plus angels Shlomo Kramer and Eran Reshef. An agentic layer over web application firewalls, built for attacks initiated by agents. Customers named include TikTok and Hugging Face.
Easy Aerial — $20M Series B, with Insight Partners, Entrée Capital and private US investors participating. Brooklyn-based, Israeli-founded, tethered autonomous ISR drones, run by ex-Roboteam’s Shahar Abuhazira.
Particula — $15M led by Haifa’s Elah Fund with Grove Ventures and Flashpoint. Udi Dor’s connected-toy company — GoCube, GoChess, GoBalance, GoDice, over a million units sold — is shifting from direct-to-consumer to retail and licensed IP, including an official Warner Bros. Harry Potter edition of GoChess first announced last year.
DataAgent — $10M pre-seed out of stealth, led by MizMaa and Alicorn. Founded January 2026 by Ishay Yaari and Nati Shalom, both ex-Cloudify (Dell, 2023). Fifteen people building agents that detect and autonomously repair cloud infrastructure failures.
Newlight — $9M seed from lomarlabs, BIRD Energy, Undeterred, CiRi and Fusion VC. Two ex-Israeli Navy officers retrofitting diesel ships with AI-guided hydrogen injection — 24% less fuel, 28% less CO₂ by their numbers — with an 8,500-nautical-mile Singapore-to-Ghana commercial voyage already run and twelve vessels under agreement.
Catch — $5M seed co-led by Entrée and Pitango, deliberately small: an AI executive assistant at $99/month. Flamingo — $4.5M seed led by Vertex: an open-source, AI-native IT platform for managed service providers at $1 per device per month.
Ten rounds spanning physical AI, agent security, defense autonomy, consumer cyber, cloud infrastructure, connected toys, shipping, productivity and MSP tooling. Only two are straightforwardly cyber, and almost none could have existed in this form three years ago.
Three Different Doors Out
Three exits, none of them a conventional trade sale, and between them they map every route to liquidity an Israeli company currently has.
Sell to the strategic that already knows you. ServiceNow acquired Sweep in a deal Calcalist estimates at hundreds of millions. Sweep was founded in 2021 by Ido Gaver and Eran Kirshenboim, who previously built Flok and sold it to Wix; it raised $46 million and builds an agentic workspace that monitors metadata across ServiceNow, Salesforce and HubSpot and increasingly implements fixes rather than flagging them. This is ServiceNow’s fifth Israeli acquisition of 2026 — after Pyramid Analytics, Traceloop, Armis at roughly $7.75 billion, and ai.work, which we covered in issue #1. Eight Israel-linked deals all-time, more than $8 billion. ServiceNow isn’t shopping in Israel anymore, it’s assembling here.
Reverse into a listed shell. CollPlant completed its acquisition of LightSolver on Wednesday — equity-centric consideration, no headline cash price. LightSolver was founded in 2020 by Weizmann physicists Dr. Ruti Ben Shlomi and Dr. Chene Tradonsky and builds an all-optical Laser Processing Unit; Ben Shlomi joined CollPlant’s board the next day. A regenerative-medicine company pivoting into photonic computing is an odd sentence, but look at what it does for LightSolver: a deep-tech company with a multi-year horizon and few natural acquirers just got a public-market vehicle without an IPO. Issue #7 watched SCiO end as a bolt-on after a decade and $64 million raised. Same structural problem — Israeli deep-tech outlives the funds backing it — solved a completely different way.
Merge into a listed company and become one. XTEND began trading on the NYSE as XTND on Friday after closing its all-stock combination with JFB Construction Holdings, a Nasdaq-listed Florida contractor. The combined entity is XTEND AI Robotics, Delaware-incorporated, headquartered in Tampa, valued around $1.5 billion. It raised nothing from public investors: a private investment planned at $152 million finalized near $100 million, $42 million of it a SAFE from the February signing, with the $60 million minimum-cash condition confirmed on September 3rd. Founded 2018 by brothers Aviv and Matteo Shapira with Rubi Liani and Adir Tubi; 267–300 employees, roughly 60% still in Israel.
Then the numbers. XTEND did about $20 million of revenue in 2025 and $5.8 million in Q1 2026 against an $11.5 million loss — against a prospectus projecting $85.6 million for 2026. Because this is a SPAC-shaped combination rather than an IPO, it was allowed to publish those forecasts, and it is visibly behind them. What the valuation actually rests on isn’t in the filing: XTEND’s X-Strike package was accepted into the US Department of War’s Drone Dominance Program in August, and it’s one of 19 companies competing at Fort Carson for a share of 60,000 attack drones. That leaderboard hasn’t been published. Issue #6 framed UVision as the test of whether public markets would pay private defense-tech prices. XTEND found the answer that avoids asking.
Seventy Percent
Issue #7 closed on IVC’s finding that 57.9% of Israeli-tech capital in H1 2026 went to companies headquartered outside Israel. This week Dealroom added a second number, circulated via VC Cafe: 70% of unicorns founded by Israelis were built in the United States.
Then the week demonstrated it. Wonderful datelines from Amsterdam. Upwind is San Francisco per Bloomberg (Tel Aviv per Israeli outlets; genuinely dual). HiBob is London. Lyte is Sunnyvale. Add XTEND formally becoming a Delaware corporation in Tampa, Easy Aerial in Brooklyn, Flamingo in Miami Beach.
Oren Zeev made the case explicitly in CTech this week, without hedging: “I’ve never met a founder who regretted a US incorporation, and I often meet founders who regret an Israeli incorporation.”
He’s probably right as advice, and that’s what makes it uncomfortable. The best counsel for an individual founder and the best outcome for the Israeli economy have come apart. One thing the aggregate numbers hide: the Israeli-domiciled companies in this issue — AIR, AITAN, Huskeys, DataAgent, Newlight, Particula, Guardio — are all small-to-mid. The pattern isn’t that Israeli companies leave. It’s that they leave when they get big, which means the domestic column is structurally made of earlier-stage companies and the foreign column of the valuable ones.
Related, and the reason the domestic column still gets funded: Swish Ventures closed a $250 million third fund, taking Omri Casspi’s firm to roughly $800 million AUM in under two years, with Sequoia Capital among the LPs. The strategy is twelve companies, ~$20 million average check, seed and Series A, in cyber, infrastructure, physical AI, defense and govtech — almost exactly this week’s sector list. Two caveats: that $800 million is AUM including unrealized appreciation, not committed capital, which is a materially smaller number; and Swish holds Wonderful, Upwind and AIR, so a new fund was announced the same week the existing book re-rated by billions. Nothing improper — that’s venture timing — but it’s context for the “eight of twenty portfolio companies are unicorns” line. It’s also the first significant Israeli fund close since Team8’s $365 million, a gap issue #6 flagged specifically.
Six Billion Shekels, Still Owed
The most important defense item this week was an interview nobody planned. IAI chairman Boaz Levy went on 103FM on Thursday and said the Defense Ministry’s debt to the company “currently stands at close to NIS 6 billion.” He said maintaining wartime production gets harder daily, that IAI is absorbing the burden itself, and called on the Finance Ministry to release budget.
Follow that back. Issue #6 reported Israel adding roughly NIS 40 billion to the 2026 defense budget precisely because production lines were weeks from shutting down, with NIS 15.5 billion owed across the three primes — NIS 5.5 billion of it IAI’s. Three weeks later IAI’s number has gone up. Whatever that NIS 40 billion is doing, it hasn’t reached the balance sheets building the munitions. On the listing, Levy said the IPO process “could still take time”; CTech separately reported that a securities-disclosure problem specific to defense companies could push IAI out of the reporting regime rather than into a listing. With an October 27th election, 2027 is the honest base case.
Pointing the other way: Elbit signed roughly $270 million in ISR and targeting contracts Tuesday with an undisclosed international customer over up to six years. Against a $32 billion backlog it’s incremental, but it’s the right kind: issue #5 noted 73% of that backlog is now non-Israeli and argued the cycle Israeli startups are riding is an export cycle, not a wartime-procurement one. A foreign customer signing a six-year ISR deal is that thesis holding.
And Anduril is putting a permanent operation in Israel, appointing a former air force commander — rendered Amiram Norkin in the Israeli reporting — to run it, moving from market review to R&D and possibly manufacturing. Connect that to issue #2, where ASIO Technologies disclosed a tens-of-millions component deal with Anduril. That looked opportunistic then. With a country lead, it looks like the first of a program.
If you’re modeling defense demand: model the export line and the Ministry of Defense line separately. They are not behaving the same way.
Everything Else Worth Your Time
No standalone Israeli tech layoff, shutdown or insolvency was announced this week. The last time this digest could say that was issue #4, and a week later there were two. Noted, not celebrated — a $1.4 billion news cycle crowds bad news out of the feed, not out of the market.
August closed at $577.5 million across 12 disclosed rounds per CTech, near $600 million per Globes — the strongest August since 2022, which is a strange sentence after three issues of drought coverage, and a reminder that “worst week on record” and “above-average month” can both be true. Year to date: $9.6 billion. IVC and LeumiTech also count 575 active multinational R&D centers in Israel, 15 opened against 8 closed in H1. Given how much of this digest covers companies leaving, that deserves a line.
Anthropic pushed back its IPO to a late-September prospectus at the earliest. Per the editorial decision in issue #7, Decart’s pending acquisition isn’t revisited until there’s a signature — but the deal was reportedly timed ahead of that listing, and the timetable moved.
UVision’s Nasdaq IPO had no new milestone. Third consecutive issue. Noting it so it doesn’t quietly disappear.
What I’d Do With This Week
If you’ve been waiting out the summer: stop. But be honest about which line you’re in. The four mega-rounds went to obvious category leaders, and in three of four an existing investor led again. The other ten averaged $20 million.
If you’re raising a seed in AI security: the number is now $50 million within six months of incorporation, pre-revenue, from Sequoia and Greenoaks. Four years ago $51 million was Minimus’s war chest and it wasn’t enough to finish. Same dollar figure, opposite meanings, and the difference is entirely which side of the category-leader line you’re on.
If you’re a deep-tech company with a ten-year horizon: LightSolver went public by reverse-merging into a Nasdaq biotech. XTEND went public by merging into a Florida construction company. Neither is elegant and both carry real costs, but for categories where fund lives are shorter than the technology’s timeline, “find a listed vehicle” now belongs on the same page as “raise a Series C.”
If you’re a defense-tech founder: Elbit signed a six-year export contract this week; IAI disclosed its domestic customer is NIS 6 billion behind after an emergency NIS 40 billion top-up. Two different businesses, two different payment behaviors, and only one is reliable right now.
If you’re deciding where to incorporate: take Zeev’s advice, probably. Just take it with open eyes about the aggregate — 70% of Israeli-founded unicorns built in the US, 57.9% of first-half capital to foreign-HQ entities, and all four of this week’s biggest winners registered somewhere else. Nobody is doing anything wrong, which is what makes it structural rather than scandalous.
If you’re watching the ecosystem number: $1.4 billion, fourteen rounds, 84% to four companies. Best funding week this digest has recorded and its most concentrated one — a sentence I have now written in some form five issues running. The drought broke. The concentration didn’t.
See you next week. Same lazy hands, same working head.
With 💜, Alex Sherman
Research assisted by AI, every claim read and filtered by a human. Primary sources: CTech / Calcalist — including its running funding, M&A and layoffs trackers — Globes, the Jerusalem Post, Bloomberg, TechCrunch, Business Wire, PR Newswire, GlobeNewswire, SEC and TASE filings, VC Cafe’s Firgun newsletter and IsraelVC. The $1.4B / 14-round weekly total is our own tally and matches Firgun’s and IsraelVC’s independently; CTech’s $1.164 billion figure covers only the nine rounds announced September 1–2 and excludes Easy Aerial (Aug 31), Lyte, Flamingo, Newlight and Guardio. Where sources disagree: Wonderful’s valuation is $5B per the company, TechCrunch, CTech, Bloomberg and Dealroom, and $5.5B per Globes; its March Series B is confirmed at $150M by CTech’s own March reporting, correcting the $165M figure carried in issue #5. HiBob is $3.2B per Bloomberg, CTech’s article on the deal and Dealroom — CTech’s separate weekly roundup says $3.5B, and we use the former. Swish Ventures’ portfolio is reported as 20 companies / 8 unicorns (Globes, JPost, Forbes) and 21 / 9 (IsraelVC). The Anduril appointment is rendered “Amiram Norkin” in Israeli reporting; the former air force chief is usually written Amikam Norkin, and sources differ. Estimates, not disclosed figures: Sweep’s price (Calcalist, “hundreds of millions”), Upwind’s ARR ($50–60M, press-estimated), LightSolver’s consideration (undisclosed, equity-based), XTEND’s final private investment ($100M against a $152M plan). Company-provided and unaudited: Wonderful’s ~$70M run-rate, Guardio’s $150M ARR and 1M+ customers, HiBob’s $400M+ sales, AITAN’s 500,000 flight hours, Newlight’s fuel and emissions figures, Particula’s unit sales, AIR’s add-on research. A reported $170M secondary alongside Wonderful’s round could not be verified against any primary source and is not included. Easy Aerial’s release names Insight and Entrée as participants without naming a lead, and we describe it that way. Particula’s Warner Bros. Harry Potter license predates this round. Not confirmed: XTEND’s Gauntlet II result, which has not been published. Per the editorial decision in issue #7, Decart’s pending acquisition is not revisited and returns only on a signed agreement; the Anthropic IPO delay is included because it is independently reported and dated within this window. Previous issues: #1, #2, #3, #4, #5, #6, #7.