Augury

Snapshot

Augury is the archetypal Industry 4.0 unicorn whose story has gotten more complicated with age. Founded in 2011 by two Technion graduates, it bolts a wireless vibration-and-temperature sensor onto rotating industrial machinery and runs the stream through a cross-customer AI model trained on more than 100 million machine recordings. CPG and food-and-beverage majors — PepsiCo, Colgate-Palmolive, Heineken, Hershey, DuPont, ICL — buy it to cut unplanned downtime on pumps, motors, compressors and fans. Augury crossed a $1B valuation in October 2021 with a Baker Hughes–led Series E, then raised a flat-to-slightly-up $75M Series F from Lightrock in February 2025 at the same $1B+ level — a round followed five months later by its third layoff in thirty months and, four months after that, a CEO transition. The question the company is now trying to answer is whether an "Industrial AI Workforce" built on top of its sensor data can re-expand multiples before Tractian prices it out from below.

Founding story

Co-founders Saar Yoskovitz and Gal Shaul are the Technion origin story that every Israeli enterprise-AI pitch seems to include. Yoskovitz was an analog-design logic engineer at Intel Haifa from 2006 to 2010. Shaul had been writing production software at Zoran and medical-device startup Endymed. Both had done military service around heavy, failure-prone machines — vehicles, pumps, generators — and both had the engineer's frustration with scheduled maintenance that either came too early or came after a breakdown. They started Augury in 2011 in Haifa (with a New York office later added) around the premise that an engineer's ear — the vibration patterns a seasoned millwright hears when a bearing is going — could be codified and generalized. The founding genius was not the sensor. It was committing to a labeled-data corpus of machine faults large enough to diagnose, not just detect anomalies, which is the capability most competitors still lack.

How it works

A hand-sized wireless sensor magnetically attaches to a motor, pump, compressor or fan. It captures tri-axial vibration, temperature and magnetic-flux data and streams it over the plant's wireless network to Augury's cloud. Models trained across the 100M+ recording corpus (per CB Insights, late 2021) classify the specific fault mode: outer-race bearing wear, inner-race, cavitation, loose impeller, voltage imbalance, insulation degradation, broken rotor bars. That matters because a diagnosis maps to a specific fix a maintenance team can actually schedule. A generic "anomaly detected" alert is noise; "outer-race bearing fault, 6 weeks before failure, confidence 94%" is a work order.

The catalog has layered up in product lines rather than revolutions: the original handheld Auguscope (US beta, 2014), the always-on wireless Halo platform (mid-2017), Process Health inherited from the May 2022 Seebo acquisition, and the combined Production Health that integrates machine and process signals. In November 2024 Augury announced a "first industrial-grade edge AI-native" sensing platform, and in March 2025 the first PdM product for ultra-low-RPM machinery — important because gearboxes, grinders and conveyors often turn too slowly for classical vibration analytics. The strategic reframing came in May 2026, when Augury unveiled an Industrial AI Workforce of role-based agents (reliability, maintenance, operations) at AVEVA World in Milan, running on top of its own data combined with AVEVA CONNECT's operational context and Google Gemini. ICL Group is the named pilot.

Product and business overview

Four buckets, in rough order of revenue weight: Machine Health (Halo sensors + diagnostics SaaS, the historical cash engine), Process Health (ex-Seebo, applied to batch continuous manufacturing processes), Production Health (bundle), and the nascent Industrial AI Workforce (agents). Augury pitches "Guaranteed Diagnostics" — SLA-backed fault identification — rather than selling raw tools, which gives it a managed-service tint that commoditized sensor vendors cannot match.

Business model and pricing

Subscription SaaS bundled with hardware and expert diagnostic labor. Augury does not publish prices; practitioner directories uniformly list "contact vendor." Glassdoor reviews consistently describe pricing as "high" relative to Tractian and Petasense. The company claims revenue grew roughly 150% in 2021 and "5×" since 2021 in the February 2025 Series F release. GetLatka pegs 2025 revenue around $155M, up ~107% year over year — directionally useful but unaudited.

Traction over time

Headcount crossed 250 at the Series E (October 2021) and reached roughly 400 after the Seebo close (January 2023); Augury reports 357 as of November 2025 per LinkedIn, after three layoff rounds. Named customers now span CPG (PepsiCo, Colgate-Palmolive, Essity, Frito-Lay, Hershey, Lindt, Nestlé, Heineken), pharma (Perrigo, Bausch, Alkermes, PCI Pharma), chemicals (DuPont, ICL, Novozymes), building materials (Holcim, Fortune Brands, Simpson, Louisiana-Pacific), forest products (Roseburg, Canfor, Millar Western), and steel (ArcelorMittal). The valuation trajectory tells a different story: ~$1B at the October 2021 Series E and "$1B+" at the February 2025 Series F — a flat or soft down round in a period the company claims 5× revenue growth, which strongly implies multiple compression on PdM broadly.

Market analysis

Meticulous Research sized predictive maintenance at $79.1B by 2031, 30.9% CAGR, in June 2024. MarketsandMarkets, cited in the Series E coverage, put PdM at $12.3B by 2025 at 25.2% CAGR. The spread is normal for a category whose edges blur into industrial IoT, APM, CMMS and asset-performance consulting. The structural tailwind is tangible: an aging industrial workforce, elevated downtime costs after COVID-era supply shocks, and energy-cost pressure that penalizes poorly-tuned motors. The headwind is that every APM and CMMS vendor is now bolting on AI diagnostics — the differentiation shifts from "we have AI" to the quality of the labeled corpus and the sales motion into the plant.

Competitive intel

Tractian is the direct challenger — a BR/US vertically integrated player that ships its own sensors, an AI stack and a native CMMS workflow, and prices below Augury. Its $120M Series C in December 2024 (General Catalyst-led) and Sacra's growth notes suggest it is taking down Augury deals on price and time-to-value. Petasense, Augury's closest hardware twin, was acquired by Baker Hughes in 2024 — notable because Baker Hughes also led Augury's Series E. Senseye (UK) went to Siemens Digital Industries in June 2022 and is now wedged inside Xcelerator, giving Siemens customers a bundled option. Uptake has shrunk from its 2017 peak. SKF Enlight AI rides the SKF bearings and service network. GE Vernova APM and GE Digital Predix retain big installed bases in power and oil & gas. Falkonry plays software-only on existing sensor data.

History and evolution

Seed in August 2014; Series A 2015; the $17M Series B in June 2017 shipped the Halo sensor platform mid-year. In January 2019 Augury closed a $25M Series C (Insight) and acquired Alluvium, an Austin ML-analytics startup; Qualcomm Ventures AI Fund added $8M that December. A $55M Series D (Qumra) closed in October 2020. The $180M Series E (Baker Hughes) in October 2021 crowned Augury as a first-wave Industry 4.0 unicorn. In May 2022 it acquired Seebo, an Israeli process-AI startup, for an undisclosed sum, adding roughly 85 employees and the Process Health product line; Verdantix flagged integration risk at the time, and Glassdoor reviewers reference persistent friction. Three layoff rounds followed: January 2023 (~5%, ~20 people), July 2023 (a larger cut reviewers still bring up), and July 2025 (18%, ~60 people) — the last coming five months after the February 2025 Series F. On October 22, 2025 Yoskovitz moved to Executive Chairman and Elan Greenberg became CEO, with a refreshed exec slate (Germain CRO, Franco CFO, Brown VP Marketing, Toles VP Partnerships). In May 2026 Augury unveiled the Industrial AI Workforce. A Rockwell Automation agentic-AI partnership followed in July 2026, and the company opened a Bengaluru engineering hub in October 2026.

What people say

The case for. Customers and employees both describe diagnostic accuracy as the moat. The "Guaranteed Diagnostics" SLA — Augury puts money behind correct fault calls — is repeatedly cited in competitive bakeoffs. Glassdoor reviewers rate work/life balance highly (4.2/5) and credit the technology as category-leading. Analysts (Verdantix, Gartner — Augury was named in the August 2026 Market Guide for APM Software) continue to list it among the leaders in machine-health AI.

The complaints. The same Glassdoor corpus is unusually candid about ops dysfunction. Reviewers describe billing lapses lasting more than a year, implementation improvised in the field, seven managers in a few years, and a compensation structure where "automatic raises" quietly never arrive. Three layoff rounds in 30 months — the last directly after a $75M raise — collapsed any "people first" messaging. Customers report sensor install quality issues in the field and ROI claims that outrun deployments. Pricing is called high relative to Tractian and Petasense, with the competitive gap narrowing.

Outlook: the open question

The answer-conditions: Augury's bull case turns on whether the Industrial AI Workforce actually converts Halo installations into plant-wide agentic-ops contracts worth 3–5× current ARR per site. If that happens — if ICL's pilot becomes a lighthouse that a dozen Fortune 500 operators roll out globally — the multiple compression of the Series F reverses and Augury has a legitimate category-defining product. The bear case: the agentic layer is a thin wrapper over Google Gemini plus structured machine data that SAP, Siemens, AVEVA, Rockwell, Honeywell and Emerson will replicate inside stacks the customer already owns, while Tractian keeps winning new logos on price. Flat valuation on 5× claimed revenue growth, a founder-to-operator CEO handoff, and layoffs five months after a fundraise all point to a board that believes operational discipline matters more than product momentum right now.

How to attack it

The most interesting wedge is the CMMS-native predictive-maintenance stack for mid-market plants. Augury's go-to-market is heavy enterprise: six-to-twelve-month procurement cycles, Guaranteed Diagnostics SLAs, custom installations. A new entrant that ships a plug-and-play sensor kit that lands directly inside UpKeep, MaintainX, Limble, Fiix or Infor EAM — with the diagnostic logic running on-sensor edge AI so no cloud subscription is required — could own the 250–500M-revenue plant operators who look at Augury pricing and walk. Tractian has started to show what that looks like; the opening is a US-domestic, OEM-channel version of the same play.

Augury's exposed flanks: (1) Pricing, repeatedly flagged in Glassdoor and competitive reports; (2) Install quality, cited in field reviews, suggesting services-as-product delivery is fragile; (3) Narrow moat on AI diagnosis as every APM vendor adds similar models — the labeled-corpus advantage erodes as competitors harvest customer data at scale; (4) Flat 2025 valuation with ongoing cash burn signals the next round is likely unfriendly without a step-change in product; (5) Agentic-AI exposure — if Google, Microsoft and the hyperscaler agents turn industrial-ops into a thin SI play, Augury's data advantage is a feature inside someone else's stack. Each is addressable by a seed-stage team that picks exactly one.

Adjacent-segment play

The same wireless sensor and diagnostic logic plumbs naturally into commercial and institutional HVAC: chillers, air-handling units, cooling towers, pumps, boilers and compressors are the same rotating machinery, just in office towers, hospitals, data centers and malls instead of plants. Augury has flirted with this edge (Johnson Controls partnership rumors; targeting Carrier, Trane, Daikin OEMs) but has not pushed hard. The buyer is a facilities director or REIT asset manager with budget, not a plant engineer with inertia — and data center operators now have genuine pain around chiller downtime that maps directly to AI power-density economics. Augury-for-data-centers is a credible adjacent-segment company.

A second adjacent: fleet-level health for heavy equipment. Augury's core capability is diagnosis of rotating gear; rail, mining, construction fleets (Caterpillar, Deere, Komatsu, Volvo CE) are full of pumps, motors and transmissions with telematics already instrumented but diagnostic layer under-built. OEMs own that channel but have systematically under-invested in third-party-grade diagnostics. A third adjacent: downmarket per-asset SaaS for the 500k facilities in the US with motors and pumps but no reliability engineer — a self-install kit priced like a Ring doorbell that pays for itself on one avoided breakdown.

Sources and further reading

• Augury Series F press release — augury.com, Feb 19, 2025
• Calcalist on July 2025 Augury layoffs, Jul 10, 2025
• Calcalist on January 2023 Augury layoffs, Jan 4, 2023
• CB Insights Series E coverage, Oct 2021
• Verdantix analysis of the Seebo acquisition, May 2022
• Augury CEO transition press release, Oct 22, 2025
• Tractian's own competitive landscape write-up, 2025
• Meticulous Research predictive-maintenance forecast, Jun 20, 2024
• VentureBeat on Alluvium acquisition + Series C, 2019