← Back to list
Colossus (Invest Like the Best / Business Breakdowns)Podcast26 Feb 2025Source: joincolossus.comHost: Colossus

Cognex: Vision Quest - [Business Breakdowns, EP.207]

In plain words

This podcast breaks down Cognex, a machine vision company. The guest thinks it's at a turning point: using deep learning to make its products easier to use, and training a new sales team to reach small and mid-sized customers who never used machine vision. The market view is that Cognex is near a 10-year low in valuation, making it attractive long-term but risky if you buy too early. Key holdings discussed: Cognex (own stock, revenue down but logistics up 20%, margins dropped from 30% to 13%); Keyence (Japanese rival, great sales model, margins over 80%); Amazon (once a major customer, now smaller share).

AI SummaryAI-generated · may contain errors · verify against the original

At a Glance

Guest Brett Larson (NZS Capital) provides a deep dive into machine vision leader Cognex on this episode of Business Breakdowns. The core thesis is that Cognex is not a typical recurring-revenue software company, but a cyclical enterprise skilled at capturing "S-curve" growth opportunities by focusing on specific customer segments (e.g., consumer electronics, logistics). Brett Larson argues that Cognex is currently in a critical transition phase: using deep learning technology and an "Emerging Customer Initiative" to expand its business from serving large clients at the top of the pyramid to a broader SME market — which could be the core driver of its next growth cycle.

In-Depth Analysis: Strategic Transformation of the S-Curve Hunter

1. Competitive Positioning: Technology Differentiator at the Top of the Pyramid vs. Process-Driven Sales Everywhere

Brett Larson points out that the machine vision market is dominated by two giants: Cognex and Japan's Keyence, but their competitive strategies are entirely different.

  • Cognex positions itself at the "top of the pyramid", serving the most complex customers and tasks. It has high technical barriers and offers high-value solutions rather than simple price competition. For example, in large logistics centers, Cognex's barcode reading accuracy can be 100-300 basis points higher than competitors, meaning thousands of packages per day require no manual handling. Its sales process is highly technical, requiring application engineers to customize solutions. R&D spending is in the mid-single-digit percentage of revenue, with gross margins around 70%.
  • Keyence focuses on the "mid-to-low end" market, with standardized products and process-driven sales. Its sales force is driven by strict KPIs (e.g., number of visits, number of demos) and is known for "blanket coverage." Keyence's R&D spending is only 2% of revenue, yet it boasts gross margins above 80%, reflecting the efficiency of its business model.
  • Brett Larson judges that Keyence's sales model is a "cultural meme" that is extremely effective, and Cognex is now trying to learn and imitate this strategy through its "Emerging Customer Initiative" to enter Keyence's territory.
2. Historical Pattern: Growth Engine Built on Stacking S-Curves

Cognex's growth is not linear; it continuously captures new waves of large customers and applications, stacking S-curves.

  • Phase 1 (1980s-2000s): Started with barcode reading on semiconductor wafers. By the 2000 internet bubble, semiconductor and electronic capital equipment accounted for as much as 80% of sales.
  • Phase 2 (2000s-2010s): The introduction of smart cameras allowed products to adapt to harsher factory environments, expanding the market into automotive, consumer electronics, food and beverage, etc. By 2010, semiconductor exposure had fallen to 15%.
  • Phase 3 (2010s-2020s): Deep partnership with Amazon to develop barcode reading products for logistics centers. In 2010, the company optimistically expected long-term sales of $75 million for this line; by the 2021 peak, it contributed $300 million, or 30% of total revenue.
  • Current Phase: Brett Larson believes Cognex is in the early stages of the next S-curve: AI machine vision applications based on deep learning and edge learning. This cycle coincides with a prolonged cyclical downturn, so its impact has not yet been fully priced in by the market.
3. Strategic Transformation: From "Programming" to "Teaching," From "Large Customers" to "Mass Customers"

This is the core insight of the podcast. Brett Larson breaks down the two key pillars of Cognex's current strategy in detail:

  • Technology Revolution: From Rules to Teaching. Traditional machine vision relies on rules-based programming, which is ineffective for complex, variable scenarios like detecting subtle scratches or judging chicken cutting positions. Through acquisitions of Viti and SuaLab, Cognex has acquired deep learning capabilities. Its "Edge Learning" products allow users to "teach" the system to perform specific inspection tasks by providing just 5-10 sample images, reducing deployment time from weeks to hours.
  • Business Model Revolution: Emerging Customer Initiative. This is a "disruptive" move by learning from Keyence. Cognex is recruiting and training "salesnoids" rather than traditional technical application engineers. They are given standardized sales processes and carry simple, easy-to-use edge learning products to target SME customers who have never used machine vision before. Brett Larson provides compelling data: In 2024, the first cohort of salesnoids completed 80,000 customer visits, added 3,000 new customers (against a base of 30,000), and exited the year generating $1 million in weekly sales, with accretive margins. 60%-70% of these sales came from customers who had never used machine vision before.
4. Financials and Risks: Valuation in a Cyclical Trough, and Whether the Market Is Pricing in "Permanent Damage"

Cognex's financial model is typically cyclical.

  • Financial Characteristics: All revenue comes from one-time capital equipment sales, with no recurring revenue. The company's long-term target is 15% revenue growth and 40% incremental margins. Operating margins can exceed 30% at cycle peaks, but recently fell to 13% (partly due to investment in the Emerging Customer Initiative).
  • Valuation Status: Brett Larson notes that Cognex's current price-to-sales (P/S) ratio is around 5.5x, near its 10-year low (historically normal range is 6-10x). He estimates this implies the market expects the company to achieve low-double-digit free cash flow compounding growth in the future. He argues that this comes from a cyclical trough, with significant margin recovery potential, making it a "bullish" case.
  • Key Risks:

1. Cyclical Risk: Brett Larson says bluntly, "Being early is the same as being wrong." Investors who bought in the past year or two may have felt "too early"; the key is when the cycle will turn.

2. China Market: Although it accounted for 18% of sales in 2024, two-thirds of that came from Western multinationals (e.g., Apple, Foxconn). For local Chinese manufacturers, facing domestic competitors like Hikrobot, it is a "headwind" over the long term.

3. Technology Transition Risk: While deep learning is an opportunity for Cognex, any technological change opens a window for disruptors. Brett Larson advises investors to monitor this continuously.

Mentioned Positions

Position Guest Stance Key Data
Cognex Bullish on long-term potential, but flags current cyclical trough 2024 revenue declined YoY, but logistics segment grew 20%; operating margin fell from >30% peak to 13%; P/S ratio ~5.5x, near 10-year low.
Keyence Seen as major competitor, risk warning Largest market share; R&D spend only 2% of revenue, gross margin >80%; sales model known for process-driven, high-density coverage.
Hikvision (Hikrobot) Risk warning (China market competition) Third-largest player in China, roughly half Cognex's revenue size, mainly serving local manufacturers.
Amazon Mentioned as major customer, no current position action indicated Historically accounted for up to 17% of Cognex revenue; now in logistics segment, exposure has fallen to high-single-digit to low-teens percentage.
Apple Mentioned as major customer, no current position action indicated Historically accounted for up to 20% of Cognex revenue; now down to mid-to-high single digits.

Key Takeaways to Remember

1. Cognex is a "self-proclaimed cyclical company" (Brett Larson & Matt Russell). It is not a typical software company; its growth is driven by large capital expenditure waves from downstream customers. Understanding its S-curve stacking growth model is more important than focusing on quarterly earnings.

2. The "Edge Learning" product is the key for Cognex to enter the mass SME market (Brett Larson). It shifts from "programming" to "teaching," reducing deployment time from weeks to hours, thereby enabling a non-technical sales team to close deals. This could be the start of its next S-curve.

3. Cognex's Emerging Customer Initiative is "fighting Keyence with Keyence's tactics" (Brett Larson). In its first year (2024), the initiative generated $1 million in incremental weekly sales, added 3,000 new customers, and 60-70% were "virgin" territory — customers who had never used machine vision before.

4. Cognex's "Minister of Culture" system is key to preserving its founder's spirit (Brett Larson). The company has part-time "Ministers of Culture" in offices worldwide, paid separately, to ensure the unique engineering culture endures after founder Dr. Bob's retirement. This explains why its employee turnover rate is half the industry average.

5. The specific risk of "mispricing" is being "too early" (Brett Larson). Because Cognex is highly correlated with capital expenditure cycles, even if the thesis is correct, buying too early can lead to long-term losses. This is the core challenge of investing in this stock.

6. Cognex's valuation is at a historical low, implying market concern about "permanent earnings erosion" (Brett Larson). With a P/S ratio near 5.5x, a 10-year low, the market may be questioning whether margins can recover from the current trough (13%) to the long-term target (30%). But the guest believes recent quarterly data shows that as soon as revenue growth resumes, operating leverage will quickly return.

7. Cognex's "S-curve" history shows its biggest growth often comes from unexpected areas (Brett Larson). From semiconductors to consumer electronics to logistics (partnership with Amazon), its next potential S-curve could come from AR/VR headset manufacturing, humanoid robot production lines, or new consumer electronics forms driven by LLMs.

~10 min full read
Deep Analysis