- STRL’s integrated site-development and electrical capabilities benefit as AI campuses grow more complex; power constraints expand its civil and electrical scope.
- Q3 revenue reached $1.56B (+127% YoY, roughly 70% organic), while adjusted EBITDA rose 117% to $337.8M; E-Infrastructure generated $1.30B, up 212%.
- Signed backlog reached $4.14B (+61%), combined backlog $6.14B (+78%), and scoped work approached $8B; 2026 guidance implies roughly 88% revenue and 101% EPS growth at midpoint.
- The author is watching a $300–$340 shelf and sees near “load-the-boat” territory, but 0.9x signed book-to-burn makes award timing and project phasing key questions.
Investment pitches
filter:
71 69 - LGN has roughly $5.7B of backlog against a $3–4B market cap, or about 1.4x+ backlog-to-market-cap, while trading near 1x sales versus 3.3–5x for cited peers.
- About 60% of backlog is tied to data centers and technology; 1.2x recent-quarter and 1.4x trailing book-to-bill indicate awards are outpacing recognized revenue.
- The bullish relative-value thesis depends on continued AI data-center spending, but acquisition integration, GAAP profitability, and execution remain risks; the author is still watching the numbers.
77 - DOCN is framed as an overlooked AI-cloud alternative: 11–12x trailing sales versus NBIS’s 31–33x, despite GAAP profits, ~40% adjusted EBITDA margins, and positive adjusted free cash flow.
- AI customer ARR reached $234M in Q2 2026, up 212% YoY; inference grew ~800% and supplies over 70% of AI ARR, with ~85% from inference and core cloud rather than bare-metal GPU rentals.
- NBIS has larger backlog and stronger GPU pricing power, but DOCN offers more customer diversification and a less capital-intensive path; the thesis depends on software attach and committed capacity converting into durable growth.
63 - FN is a picks-and-shovels play on AI networking: USFP targets 6.4 Tb/s per direction while extending pluggable optics rather than immediately replacing them with co-packaged optics.
- FY2026 revenue grew 36% to $4.64B; Q4 data-center revenue reached $669M, up 68% and 51% of sales, with FY2027 guidance pointing to continued program ramps.
- CSCO and NVDA represented roughly 36% of FY2026 revenue, providing exposure to major AI-networking customers but increasing concentration risk.
- The author remains bullish after the fall from nearly $749 to the low-$400s, while noting USFP is not an immediate catalyst and the chart, execution, competition, and emerging architectures remain risks.
82 - WDC is now a pure-play HDD business after the SNDK spin-off; author sees AI training, inference and agents creating durable bulk-storage demand that expensive flash cannot economically absorb.
- FY2026 revenue rose 36% to ~$12.9B, Q4 non-GAAP gross margin reached 54.4%, and FCF was ~$3.5B; Q1 FY2027 guidance supports the cash-generation thesis.
- Areal-density gains and UltraSMR let WDC grow exabytes without proportional drive output; management targets >25% annual growth, with 40TB ePMR ramping and 44TB HAMR targeted for H1 2027.
- Author bought more after the pullback, but flags customer concentration, AI-capex digestion, Toshiba oversupply, HAMR execution and cyclical margin normalization; October 22 earnings is the key near-term test.
74 - AXON has fallen from $800+ to about $423 even as Q2 2026 revenue rose 35% and ARR 39%; future contracted bookings reached $15.1B.
- That pipeline provides multiyear visibility, with management expecting 20–25% conversion within 12 months; 2026 guidance is 32–34% growth at roughly 25.5% adjusted EBITDA margins.
- AI Era products grew nearly 700% year over year and Platform Solutions 123%; the author sees software, drones and emergency systems expanding a roughly $159B estimated TAM.
- At about $423, AXON trades near 55x estimated 2026 non-GAAP EPS; upside depends on margin expansion and cash conversion, so the author wants a durable base before buying and cites $300s downside if support fails.
70 - COHR generated $7.12B of FY26 revenue, up 23%, while datacenter and communications grew 40%; management says FY27 is essentially booked, with visibility into 2027.
- The thesis is that 800G/1.6T hyperscaler deployments make InP lasers, transceivers, and photonics a bottleneck; COHR’s vertically integrated InP and optical-module capabilities could convert demand into growth.
- Capacity expansion, rather than orders, is the constraint; management’s path toward $3B in quarterly revenue by late FY27 is the growth case, dependent on InP yields, capacity, margins, and hyperscaler spending.
- The author would accumulate weakness rather than chase: watching 50-day support, ~$290 at the 200-day, and a $330–$360 reclaim; valuation matters, and a breakdown could send COHR lower.