YB new stock pitches (Tue, Jul 21)

Upslope Capital Management has taken a new long position in The Magnum Ice Cream Company (MICC), a pure-play global ice cream company spun out of Unilever at the end of 2025 and by far the world's largest ice cream player (~21% share, nearly double #2 Froneri, with all other players holding at most 2%). Magnum owns 4 of the top 5 global brands, including flagship Magnum, Ben & Jerry's, Breyer's, Cornetto, and Wall's, with geographically balanced sales (~40% Americas, ~40% Europe/ANZ, 25% Asia/Middle East/Africa), Emerging Markets contributing ~30% of sales, and only ~25% U.S. concentration, giving it manageable GLP-1 risk. The thesis rests on its dominant, defensive business with competitive advantages from leading brands and a complex global frozen supply chain, top-line growth acceleration and margin gains as a standalone ice cream-focused business, and the Froneri 'comp' providing both margin and valuation upside—Froneri has reportedly expanded EBITDA margins to ~20% versus MICC's ~16%, and a recent transaction valued Froneri at an estimated 10-11x EBITDA. Despite a recent re-rating partly driven by rumors of private equity interest (though a near-term deal is unlikely due to post-spin-off tax hurdles), MICC still trades at a discount at ~9.5x 2026E EBITDA (17x EPS) with net leverage under 2.5x. Key risks include its limited history as a standalone public company, GLP-1-related volume/consumer uncertainty, FX exposure (sales are mostly ex-U.S.), and weather impact on short-term results.

Read the full article here. Read time: 2 min

Diversified Energy Company (DEC), a natural gas and oil producer, trades at $13.35, essentially the same price as early 2024 when Henry Hub gas was below $2/MMBtu (35-40% lower than today's $2.89/MMBtu), despite the company having significantly expanded—doubling its PDP, increasing liquids (oil and NGLs) from 14% to 28% in 1Q26, and completing a series of acquisitions (Sheridan, Canvas, Maverick, Summit, East Texas II, Crescent Pass, and the remaining 50% of Oaktree's holding). The bull case rests on rising gas demand catalysts: the EIA forecasts prices returning to $4+ heading into winter, growing LNG export demand (including from Mexico and Canada), AI-driven power generation demand (US CCGT demand up 1000% between 2023 and 2026, gas turbine order books at four-year waits), and reshoring of urea production (potentially adding 450 BCF/year, or 1.3 bcf/day, growing 5% annually), while oil extraction slows at the Permian and most US onshore fields. DEC currently yields 8.06% annually (~2% quarterly) and has paid $21.84/share in dividends over nine years (36 quarters since June 2017), meaning dividends plus the current share price total $36.19—undermining the argument that the stock hasn't 'paid for itself.' The author notes DEC can also be actively traded (an imaginary $10,000 on Jan 1 could have become $22,345 by July 17 via 16 trades). Additional catalysts include the Camino acquisition (second largest behind Maverick, due to close 3Q26, with strong synergies and off-balance-sheet financing support from Carlyle, which currently prevents buybacks) and the 1H26 interim report due August 11. The long-term thesis includes DEC's low-decline asset optimization model (now being copied), a valuable standalone decommissioning business (2m orphan wells, earning $150k-$200k per well, offsetting its own ARO), potential geothermal upside turning liabilities into assets, and its positioning to steward declining resources. Bear cases/risks include the flat share price frustrating detractors, growing US gas supply, past exposure to a 2022 gas price spike and reversion, prior short-seller attacks (Snowy, the Bloomberg well-leak documentary), and Congressional scrutiny (Rep. Pallone's letters to CEO Rusty). The author remains bullish, viewing DEC as an undervalued high-yield cash monster worth well above $13.35.

Read the full article here. Read time: 5 min

indie Semiconductor ($INDI), trading at ~$3.7-$4.10/share with 227M shares (~$770M market cap, ~$1B EV after ~$230M net debt), is an automotive ADAS chip supplier valued almost entirely on its car business (<4x 2026 consensus sales of $266M, <2x 2028 sales of $521M, versus peer hypergrowth semis at 10-15x forward sales), leaving its humanoid robotics and quantum photonics optionality essentially free. The near-term catalyst is Unitree Robotics' Shanghai STAR Market IPO (targeting ~$618-620M raise, ~$6.18B valuation, listing possibly late July 2026), which should force generalist investors to ask who supplies humanoid sensors—and $INDI has confirmed multi-quarter design-ins at both Unitree and Figure AI (using iND880-powered cameras), and is the only US-listed Western semi with disclosed content at Unitree. The core auto thesis rests on rising content per vehicle (from ~$25 today toward $150+ by 2030, a 6x expansion), 550M+ cumulative SoCs shipped, a $25M Q1 2026 radar order (industry-first 4TX 8RX), a >50M unit/year opportunity, and eMirror/camera mirror wins (NIO, Buick GL8, AITO M9, Cadillac VISTIQ) generating tens of millions annually and possibly exceeding radar revenue in 2026; the key moat is the DRAM-less iND880 vision processor, uniquely valuable amid a 2026 DRAM shortage (contract prices up 50-95% QoQ, AI datacenters consuming ~70% of memory output, automotive-grade DDR3/LPDDR4 nodes being wound down) where management says customers often 'cannot source memory at all,' while the newer iND881 (2 TOPS NPU, targeting humanoids/AMRs directly but requiring external LPDDR4x, forfeiting the moat) could carry more content per unit. A pending €40M AMS OSRAM CMOS image sensor acquisition completes the sensor stack and enables the content-per-vehicle climb, alongside a new per-vehicle software royalty stream (Mahindra in production, near-100% incremental margin) and a nascent photonics/quantum business (~$1M 2025 revenue, tripling in 2026, first UV DFB laser). indie is divesting its stagnant, lower-margin Wuxi subsidiary (~40% of consolidated revenue) for ~$135M gross (~$108M net), which causes a one-time 2026-to-2027 revenue optics dip but improves margins and takes pro forma net debt to ~$122M. The SOTP models Physical AI at $150 content/unit across all cases (base: Unitree 500K + Figure 250K units by 2030), yielding a base-case 12-month price target of $11.02 (~2.7x upside) and bull-case $15.74 (~3.8x upside), both derived from 2027 outputs, with four-year targets reaching $19.28 (base) and $44.28 (bull) versus a $4.44 bear case; management targets 60% long-term gross margin (base assumes 57% by 2030), and the author sits below consensus 2027/2028 sales while still getting ~200% base-case upside. Key risks: slower humanoid adoption keeping Physical AI a rounding error, an ADAS/auto downturn (auto carries most EV in every scenario), a China-exposure valuation discount (NIO/AITO revenue, Wuxi), and ADR-adjacent regulatory/delisting scrutiny; other caveats include above-consensus bull shipment assumptions, unverified per-unit content estimates, the precision-component (harmonic drives/magnets) supply bottleneck, potential faster DRAM normalization, and Wuxi still needing Chinese regulatory approval.

Read the full article here. Read time: 22 min