Most important take away

AMD is emerging as a credible AI infrastructure winner thanks to its dominant CPU position (a key differentiator from Nvidia) plus accelerating data center growth, while Arm Holdings, despite a bold pivot into making its own chips, is priced at a stretched valuation (roughly 10x four-year forward sales) that leaves little margin for error. For investors, AMD looks like the more compelling AI semiconductor pick today, while DoorDash continues to execute well operationally despite optical headwinds from the Deliveroo acquisition.

Summary

Actionable insights and investment notes by company:

• Arm Holdings (ARM): The hosts view shares as overvalued at >$250B market cap against management’s own ~$25B revenue / ~$9 adjusted EPS guidance for fiscal 2030/31 (about 10x four-year forward sales, ~23x earnings four years out). The bull case is real: Arm’s CPUs are claimed to be 2x more energy efficient than X86, saving AI data centers ~$10B per gigawatt in capex, and the shift from licensing to making its own chips could deliver 10x the gross profit per chip with $2B in custom-chip demand already booked over two years. Action: Be cautious about chasing on momentum; the valuation has run ahead of the fundamentals, and execution risk on the in-house chip ramp is real (capacity, ASML equipment lead times).

• Advanced Micro Devices (AMD): Stock jumped ~20% on earnings and has tripled over the past year. Data center revenue grew 57%, Q2 guidance implies accelerating growth, and CEO Lisa Su projects “tens of billions” in data center AI revenue next year. The differentiated angle is AMD’s CPU leadership (vs. Nvidia’s GPU dominance), which matters more in the agentic AI era; the client/PC segment is also taking share from Intel. The Helios full-rack AI system ships later this year at ~$3M per unit, with OpenAI and Meta already placing large orders (Meta’s reported as one of the largest AI infrastructure deals ever). Action: AMD remains attractive even after the run, with a clear product roadmap and customer commitments backing the growth thesis.

• Semiconductor equipment (“bottlenecks behind the bottlenecks”): ASML, Lam Research, and KLA are gating factors for any company (Arm, Alphabet, Tesla, etc.) wanting to build new fab capacity. This actually mitigates near-term oversupply risk because new capacity takes years to come online. Action: Keep these names on the watchlist as essential picks-and-shovels exposure to AI capex; also monitor for cyclical risk if demand ever plateaus while supply ramps (a 5-7 year concern).

• DoorDash (DASH): Revenue, GOV, and orders grew ~25-33%, but operating profit and cash flow declined. Most of the cost pressure is a non-recurring D&A jump (>50% YoY, ~$450M of amortized acquired intangibles) tied to the late-2025 Deliveroo acquisition. Stripping that out, opex (S&M +27%, R&D +30%, G&A +30%) is growing slower than revenue, so operating leverage is actually appearing. Positive signals: accelerating DashPass membership growth, all-time-high engagement on non-restaurant categories (groceries, drugstore), and early ROI from the increased tech/marketing spending that originally hit the stock in late 2025. Concerns: trades at ~40x free cash flow, and the long-term TAM is uncertain. Action: On the watchlist; wait for Q2 to confirm the engagement and ROI trends are real before buying.

• Negative shareholder equity (Starbucks, Domino’s, plus mentions of Moody’s/MSCI): Negative equity can come from sustained losses or, more commonly here, from aggressive buybacks and dividends exceeding retained earnings. For Starbucks and Domino’s, it reflects shareholder-friendly capital return rather than poor capital allocation. Action: Don’t reflexively view negative equity as a red flag; check whether it’s driven by losses (bad) or by buybacks/dividends from a profitable business (often fine).

• “SaaS apocalypse” - Wix (WIX): John believes Wix is less exposed to AI disruption than feared because it bundles domain hosting, memory, and infrastructure that AI tools don’t replace, and AI can be additive to its website-builder offering. However, real concerns include free cash flow add-backs (excluding HQ buildout) and a rising share count despite buybacks. Action: Holdable but not a high-conviction buy; consider the capital allocation red flags.

• “SaaS apocalypse” - Salesforce (CRM): Down ~35% from 52-week highs. Mixed picture: top line still growing 10%, Agentforce ARR at $800M (up 170% YoY), and >60% of Agentforce/Data 360 bookings come from existing customers (good expansion signal). But core CRM growth is single-digit and a $25B buyback hints management can’t find better uses for the cash. Action: Likely “relatively unscathed” near-term but legitimate longer-term AI questions; not a clear buy here.

Chapter Summaries

• Arm Holdings earnings reaction: Arm beat earnings but shares retreated on weak mobile commentary. The hosts argue the bigger story is Arm’s pivot to making its own chips (2x more efficient than X86, $2B in custom-chip demand booked) but conclude the valuation is stretched at ~10x four-year forward sales.

• AMD earnings reaction: 20% post-earnings pop, stock has tripled in a year. Data center grew 57%, Q2 guidance accelerates, and CPU leadership differentiates AMD from Nvidia in the agentic AI era. Helios rack systems ship later this year with OpenAI and Meta orders, including one of the largest AI infrastructure deals ever.

• Semiconductor capacity and cyclicality: Discussion of bottlenecks behind bottlenecks (ASML, Lam, KLA). Equipment lead times naturally throttle capacity expansion, reducing near-term oversupply risk, but cyclicality remains a 5-7 year concern, especially if algorithmic efficiency gains (DeepSeek-style) reduce compute demand per token.

• DoorDash earnings: Revenue grew nicely but operating profit fell. Most of the margin compression is the Deliveroo acquisition’s amortization. Underlying operating leverage is appearing, DashPass and non-restaurant engagement are accelerating, and the late-2025 spending bump is showing ROI. Valuation at 40x FCF is the main concern.

• Mailbag - negative shareholder equity: Tyler explains that negative equity at Starbucks and Domino’s stems from buybacks (treasury stock) and dividends exceeding retained earnings, which is generally a sign of strong shareholder returns rather than poor capital allocation.

• Mailbag - SaaS apocalypse (Wix and Salesforce): John defends Wix as less AI-exposed thanks to hosting/infrastructure bundling, though flags FCF and share count concerns. Matt is more cautious on Salesforce, noting strong Agentforce metrics but slowing core CRM growth and a $25B buyback that signals limited reinvestment opportunities.