Tesla Q3 Preview: FSD Could Be The Key Driver Of Profitability (Downgrade)

The analyst expects Tesla’s FSD subscriptions to grow 56% YoY in Q2, reaching $791 million in ARR, and forecasts up to 750,000 subscribers and $860–$890 million in ARR for Q3. Despite the projected subscription growth and high-margin business model, the analyst reduced the rating from Strong Buy to Buy due to higher CAPEX guidance and projected negative free cash flow through 2026.

Analysis

The key underwriting question is cash conversion, not the headline subscription run-rate. Annualized subscription metrics can overstate durable economics if they omit churn, promotional pricing, regional availability, or conversion from one-time FSD purchases; verify paid net adds and recognized software revenue before assigning a software-like multiple. Rising compute, training, and deployment spend could absorb incremental gross profit, leaving equity sensitive to funding the autonomy roadmap rather than near-term subscription growth.

Over the next 1–3 months, reported subscriber growth and CAPEX/FCF guidance are the likely repricing catalysts. A miss on paid net adds or a further cash-spend increase could outweigh a record ARR figure. Over 6–18 months, the upside case requires demonstrated retention and broader regulatory permission to monetize autonomy; until then, “physical AI” is an option, not an established earnings stream. Autonomous-driving competitors may pressure pricing or accelerate expectations, but no competitor’s technical progress alone establishes Tesla’s monetization.

Contrarian angle: the market may over-credit ARR as recurring software while underweighting the possibility that subscription penetration improves economics of the installed fleet over time. Conversely, treating FSD subscriptions as proof of scalable robotaxi economics is premature. With mixed signal and no valuation or cash-flow detail supplied, the evidence supports event-driven caution, not a high-conviction directional bet.

Key Decisions for Investors

• Do not add TSLA solely on the subscription run-rate. Before upgrading conviction, verify paid net additions, churn/retention, recognized software revenue, and whether growth is incremental rather than shifting customers from one-time purchases.
• For the next quarterly update, track the combination of subscription monetization and cash burn: stronger paid growth alongside stable or improving FCF would support the software thesis; rising CAPEX without monetization progress would falsify it and increase downside sensitivity.
• Keep TSLA exposure sized for automotive and execution risk rather than valuing autonomy as a proven software business. With no supplied valuation or position context, there is no basis for a numeric target or a forced long/short trade.
• Reassess over 6–18 months against regulatory expansion, retention, and evidence of profitable autonomous deployment. If those catalysts fail to materialize, treat the AI premium as vulnerable; if they do, the recurring-revenue model may be underappreciated.