SpaceX Bought $295M in Tesla Megapacks. That Is Not the Bull Case You Think.

I always keep an eye out for the false narratives that take the market by storm, and this one is dressed up as ecosystem validation. SpaceX spent $295 million on Tesla Megapack batteries in the second quarter - $329 million in the first half - and the market is treating it as proof that Tesla's energy storage business is a structural growth engine with expanding economics. The evidence points in the opposite direction.

Tesla's energy division deployed a near-record 13.5 gigawatt-hours of storage in Q2, its second-best quarter ever. Revenue from the energy generation and storage segment reached $3.14 billion, up 13% year over year. That is genuinely strong volume growth. But energy gross margins were cut nearly in half, collapsing from 39.5% to 20.4%. CFO Vaibhav Taneja attributed the decline to a $240 million warranty true-up on legacy vendor cell issues, a $200 million tariff benefit from Q1 that did not repeat, and declining average selling prices as competition intensifies. He then guided that long-term margins should normalize in the "mid-to-low 20% range." That is the clearest signal yet that the near-40% margins investors became accustomed to were not the new floor - they were a high-water mark that competition has permanently repriced.

The SpaceX Megapack order does not change that trajectory. At $329 million for the first half, it represents about 2.6% of the energy segment's annual run rate. But more importantly, it's a related-party transaction - classified as such in SpaceX's regulatory filings - between two companies sharing the same CEO. When your largest new customer doesn't negotiate like an arm's-length buyer, the question isn't whether the revenue is real. It's whether it's real at a price that sustains the margins the valuation demands.

Step back and look at Tesla's full financial picture from Q2. Revenue was a record $28.24 billion, up 26% year over year, but operating income fell 57% to $398 million, with the operating margin compressed to 1.4%. Capital expenditures more than doubled to $5.79 billion, up 142%, as Tesla poured money into AI compute, a semiconductor fab, battery capacity expansion, and Cybercab production. The result was a free cash flow deficit of $1.09 billion - Tesla's first negative free cash flow quarter since early 2024, compared to a $1.44 billion surplus in Q1. The company crossed $100 billion in trailing-twelve-month revenue for the first time, but the cash conversion engine that justified its valuation is sputtering.

Now look at the valuation. Tesla trades at 334 times trailing earnings and 402 times forward earnings. Its EV/EBITDA multiple - enterprise value divided by earnings before interest, taxes, depreciation, and amortization - sits at 114x. It pays no dividend. Nvidia, the actual AI compute winner, trades at 33 times earnings and 32 times EV/EBITDA. Alphabet trades at 18 times earnings. Amazon at 22. The gap is not a difference of opinion. It's an arithmetic problem. Tesla's valuation is pricing in a future where energy storage and AI infrastructure produce massive, high-margin cash flows that the stock can never afford to wait for if margins are normalizing into the low 20s.

The Musk ecosystem interconnection is worth examining on its own terms. SpaceX, which went public in June and merged xAI into its operations in February, reported $7.81 billion in revenue but posted a net loss of $541 million. Its AI segment generated $2.56 billion in revenue but lost $1.26 billion from operations. Total SpaceX capital spending reached $18.37 billion, including $15.83 billion for AI infrastructure. This is a company spending $295 million per quarter on batteries while burning $1.26 billion per quarter running the data centers those batteries power. The Megapack purchase is infrastructure spending for a cash-negative operation - capital consumption on both sides of the transaction.

The supply chain dynamics underneath this add another layer of structural risk that the market is not pricing in. Tesla's Nevada LFP (lithium iron phosphate) cell line - the one meant to provide domestic, tariff- and tax-credit-compliant cells for Megapack production - remains stuck in an early ramp phase with no updated timeline. Most of Tesla's US-deployed BESS cells still come from third-party Chinese suppliers. Foreign entity of concern restrictions on the US investment tax credit, in effect since the start of the year, make using those imported cells effectively incompatible with claiming the credit on new projects. That's why Tesla signed a $4.3 billion LFP supply agreement with LG Energy Solution in March to secure US-made cells for the Megapack 3 line at its new Texas Megafactory. The domestic supply chain story - the one that would support long-term margin expansion - is years from resolution, not quarters.

The competition is real, too. BYD held 13% of global system integrator market share in 2025, ahead of Tesla's 10%. Wood Mackenzie ranked Tesla outside the top three in Europe, Asia-Pacific, the Middle East, and Latin America. The volume story is intact. The margin story is not. Taneja's own long-term guidance effectively concedes that the near-40% margins were an anomaly, not the trajectory.

I rate Tesla as a Hold. The company has genuine advantages in energy storage deployments, a robust order backlog, and a real role in the AI infrastructure buildout. The SpaceX Megapack orders are evidence of ecosystem cross-pollination, and the energy business is growing faster than automotive. But at 334 times earnings with no dividend, collapsing free cash flow, and an energy margin the CFO has guided into the low 20s, the stock has priced in years of margin expansion, AI dominance, and energy leadership that the current data doesn't support. Investors who own Tesla should not sell in panic. Those looking to add should wait for free cash flow to stabilize and energy margins to demonstrate a floor - because right now, the market is treating a related-party capital expenditure between two cash-burning companies as a bullish fundamental. In my opinion, it isn't.