Backlog isn't revenue — reading SB Energy's $439B number
SB Energy's S-1 lists $439B in contracted backlog, but only ~$1B is expected to become revenue in 24 months. Here's how backlog actually converts.
A headline number like "$439 billion" is built to be repeated, and in the days after SB Energy filed its S-1 for a Nasdaq listing under ticker SBE, it was — attached to nearly every writeup of the filing. What gets lost in the repetition is that $439 billion is not a revenue figure, a valuation, or anything resembling cash the company has earned or will soon collect. It's contracted backlog, a specific accounting term with a specific and much smaller near-term payoff, and the gap between the headline and the mechanics underneath it is worth understanding in detail — because backlog shows up in plenty of IPO filings beyond this one.
What "backlog" actually means in an S-1
Backlog, as SB Energy's filing defines it, is customer commitments made under binding contracts that haven't yet been recognized as revenue under GAAP. That includes two different things bundled into one number: revenue that's been billed but deferred, and revenue that's been contracted but not yet billed at all. Critically, the filing itself states that backlog is neither recognized revenue nor evidence that the underlying facilities have actually been built. A contract can be binding, real, and worth billions, while the power plant or data-center campus it depends on is still an empty lot.
That distinction matters more for SB Energy than for most companies, because none of its data centers are operational yet, and the company has generated zero revenue from that side of the business so far. The backlog is a claim about future capacity under contract — not a claim about anything currently generating cash.
The $439 billion number, and the $1 billion behind it
Of SB Energy's total backlog, roughly $430 billion comes specifically from data-center contracts — the AI-power-demand story that's driving investor interest in the deal. But the filing also discloses how that backlog is distributed across time, and the distribution is the part that didn't make most headlines: only about $1 billion of it is expected to convert into revenue over the next 24 months. Roughly $357 billion of the total sits beyond year eight. The data-center contracts carry a weighted average remaining term of 19.6 years.
Put those numbers side by side and the picture changes completely. A $439 billion backlog sounds like a company sitting on a near-term revenue avalanche. A $1 billion near-term conversion against a 19.6-year average contract term describes a company whose economics depend on decades of future execution — construction schedules, power-grid interconnections, and customer commitments all holding up over a time horizon most public-market investors don't model with any precision. Both descriptions are accurate. Only one of them is what the $439 billion headline implies.
Nvidia's involvement follows a similar pattern of a large number requiring a second look. Nvidia has committed $3 billion around the IPO — split between a private placement at the IPO price and a prepaid forward contract — plus guarantees that support the first phase of SB Energy's data-center buildout, capped at $105 billion. That's real, disclosed, binding capital, and it's meaningfully different from a $3 billion cash purchase order. Guarantees back specific construction obligations; they aren't revenue either, and they don't convert into revenue on any fixed schedule.
Why the conversion timeline is the actual risk
The reason backlog-heavy filings deserve extra scrutiny isn't that the contracts are fake — it's that a contract's face value and its cash-flow timing are two separate questions, and only the second one tells you what the business is worth today. A data-center lease signed now but not billable until a facility finishes construction three years out carries construction risk, financing risk, and interconnection risk between signing and the first invoice. None of that risk shows up in the backlog figure itself; it only shows up once you ask what fraction of the backlog converts on a horizon short enough to matter to a current shareholder.
This is the same discipline that matters when evaluating free cash flow yield on already-public companies — a metric is only useful once you know exactly what it's measuring and what it's silently excluding. Contracted backlog measures commitment, not cash collected; a $439 billion backlog and a 2% two-year conversion rate are two different facts, and a filing that leads with the first number is not lying by including it, but a reader who stops there is missing the one that actually bears on near-term financial performance.
Reading backlog in any filing, not just this one
SB Energy won't be the last capital-intensive infrastructure company to file with backlog measured in the hundreds of billions — power, semiconductor capacity, and long-duration infrastructure deals all tend to disclose the metric, because a large backlog is a more flattering number to lead with than early-stage revenue. The questions worth asking are the same every time: what fraction of the backlog is expected to convert to revenue in the next 12 to 24 months, what's the weighted average remaining contract term, and are the contracts contingent on construction milestones the company hasn't hit yet. A filing that discloses those breakdowns, the way SB Energy's does, is giving you the tool to correct for the headline. A filing that discloses only the aggregate number is asking you to take the headline on faith.
SB Energy hasn't priced yet — no share count or price range is set — and it's currently listed among the still-unpriced names on Zorro's IPO pipeline tracker, alongside the rest of the AI-power-infrastructure cohort filing this quarter. Once it prices, it moves to the recent-IPO tracker, where the numbers that matter shift again: not backlog, not guarantees, but actual quarterly revenue against whatever growth rate the roadshow priced in. That's the number the $439 billion headline was never going to be a substitute for.