CSAI — Q2 2026 Financial Report Analysis

Cloudastructure, a Palo Alto company that sells cloud-based video surveillance plus AI-assisted "remote guarding" (trained staff watching camera alerts and talking down intruders over a speaker) mainly to apartment-building operators, grew second-quarter 2026 revenue 13.5% to $1.23 million. The mix changed far more than the total: recurring subscription revenue from its two core services almost tripled, while one-time hardware and installation sales fell. The net loss narrowed to $1.73 million from $2.19 million, but that improvement came from a non-cash accounting gain; the loss from actually running the business widened, and with $3.81 million of cash against roughly $0.74 million a month of operating cash burn, the 10-Q carries a going-concern warning.

At a glance

• $764,000 of subscription revenue, up 164% (cloud video $411,000 + remote guarding $353,000). It is now 62% of revenue versus 27% a year ago, and subscribing properties rose 140%.
• Gross margin 49.5%, up from 36.8% — gross profit (revenue minus the direct cost of delivering it) rose 53% to $610,000 as low-margin hardware sales shrank.
• $3.81 million of cash at June 30, down from $8.45 million at December 31, after $4.46 million of operating cash outflow in six months. Management says these conditions raise "substantial doubt" about the company's ability to continue as a going concern.

The numbers

The Q2 2025 column is "as revised": the company found errors in how it accounted for its Series 1 and Series 2 convertible preferred stock (see below) and restated the prior-year comparison. Per-share figures reflect the 1-for-30 reverse stock split of July 31, 2026.

Where the growth came from

Revenue comes from four lines, and they moved in opposite directions:

The subscription lines are billed per camera view and mostly paid annually in advance, so they recur; hardware and installation are one-off sales tied to new sites. The filing attributes the shift to a 140% increase in subscribing properties. Because hardware carries thin margins, its decline cut cost of goods sold by 9% even as revenue rose: hardware costs fell $177,000 (69%), while hosting/data-center costs rose 91% and remote-guarding costs 96% as the service base grew. That mix shift is the whole story behind the 12.7-point gross-margin gain.

Over the first half, revenue rose 40% to $2.55 million and gross profit 81% to $1.32 million, so the second quarter was slower than the first: H1 revenue implies a Q1 of about $1.32 million, so Q2 revenue fell roughly 6% sequentially: hardware dropped from about $382,000 in Q1 to $233,000, while subscription revenue rose about 20% quarter on quarter (from roughly $638,000 to $764,000).

Spending still outruns the business

Operating expenses rose 16.7% to $2.71 million, faster than revenue. Sales and marketing rose 52% ($321,000), mainly from salaries and benefits (+$194,000), and the "operations" department more than doubled (+$188,000, 110%) on added headcount. General and administrative fell 16% on lower salaries and professional fees, and R&D slipped 6%. The result: the operating loss widened 9.3% to $2.10 million. Operating margin — operating loss as a share of revenue — was -170.5%, meaning the company spent about $2.70 in costs for every $1 of revenue.

What the headline numbers hide

• The narrower net loss is an accounting swing, not an operating improvement. Q2 2026 includes a $319,000 non-cash gain from remeasuring derivative liabilities tied to the preferred stock; Q2 2025 had a $334,000 non-cash loss on the same line. Strip that line out of both years and the loss before it widened from about $1.85 million to $2.05 million (+11%), in line with the wider operating loss.
• The per-share improvement is mostly dilution and smaller preferred charges. The loss per share fell 58%, but the weighted share count rose 40% (570,563 to 797,585 post-split) and the "deemed dividend" charged to common holders for the preferred stock's original issue discount fell to $122,000 from $961,000. Spreading a similar loss over more shares is not the same as losing less money.
• Prior-period figures were restated. While preparing its Q1 2026 statements, the company found it had not separated out the variable-price conversion feature in its preferred stock as a derivative liability, and had classified the preferred as permanent equity instead of "temporary equity." Management deemed both errors immaterial and revised the comparatives; they do not affect revenue, operating loss or cash. The Q1 10-Q was not filed until July 17, 2026. On June 29 the Series 2 terms were amended to remove the features behind both errors.
• Cash burn roughly equals the loss. Six-month operating cash outflow was $4.46 million against a $4.51 million net loss; $791,000 of stock-based compensation added back was offset by working-capital outflows ($335,000 into other current assets, $294,000 paid down on payables). Receivables rose 16% since December to $1.01 million and deferred revenue (cash collected in advance for service not yet delivered) 14% to $674,000; both are small next to the cash burn.
• Debt replaced part of the preferred stock. On June 30, holder Streeterville exchanged 1,170 Series 2 preferred shares for a $1.30 million unsecured note at 9.5% interest, due July 30, 2027. Streeterville may convert up to $108,332.50 a month into stock; the first such exchange, on August 6, issued 22,297 shares at about $4.86 each. A Nasdaq delisting would be an event of default that could accelerate the note and raise the balance by 10%.
• No guidance. The filing gives no revenue or profit outlook, so there is no prior guidance to compare against.

Takeaway: The business is turning into what investors want — recurring subscription revenue up 164% and gross margin near 50% — but at about $1.2 million a quarter that revenue is a fraction of $2.7 million in quarterly operating costs, and $3.81 million of cash covers only around five months of the first half's burn rate without new money.

Funding and listing: the real near-term question

Management's plan to fund operations is to sell stock through a $50 million equity line with Atlas (no shares sold under it so far, other than commitment shares) and a $9 million at-the-market program with Maxim. Both require the stock to stay listed on Nasdaq. The company received a minimum-bid-price deficiency notice in February 2026, did a 1-for-30 reverse split on July 31, and regained compliance on August 14. It also flags that its market value of listed securities was about $5.7 million at June 30, close to a proposed $5 million Nasdaq minimum that is currently stayed. The filing itself calls this "a circular dependency": raising money requires the listing, and keeping the listing may require the money.

What to watch

The Q3 2026 10-Q, due around mid-November, will show whether subscription revenue keeps compounding at triple-digit rates and whether hardware stabilises, so that total revenue resumes growth after the sequential dip. The bigger tests are financial: whether the company draws on the equity line or ATM (and at what dilution), how much of the Streeterville note converts into shares each month, and whether the board implements the stockholder-approved repricing of about 515,000 stock options, which would add non-cash compensation expense. With cash falling about $2.3 million per quarter, a capital raise before year-end looks likely on the current trajectory, and its terms will matter more to existing shareholders than the quarter's revenue growth.