Fusemachines Inc.FUSE

Sameer Maskey’s Fusemachines Inc. filed another shelf-registration S-8 on 9 July 2026 to issue up to 5 million additional ordinary shares, this time “subject to stockholder approval,” a phrase that in plain English means the company needs fresh equity to stay solvent. The timing is unlucky: the latest 10-Q for the quarter ending 31 March 2026 shows $1.9 million in revenue against a $928,000 net loss for the year to 31 December 2025, while total liabilities have ballooned to $20 million against just $4.9 million in assets. Shareholders’ equity sits at a neat minus-$15 million, the kind of balance-sheet hole that usually precedes a rights issue or a fire sale of the furniture.

Fusemachines sells itself as an enterprise AI studio with product pages promising sub-50-unit forecast errors for retailers, 90%+ CSAT for media clients, and AI-driven compliance for healthcare. The marketing site is slick; the 10-Q is brutal. Revenue of $1.9 million is spread across retail, media and entertainment, and healthcare, but the operating loss is $3.0 million, and operating cash flow is negative $2.2 million, meaning every dollar of revenue is being soaked up by losses and burned by the business. Research and development cost only $298,000 in the quarter, so the company is not even investing at scale-it is merely treading water while regulators count the paper.

The board’s latest 8-K on 24 July 2026 is the fourth regulatory filing in three weeks, an administrative blizzard that underlines the gap between Maskey’s public platform story and the company’s thin financial reality. The 2 July ownership 4/A confirms insiders already hold a large stake, so any new shares will almost certainly dilute existing holders rather than fund buybacks. Diluted EPS for the March quarter is minus 3 cents on 28.9 million diluted shares, and the company burned $2.2 million in operating cash during the same three months. Capex was a cosmetic $20,000, signalling that the firm is not scaling hardware or data centres-it is scaling filing trail. The S-8 shelf itself is a placeholder; approval is not guaranteed, but the optics are clear: the company is preparing to issue, not repurchase.

The company’s product pages trumpet “AI-enabled healthcare delivery” and “AI studio accelerating development,” but the filing rail contains no audited customer contracts, no named Fortune-500 logos, and no revenue recognisable in GAAP form beyond the $1.9 million headline. The healthcare page claims a “data pipeline that efficiently handles healthcare price information,” yet the 10-K for 2025 shows a net loss of $928,000 and no segment-level breakouts that would let investors trace which vertical is carrying the load. The media-and-entertainment page boasts 90%+ CSAT, but the filing evidence contains no third-party verification or customer names. Retail claims “under 50 units error in weekly sales,” yet the filings do not disclose the size of the client roster or the revenue per client. In short, the product pages are marketing, while the filings are the only evidence that counts.

For readers who like to trace the company’s claims back to the source rail, the SEC submissions JSON for CIK 2033383 lists the S-8, 8-Ks, and ownership 4/A filings in chronological order, and the company facts JSON confirms that the only audited financials are the 10-Q for 31 March 2026 and the 10-K for 31 December 2025. The EDGAR filing index is the single source-of-truth for every document linked above.

There is The source-change rail is filing/source-led and does not use chart-derived signals. Chart context is available separately on this page through the Atlas chart snapshot or TradingView route where available. The stock is quoted on a thin regional venue; liquidity data and order-book depth are not disclosed in the filings and therefore cannot be asserted. Without a primary-market venue identifier or a depth-of-market feed, any reference to “liquidity” or “bid-ask” would be conjecture. The filings do not include broker quotes, exchange market-data links, or volume statistics, so the only measurable facts are the SEC numbers already cited.

What changed this week: the S-8 shelf registration, another 8-K, and the latest ownership disclosure. What remains unproven: customer concentration, contract sizes, recurring revenue, and any pathway to positive operating cash flow. What would deepen the file: a segment-level revenue breakout, audited customer contracts, or a credible cash-flow bridge showing how the company converts $1.9 million of revenue into operating cash rather than burning it.