-$1.58
$0.95
bear -$1.58
base -$1.18
bull $0.95
margin of safety $0.40
market price* $8.98
Business condition attached to the margin of safety
Four consecutive quarters of positive operating cash flow after capital expenditures, capitalized software and equipment-financing repayments, with no new share issuance during those quarters, reported in 10-Q/10-K filings.
QuickLogic is a small fabless chip company in San Jose. It earns money three ways: licensing embedded programmable logic (eFPGA) that customers build into their own custom chips, together with the engineering services to integrate it; selling its own small low-power programmable chips such as EOS S3, ArcticLink III and PolarPro; and a slowly shrinking stream from mature parts such as PASIC 3 made on old manufacturing processes. In 2025 eFPGA licensing and services were $9.5 million of $13.8 million revenue, new hardware about $1.0 million and mature products $3.3 million. A single customer, tied to US government and defense trusted-microelectronics work, supplied 44% of 2025 revenue and 56% of 2024 revenue, so results move with that customer's contract awards. The company also exited its SensiML software subsidiary in 2025, writing off $2.4 million. The cash record. We measure what the business leaves for its owners after the cash it must spend to stay in business: operating cash flow, with working-capital timing swings removed, minus the cost of paying staff in stock, minus the money spent on test equipment, tooling and capitalised software. We count the repayments on equipment bought with vendor financing as capital spending, because that equipment was bought on credit instead of with cash. On that basis QuickLogic has not had a single year of positive owner cash flow in the ten years from 2016 to 2025. The shortfall narrowed from $19.4 million in 2016 to $3.5 million in 2022 and $3.9 million in 2023, the best years. It then widened to $7.6 million in 2024 and $12.0 million in 2025, when revenue fell 30% and gross margin fell from 62% to 22% because a larger engineering and tooling cost base had far less contract revenue running over it. Over the same decade the share count rose from about 4.7 million (adjusted for the 1-for-14 reverse split of December 2019) to more than 18 million. Shareholders have paid for the losses by being diluted, through an at-the-market stock sales program and private placements ($9.3 million raised in 2025 and $9.8 million in the first half of 2026). The first half of 2026 was better: revenue rose 31% to $10.5 million, gross margin recovered to about 40%, and operating cash flow was a positive $2.2 million. Because the main cost here is engineering staff and test tooling, not a factory, we treat 2021 to 2025 as the period that describes today's licensing-led business. The middle year of that window is a shortfall of about $6.8 million, both before and after interest, since interest was under $0.1 million that year. In 2025 the shortfall grew by 58%. For the coming year we assume it shrinks back to $6.8 million, and we have a dated reason: the first-half 2026 results above, plus the end of SensiML's losses. Balance sheet. At June 28, 2026 the company held $18.5 million of cash. It owed $5.0 million on its revolving credit line, which matures December 31, 2026, plus about $2.8 million of equipment financing, so it held net cash of about $10.7 million, roughly $0.59 a share on 18.3 million shares. The 2025 annual report says the company signed a term sheet with Sunflower Bank, N.A. for a smaller $10 million replacement credit line. Its auditors raised no going-concern doubt. Bear case. The shortfall averages the 2024-2025 level of about $9.4 million a year before interest. It lasts five more years, which is how long the record shows it can last before cash and the market's willingness to fund it run out. After that the business has nothing left to sell. Here the business is worth about minus $39.6 million, the net cash adds $10.7 million, and the result is about minus $28.9 million, or roughly minus $1.58 a share. In plain terms, the owners would have to put in about $29 million more than they get back. In practice that means the losses would keep being paid for with new shares, sold at whatever price the market will bear. Base case. The shortfall stays at the recent median of about $6.8 million a year for three years. It then narrows to about $3 million a year through year ten, as the tooling bought in 2023-2024 is used for more contracts and SensiML is gone. Nothing we can name and evidence carries earning power beyond that, so the case ends there. The business is worth about minus $32.2 million; add the $10.7 million of net cash and the result is about minus $21.5 million, or about minus $1.18 a share. The licensing royalties the company has been building toward for years do not appear here. They have never shown up as cash, and we credit only what the record shows. Bull case. The first-half 2026 recovery holds. The business loses about $3 million in the coming year and then settles at a small positive $0.9 million a year for twenty years. That level of profit is evidenced, not hoped for: 2023, the strongest year on record with $21.2 million of revenue, produced a small operating profit. The business is worth about $6.6 million; add the $10.7 million of net cash and the result is about $17.4 million, or roughly $0.95 a share. Even this case contains no royalty windfall and no growth beyond the 2023 level. Accounting judgements. The $2.4 million SensiML write-off and the $5.4 million of 2025 depreciation and amortisation are non-cash charges and are added back. The cash spent on tooling, capitalised software and equipment-financing repayments is charged in full, since revenue fell in 2024 and 2025 and none of it can be called growth spending. Our earlier figures used a different method and are not comparable. What could make this wrong. On the upside, eFPGA licences signed for custom chips could start paying royalties on production volumes. Government trusted-microelectronics programs could also turn into steady multi-year funding. Neither has yet produced lasting cash, which is why neither is credited. On the downside, the business depends on one customer for nearly half its revenue and on raising new equity to fund the losses. Losing either would speed up the bear case. Margin of safety price. We have low confidence in any positive earning power, because there is none in ten years of filings. The worst plausible outcome is the 44% customer cutting back while the credit line shrinks from $15 million to $10 million. The business would survive that only by selling shares at depressed prices. If we are wrong about the 2026 recovery lasting, about the customer concentration, or about the company's access to equity, the shares are worth less than nothing to holders who must fund the losses. Over the holding period the business pays nothing: no dividend, and steady dilution. No margin-of-safety price applies on price alone. If the company shows a full year of self-funded positive cash flow, a price of $0.40 a share would be less than the net cash alone ($0.59 a share), with any earning power carried at no value. That gives a realistic route to a 20% annual return even if the business only holds the small 2023-level profit, and that is why the margin of safety price is $0.40, conditional on that proof.