Why a chipmaker's operating profit swelled 375.0% in a year, and what its design costs suggest

Sales up by more than half, profit up several times over

Most people have never heard of Lattice Semiconductor, yet its chips sit inside things they use: servers in data centres, factory machines, cars and network gear. The American company designs small, low-power programmable chips, known as FPGAs, that engineers can configure for a specific job after the chip leaves the factory. Think of them as the adaptable glue between a device's bigger, more famous processors.

In the quarter that ended in early July 2026, Lattice took USD 201m in revenue. In the same quarter a year earlier, ending in late June 2025, it took USD 124m. That is a rise of 62.2%, or USD 77.1m more in a single quarter. The figures are consolidated, in US dollars, under US GAAP.

Operating profit, what the business earns from making and selling chips before interest and tax, moved much further. It went from USD 4.7m to USD 22.4m, an increase of 375.0%. The operating margin, the share of each dollar of revenue left as operating profit, widened from 3.8% to 11.1%, a gain of 7.3 pp.

Net profit followed the same path, from USD 2.9m to USD 19.4m. A year earlier Lattice kept barely 2.3% of revenue as net profit. Now it keeps 9.6%.

So the quarter was clearly better. The more interesting question is why a strong but not extraordinary rise in sales did so much more for profit, and what that says about the quarters when sales go the other way.

A chip costs a fortune to design and little to copy

The answer lies in how a chip company spends its money. Lattice designs its chips and has them manufactured by contract foundries. The expensive part is not the silicon in each unit. It is everything that happens before the first unit ships: years of engineering salaries, design software, testing, the software tools customers need to program the chips, and the sales engineers who help a carmaker or a server builder put the chip into a new product.

Nearly all of that is fixed in the short run. The engineers are paid whether the company ships many chips in a quarter or far fewer. Once revenue covers that base, each extra chip sold carries only the cost of making it, and most of the rest falls through to operating profit. In a lean quarter the logic flips: the engineering bill stays where it is, and the shortfall in sales comes almost straight out of profit.

That is exactly the pattern in these quarters. A year earlier, with revenue of USD 124m, the fixed base swallowed nearly everything and left an operating margin of 3.8%. With USD 77.1m more revenue on a cost base of the same kind, the margin climbed to 11.1%. The business did not change its nature. More sales simply spread the same fixed costs thinner.

The break-even check is simple to run on Lattice: operating profit is now 11.1% of revenue, so if every cost were fixed, a fall in revenue of about 11.1% would take it to zero. Some costs, such as the foundry bill for each chip, do shrink with sales, so the real cushion is somewhat wider than that. The year-earlier quarter shows how thin it gets: with USD 77.1m less revenue, the margin was just 3.8%. A drop in sales the size of the one Lattice has just recovered would take the bulk of its operating profit with it.

The balance sheet grew alongside the profit

The extra profit shows up in the bank too. Cash rose from USD 107m to USD 173m, an increase of USD 66.1m, or 61.7%. Equity, what would be left for shareholders if every asset were sold and every debt paid, rose by USD 100m to USD 788m.

Lattice remains a company funded mostly by its owners rather than lenders. Equity covers 80.7% of total assets of USD 976m. That share slipped by 4.3 pp from 85.0% a year earlier, because total assets grew faster than equity. With a cushion that thick, I would not read much into the dip. A company this lightly indebted can ride out a weak quarter without anyone at the bank getting nervous, which matters for a business whose profit moves this sharply.

What a lever this long means for customers and for the rest of us

For anyone watching the semiconductor industry, the pattern is a reminder that a chip company's profit is not a smooth line. A year of modest sales can look like a crisis on the profit line, and a year of recovery can look like a transformation. Both are often the same cost base meeting different volumes.

For anyone running a business with a heavy fixed base, whether it is a software house, a design studio or a workshop full of skilled staff, the check is the same one that explains Lattice. Split your costs into those that rise with sales and those that do not, then work out how far revenue could fall before operating profit reaches zero. If the answer is uncomfortably small, your best quarter is also showing you the size of your worst one.

Lattice had a genuinely better quarter, and it earned it the way chip designers usually do: more units through the same engineering bill. It did not become a calmer business. A profit that grows 6.03× as fast as sales can shrink just as quickly, so I would watch revenue first and the margin second, because the second will simply follow the first.