Since AMD swallowed Xilinx in 2022, field-programmable gate arrays have had no public price. The only listed pure-play left was LatticeLSCC--, which is less a proxy for the category than a special case. That is what makes Altera's listing the first real test in four years of a question the two stocks in play answer in opposite directions: whether FPGAs are a standalone growth asset or a business that only works bolted inside a larger chipmaker.
First, the timing. On September 10, 2026, Reuters reported that Altera is preparing an IPO that could raise more than $2 billion, with a listing possible as early as late 2026, and that Silver Lake has named Barclays, Citi, JPMorgan and Morgan Stanley as underwriters. The detail that matters for anyone tempted to act on a headline: no per-share range has been filed. Altera has not priced a single share. So the test is defined, but the verdict is not in — and the way to read a listing that does not yet exist is to lock in the two anchors we do have and see what a price must clear to mean anything.
The two anchors
Quick Backtesting Tool
The private anchor is the $8.75 billion mark Silver Lake set when it bought 51% of Altera for $4.46 billion, a deal agreed in April 2025 and closed last September, with Intel retaining 49%. Hold onto the context: Intel paid roughly $16.7 billion for Altera in 2015. The carve-out value is about half the entry price — Exhibit A in the case that Altera only worked, or failed, inside IntelINTC--.
The public anchor is Lattice. At a market cap near $16.2 billion on about $650 million of trailing-twelve-month sales, Lattice trades at roughly 25x revenue, carries a 68% gross margin, and is growing: revenue climbed from $124 million in the second quarter of 2025 to $201 million a year later. That is the multiple the "FPGA asset class" story rests on.
Line Altera up against it
Altera's numbers are not in the same zip code. In fiscal 2024 Altera generated $1.54 billion of revenue, a $361 million GAAP gross margin, and a $615 million GAAP operating loss. Under CEO Raghib Hussain the business has turned: revenue rose about 20% in the prior year and operating income more than doubled, with mid-20% growth expected in 2026.
Run the arithmetic at the private mark. Against fiscal 2024's $1.54 billion, the $8.75 billion valuation is 5.7x revenue; push it to a roughly $2.3 billion 2026 run-rate and it drops toward 4x. Against Lattice's ~25x, that is not a modest discount. It is a different asset class wearing the same label.
This is the two-market split, and it is the point a single "pure-play FPGA" ticker blurs. Lattice is a $650 million-revenue, 68%-gross-margin, edge-and-secure-compute niche growing at a third a year. Altera is a multi-billion-dollar, cyclical data-center and telecom business that just clawed out of a loss-making trough. The gross-margin gap — Lattice's 68% against Altera's ~24% in fiscal 2024 — is the cleanest evidence that the two are not comparable economics. Slap Lattice's 25x sales on Altera and you get a valuation above $35 billion, which is nonsense on its face.
What a price must do to validate or to confirm
The listing only answers the question if it diverges clearly from the $8.75 billion mark.
Price meaningfully above it, and the carve-out was underpriced, FPGAs are a viable standalone growth business, and Intel's 49% retained stake becomes a liquid public mark that strengthens the sum-of-parts — an outcome reinforced by how much Intel has already re-rated, with its own stock up more than 170% this year.
Price at or below the mark, and the carve-out reads as a distressed exit and the "FPGA growth asset" story loses its reference point. That is the outcome that undercuts the ~25x-sales justification Lattice's holders are implicitly paying for. For Lattice, the structural risk in this listing is blunt: the market gets a far larger FPGA comp at a far lower multiple, and the category reprices down toward it.
What to watch
Two numbers, not one. First, the per-share price multiplied into a total valuation once a range is filed. Second, whether that figure sits clearly above the $8.75 billion mark (validate) or at or below it (confirm). The reported $2 billion raise tells you nothing by itself — at a 15% float it implies roughly $13 billion; at 25%, near $8 billion. The float is the missing variable. And if the IPO is pulled, or the pricing lands so close to the mark that there is no discernible divergence, the test produces no signal and the headline is noise.
For a holder, the action is to hold off. This is a watchlist item defined by two numbers, not a trade. The listing has not priced. But because Altera has no choice but to price its business for the first time since Xilinx disappeared, its filing will tell anyone holding Lattice or Intel which side of the FPGA split is real.
