Semiconductors Are Winning 2026. Our AI Score Says One Beaten-Down Sector Is Next.

By AltIndex Research · 10 min read · August 9, 6:44 am

Semiconductors were the best-performing sector of the last six months, with the median chip stock up about 29%. The surprise is who led: not Nvidia, which returned 18% and trailed the sector it defines, but the memory makers, custom-silicon designers, and equipment suppliers one tier down. And the run may not be done. Our AI Score ranks chips second of every sector looking forward, while the highest score of all belongs to a sector the market has overlooked this year, Engineering and Construction, the builders of the data-center boom.

Semiconductors are winning 2026, and a specific set of names did the winning. Micron doubled. AMD gained 124%. Marvell ran up 166%. When companies this large move like small caps, the capital behind them is enormous, and it is all chasing the silicon that turns AI spending into compute. That is where this story starts. It ends in the sector our data likes even more, one almost no one is watching. First, the run that already happened.

The Sector That Left Everything Else Behind

We ranked every sector by the median six-month return of its stocks rather than the average, because a single 300% outlier can drag an average upward and make a mediocre sector look like a winner. On a median basis, semiconductors sat at the top at roughly 29%, and they did it with breadth: 67 active chip stocks in the sample, of which 56 finished positive. Cloud computing followed near 24%, and digital health near 27% on a much smaller set of names.

The point of the median is breadth, and semiconductors had it. This was not two rockets and a field of dead weight. Twenty-three chip stocks rose more than 50% over the six months, and eleven of them at least doubled. When that many names in one sector double at once, the story is usually macro, and here the macro is simple: the capital going into AI data centers has to pass through silicon on its way to becoming compute.

6-month sector returns, median per stock

The Biggest Names Were the Slowest

Here is the part that runs against intuition. If you owned the semiconductor sector through its best six-month stretch in years by buying the household names, you underperformed the sector badly. Nvidia rose 18%, Broadcom 25%, ASML 22%, Qualcomm 22%, TSMC 19%. Every one of those trailed the median chip stock.

The leaders were the names one rung down the ladder. Marvell Technology gained 166%. Micron gained 129%. ARM Holdings gained 127% and AMD 124%. Intel, left for dead by much of the market a year ago, doubled. And the single best performer in the entire sector was a company most investors have never heard of: Aehr Test Systems, a maker of semiconductor test and burn-in equipment, up 274%.

Top semiconductor stocks

Sorted by six-month return. Check the best semiconductor stocks for a real-time toplist.

Memory, Custom Silicon, and the Picks and Shovels

The winners cluster into three clear groups, and each one maps to a different part of the AI buildout.

The first group is memory. AI accelerators are starved for high-bandwidth memory, and Micron is the primary US-listed way to own that shortage. The stock rose 129%, yet it still trades at under 20 times earnings. That combination, a stock that doubles while its valuation multiple stays low, only happens when profits are rising as fast as the price. The earnings grew into the run rather than the other way around, which is why our fundamental read on Micron remains firm even after the move.

The second group is custom silicon and interconnect, the plumbing that ties thousands of chips into a single training cluster. This is where Marvell, Credo Technology at 102%, Astera Labs at 78%, and ARM all live. As hyperscalers design more of their own accelerators, the value shifts toward the firms that supply the connective tissue, and the market repriced that entire group upward.

The third group is equipment, the literal picks and shovels. Every new fab and every capacity expansion flows through the same short list of tool makers. Aehr led at 274%, Ichor rose 100%, Veeco 73%, Applied Materials 64%, Axcelis 64%, with KLA and Lam Research both up in the mid-30s. When chipmakers commit to years of capacity growth, the equipment suppliers book the orders first, and their stocks tend to move before the volume shows up in revenue.

What our Data Says Comes Next

Six-month returns are history. The more valuable question is which sectors are set up to keep working, and that is what our AI Score is built to answer. The score blends fundamental, employment, audience, and customer signals into a single 0-to-100 read, where 60 and above is a buy signal, 40 to 59 is a hold, and below 40 is a sell.

We averaged the AI Score across every stock in each sector. One caveat before the ranking: no sector averages a clean buy, because averaging across every name, including the weak ones, pulls the result toward the middle. The signal here is the relative ranking, not the absolute number.

AltIndex AI Score, sector average

Semiconductors rank second of every sector on forward AI Score at 55.0, which is the real headline for anyone worried they missed the trade. A sector that just ran this hard would normally look exhausted on the data. It does not. The fundamental and employment signals under the chip names are still firm, so the score is telling us the run has support rather than froth.

The sector that outranks it is the surprise. Engineering and Construction posts the single highest average AI Score at 57.4, and it did so while its median stock fell 6% over the same six months. That is the mirror image of semiconductors: strong forward signal, weak recent price. It is worth remembering what these companies actually do in 2026. The AI data-center boom is, before it is anything else, a construction project. Someone has to pour the concrete, run the power, and build the shells, and the market has not yet paid for that work the way it has paid for the chips inside.

Inside engineering and construction, the highest scores cluster around the companies that physically build and power the boom. Granite Construction, a heavy civil builder, tops the group with an AI Score of 86 on rising fundamentals and hiring. Quanta Services, which builds electrical grid and power infrastructure, scores 75, and it may be the cleanest data-center play in the group, because every server hall needs power delivered to it before it can run a single chip. KBR sits at 74 on its engineering and project work.

The materials suppliers round out the list, with Vulcan Materials, the largest US producer of construction aggregates, at 65, and Owens Corning, in insulation and roofing, at 64. What ties them together is fundamental strength tied to the buildout, not a shared price trend. Over the last six months the group split hard: Vulcan fell 12%, KBR 9%, and Granite 6%, while Quanta ran up 31% and Owens Corning gained 12%. The market has already rewarded the names closest to data-center power and left the rest behind, even as our AI Score rates the whole group among the strongest in the market looking forward.

Top engineering & construction stocks

Sorted by AI Score. Six-month price change shows the split within the group: the market has rewarded Quanta and Owens Corning while leaving Vulcan, KBR, and Granite behind.

The Bottom Line

Semiconductors were the trade of the last six months, led not by Nvidia but by memory, custom silicon, and the equipment makers, and our forward data still ranks the sector second of all, so the momentum has fundamentals behind it rather than just enthusiasm.

The call the market has only half-made is the more interesting one. The highest forward AI Score in our entire universe belongs to the builders of the very boom that made chip investors rich this year, a sector still down at the median even as a few of its names, Quanta and Owens Corning among them, have started to move. When the data rates the diggers of the foundation among the strongest in the market and the price has only begun to follow, that gap is worth watching.

Track AI Scores and alternative data for 3,000+ stocks

Get started with AltIndex →

Disclosure: This article is for informational purposes only and does not constitute investment advice. AltIndex aggregates publicly available alternative data signals. Past signal performance does not guarantee future results. Always do your own due diligence before making investment decisions.