Three small-cap ways to own the AI hardware buildout — a profitable defense play, an O-RAN doubling, and an accelerating test ramp. All three are growing. Only one combines the growth with the quality. Here's the pick.
Three small-caps, three different bets on the same theme. AI hardware — but at three very different points on the growth/quality curve.
All three ride the AI buildout, but they're not interchangeable — each sells something different, makes money differently, and carries a different risk. The whole comparison comes down to one question: which one pairs the growth with the quality to back it up?
The chart is the thesis in one frame. AMPG has the fastest growth (far right) but sits below the profitability line. TRT has fast acceleration too, but the lowest margins of the three. OSS is the only name in the upper half — solid growth and real profitability. In a market that has spent this whole research cycle punishing growth-without-profit (CRWV) and rewarding profitable growth (NBIS, CRDO), the name above the line is the structurally favored one.
Picture a graph: side-to-side is how fast sales grow, up-and-down is how profitable the company is. AMPG and TRT are far to the right (fast growth) but low down (little or no profit). OSS sits up high — it grows well AND makes money. That upper-right zone is where you want to be, and only OSS is there.
Every metric that matters, in one table. Same lens applied to all three.
| Metric | OSS | AMPG | TRT |
|---|---|---|---|
| What it sells | Rugged defense edge-AI compute | O-RAN 5G radios / RF | Burn-in / chip reliability test |
| Latest revenue growth | +55% | ~100% guided (after +163%) | +124% (accel) |
| Gross margin | 51.6% (record) | 48% | 16% (fell from 27%) |
| Profitable? | Yes — +EBITDA, +FCF | No — net loss | ~Breakeven — Q3 net loss |
| Balance sheet | No debt · ~$34M cash | No debt · diluting | D/E 0.11 · diluted ~$10M |
| Demand signal | 1.8x book-to-bill | 2 LOIs shipping + 3rd >$70M | ~$10M AI-GPU orders |
| Backlog visibility | $65M+ P-8 thru 2027 | $20M+ · LOI tranches | Order-by-order |
| TAM tailwind | Defense edge-AI ($10B+) | O-RAN re-shoring + AI-RAN | Burn-in test (~$1B by 2027) |
| The catch | Ran 410% · lumpy | Loss + dilution + H2-weighted | 16% margin layer · narrative overreach |
| Growth-lens score | 8.5 | 8.0 | 7.5 |
Laid out together: AMPG grows fastest, TRT accelerates hardest, but OSS is the only one already profitable with a fat margin, no debt, and the clearest forward orders. Each has a catch — OSS already ran up a lot, AMPG loses money and dilutes, TRT's work is low-margin. But OSS's catch is about price; the other two's catches are about the business itself.
The only one where everything points the same way. Growth + margins + balance sheet, all aligned.
Why it wins the comparison: OSS is the single name here where you don't have to choose between growth and quality. It grew revenue 55%, posted a record 51.6% gross margin, swung to positive EBITDA and record free cash flow, carries no debt and ~$34M cash, and prints a 1.8x book-to-bill on a layered multi-year pipeline (P-8 through 2027, a $25M+ medical deal, autonomy, aerospace). It already did $72M revenue in a prior cycle, so the runway back up is proven — now at much higher margins after shedding the low-margin Bressner unit.
The catch is price, not the business: it's up ~410% and bumps against analyst targets, and three directors sold into the strength. That's a real consideration — but it's a valuation caveat on a proven business, which is a categorically better problem than AMPG's "still loss-making" or TRT's "16% margins." Growth-lens rating: 8.5 quality / 7.0 entry.
OSS is the one that's firing on all cylinders — growing, profitable, debt-free, with clear orders booked years out. Its only real downside is that the stock already went up a lot. That's a much nicer problem to have than "loses money" or "thin margins," which is why it's the pick.
The steepest top line — if you can stomach the loss and dilution. The aggressive second pick.
The bull edge: AMPG has the most violent revenue ramp of the three — +163% in FY25, guided to roughly double again to $50M+ in FY26 — and the catalysts are converting, with both the $100M and $40M LOIs now in concurrent production shipments and a third LOI >$70M referenced. Gross profit grew 116% on 49% revenue — operating leverage is starting. If the H2 ramp lands, this has the highest top-line torque in the group.
Why it's the second pick, not the first: it's still loss-making ($(1.52)M Q1) and funds the build with dilution — share count is rising. So you're paying for a steeper ramp with worse profitability and more shares. That's a legitimate higher-risk/higher-reward trade for someone who wants maximum growth torque and can sit through the losses, but it doesn't have OSS's "everything aligned" quality. Growth-lens rating: 8.0 trajectory / 6.5 entry.
AMPG grows the fastest and its big orders are shipping — exciting. But it still loses money and keeps selling new shares to fund growth, which dilutes owners. It's the aggressive pick: more growth, more risk. Good for someone who wants maximum upside and can handle the bumps, but not as clean as OSS.
Fastest acceleration, thinnest margins. The most speculative — one number decides it.
The bull edge: TRT has the steepest acceleration (+58/82/124%) into a real ~$1B-by-2027 burn-in TAM, with ~$10M in recurring AI-GPU orders and capacity going in. The growth is genuine and order-backed.
Why it's third: the AEHR contrast tells the story — AEHR sells the testing machine at ~60% margins; TRT supplies boards and runs the test floor at 16%. Same supercycle, low-margin layer. So revenue doubles while earnings barely move, and the entire re-rate gates on SBS gross margin recovering toward 20%+ — which hasn't happened yet. Layer on the unresolved narrative issues (the unproven "Micron is the 41% customer" inference, the quant-13F misread, the omitted dilution) and it's the most speculative of the three. Growth-lens rating: 7.5 trajectory / 6.0 business.
TRT is growing fastest of all, but it does the low-paid part of the work — running the test floor at thin margins — so the growth doesn't turn into profit yet. The whole bet is whether those margins recover. Until they do, it's the riskiest of the three.
If you buy one: OSS. The only name where growth, margins, and balance sheet all point up together.
All three are real growth stories — this isn't "one good, two bad." But on risk-adjusted growth, OSS is the pick, because it's the only one that doesn't ask you to also bet on a fix: AMPG needs its losses to narrow and dilution to stop; TRT needs a margin inflection that hasn't shown up. OSS already cleared both bars — it's profitable, debt-free, and converting a layered backlog — so the only open question is entry price, which is the easiest risk to manage (wait for a dip, size in). The ranking: OSS (best risk/reward) → AMPG (highest torque, accept the loss/dilution) → TRT (most speculative, gated on margin).
How to hold all three if you want the basket: OSS as the core (quality anchor), AMPG as the aggressive satellite (size for volatility, key to the H2 ramp), TRT as a small speculative (key strictly to SBS margin recovery — exit if it stays 16%). But if it's one name, it's OSS.
All three can work, but OSS is the best buy because it's the only one that doesn't require a second thing to go right — it's already profitable and growing. AMPG needs its losses to shrink; TRT needs its margins to recover. OSS just needs a reasonable entry price. Pick: OSS first, AMPG for aggressive growth, TRT only as a small bet on margins improving.
Want the entry levels and position sizing for all three?
Join the Discord to find out! →OSS Q1 FY26 8-K, transcript & FY26 outlook (May 2026 — +55%, 51.6% GM, +EBITDA, 1.8x book-to-bill, $34M cash, $65M+ P-8, layered pipeline; prior-cycle $72.4M FY22 peak) · AMPG FY25 & Q1 FY26 8-Ks/10-Q (+163% FY25, $50M+ FY26 guide, $100M+$40M LOIs shipping, 48% GM, $(1.52)M net loss) · TRT Q2/Q3 FY26 8-Ks & releases (+82%/+124%, SBS +141%, 16% GM, ~$10M AI-GPU orders, $10M raise) · burn-in TAM & AEHR comp (~$1B by 2027, ~60% system margins) · analyst targets via MarketBeat/Stockanalysis. Figures from SEC filings & transcripts; growth-lens comparison, not a recommendation. Not live IBKR pricing.