A-Portfolio

Bias / Risk / Action

• Bias: Bullish on the company. Ambiq nearly doubled quarterly sales, expanded the profit it keeps from each chip, and has a real technical advantage in devices that cannot waste battery power.
• Risk: Three customers supplied 77.5% of quarterly sales, one factory partner makes every wafer, the company still loses money, and the stock trades at roughly 9.4 times its current annualized sales pace.
• Action: WATCH. No position. I want the company. I do not want today's bill. Rebuild the case at $50 to $55, or after two more quarters prove growth, customer diversification, and shrinking losses.

Why this matters - in plain English

Ambiq designs tiny computer chips that let watches, medical devices, smart glasses, and factory sensors run artificial intelligence without murdering the battery. Sales nearly doubled because customers launched new products with Ambiq chips inside, and the company now keeps more money from each sale before overhead. If Ambiq becomes the standard low-power brain inside millions of always-on devices, the business could become far more valuable. Derek should care because this is a credible category-leader story, but the stock already charges us for a fat slice of that future.

The company in one breath

Ambiq sells low-power chips and the software needed to run artificial intelligence directly inside small devices. Device makers pay Ambiq because its patented SPOT design can cut power use by two to five times versus conventional chip designs, which can mean longer battery life without making the product dumb.

The company does not own a chip factory. It designs the product, then Taiwan Semiconductor Manufacturing Company, better known as TSMC, makes every wafer. That keeps Ambiq's factories off the balance sheet, but it also puts every egg in one very sophisticated Taiwanese basket.

Why now

Ambiq reported Q2 on August 11, and this was the cleanest growth print in today's screen. I looked across recent IPOs and fresh earnings names including Figma, Chime, eToro, Circle, CoreWeave, CAVA, PubMatic, Credo, and Ambiq. Several have bigger businesses. Ambiq had the sharpest combination of recent-IPO status, accelerating sales, improving gross profit, and a defensible future-category thesis without repeating an archived ticker.

The quarter gave us real evidence:

1. GAAP net sales hit $33.9 million, up 89.7%. That means customers bought almost twice as many dollars of Ambiq products as a year ago. For a young chip company, anything above 30% is strong, so 89.7% is a fire hose.

2. GAAP gross margin reached 45.0%, up from 40.1%. Gross margin is the cut left after making the chips. Ambiq kept 45 cents from each sales dollar before engineers and office costs, and a healthy specialized chip company eventually wants 50 cents or better.

3. GAAP operating loss improved only slightly to $8.7 million from $8.8 million. The sales engine grew, but Ambiq spent aggressively on engineering and corporate staff. Good growth should eventually shrink that loss much faster.

4. Management guided Q3 sales to $36 million to $37 million. That would mark a sixth straight quarter of sequential growth, meaning each quarter beat the one before it. It is management's target, not a GAAP result.

5. Management says second-half sales should more than double year over year. That is a forecast, not money in the bank. The company also says supply shortages will limit how much it can ship.

The dumb take is that 90% growth plus the letters AI equals an automatic buy. That is how people pay steakhouse prices for a sandwich. The business is improving. The valuation still gets a vote.

The numbers, translated

GAAP results

First-half sales reached $59.0 million, up 75.5%. That means the acceleration lasted longer than one lucky quarter. Good looks like several quarters above 30% growth while losses fall as a share of sales.

First-half operating cash use reached $20.7 million, versus $10.5 million a year ago. Operating cash use is money the daily business consumed. Good looks like that number shrinking toward zero as sales rise, not doubling because inventory and unpaid customer bills soaked up cash.

The balance sheet

Ambiq held $366.8 million of cash and reported only $31.0 million of total liabilities at June 30. That is a fortress for a company this size. At the first-half rate, cash used by operations, equipment, and purchased technology totaled about $26.6 million, so the balance sheet gives management years to build.

Do not confuse the cash pile with business success. Ambiq raised $245.5 million from two stock offerings during the first half. The cash register did not create that money. Shareholders did.

Shares outstanding rose to 24.15 million from 18.32 million at year-end, mostly because Ambiq sold stock at $31 in January and $78 in June. That is roughly 32% more slices in the pizza. The pizza grew, but your slice still matters.

Mechanical valuation

Using the verified $67.84 share price, 24.146 million June shares, and $366.8 million of cash:

• Market value is approximately $1.64 billion. That is what investors currently pay for all common shares.
• Enterprise value is approximately $1.27 billion. That is the operating-business price after subtracting cash.
• Q2 sales annualized are approximately $135.6 million. Annualized means multiplying one quarter by four, not forecasting the year.
• Enterprise value equals approximately 9.4 times that annualized sales pace. A young chip winner can earn a premium, but good entry math usually needs either a lower multiple or a much larger proven sales base.
• Cash equals approximately $15.19 per share. That cushion helps fund growth, but management raised most of it by selling more shares.

The stock trades 182.7% above its $24 IPO price and 13.0% below the June follow-on price of $78. IPO buyers won. June buyers have not. Neither fact makes today's price cheap.

What I like

The power advantage solves a real problem. Cloud AI can plug into a wall. A ring, glucose monitor, or security sensor cannot. Ambiq says its SPOT design uses two to five times less power than conventional designs, and more than 300 million devices have already shipped with its technology.

The product mix is getting better. Management says newer products raised average selling prices, which helped gross profit grow faster than sales. Selling smarter chips for more money beats winning by shoveling cheap units out the door.

The balance sheet buys time. Ambiq can fund engineers, secure supply, and build software without begging lenders next quarter. In a chip race, cash is oxygen.

What worries me

Three customers control the story. The top three supplied 28.9%, 24.4%, and 24.2% of Q2 sales. Lose one product slot and the growth chart can fall off the bar stool.

TSMC makes every wafer. Ambiq has no second wafer supplier. A capacity squeeze, manufacturing problem, or disruption around Taiwan can stop shipments even when customer demand stays strong.

Profit has not caught the growth yet. Operating expenses rose 50.3% while sales rose 89.7%, which is directionally good, but Ambiq still lost 25.7 cents at the operating line for every sales dollar. Good looks like that loss dropping below 10 cents, then crossing into profit.

Dilution is not theoretical. Ambiq sold stock twice in six months, executives adopted preset selling plans, and 4.38 million potential shares sat outside the diluted loss calculation. The company can build something great while repeated share issuance eats the return per share.

The crowd evidence is weak. The last-30-days scan found two tiny YouTube videos and mostly unrelated social results. One transcript nailed the real issue: "This management has sold stock twice in 6 months." I do not know whether broad investor interest will develop. The filing, not social applause, carries this call.

What the market may be missing

The market may still frame Ambiq as a smartwatch-chip supplier. The larger bet is a low-power computing platform for medical devices, industrial sensors, smart buildings, glasses, and other machines that need to think locally. Local computing means the device can respond faster, protect private data, and work without constantly calling a distant data center.

Ambiq's new heliaCORE, compressionKIT, and heliaPROFILER software matter because hardware alone rarely wins a platform war. Developers need tools that help their AI models fit and run on the chip. The best stove loses if nobody knows how to cook on it.

The catch is timing. Body-worn devices still drive much of today's revenue, and one new major customer launch helped Q2 growth. We need proof that medical and industrial demand broadens the base before paying for a universal edge-AI empire.

Bull / base / bear

These are 12 to 24 month guardrails, not targets. I apply sales multiples to three possible future annual sales levels, add June cash, and divide by June shares. I do not know which case will happen.

The base case gives us a strong business and still loses money from today's quote. The bull case needs a 62% larger sales base than the current Q2 annualized pace and a rich eight-times-sales valuation. That can happen. It is not enough upside to pay first and ask questions later.

What would make this a buy

Price path: Rebuild the model at $50 to $55. At that range, enterprise value falls to roughly 6.2 to 7.1 times the current annualized sales pace. That is still not cheap, but it gives the base case room to work.

Proof path: I would pay more after two quarters with all four conditions:

1. Sales growth stays above 50%.

2. GAAP gross margin stays above 45%.

3. GAAP operating loss falls below 15% of sales.

4. The top three customers fall below 70% of sales because new customers grow, not because an old one collapses.

A score below 80 means no position. There is no starter size today.

What makes me walk away

1. Sales growth falls below 30% for two consecutive quarters before the company reaches GAAP operating profit.

2. Q3 sales miss the $36 million to $37 million management range by more than 10% without a clearly disclosed shipment delay.

3. GAAP gross margin falls below 40% for two quarters.

4. The top three customers stay above 80% of sales through the first half of 2027.

5. A major customer removes Ambiq from a device design or delays a launch without a replacement win.

6. TSMC cannot provide capacity and Ambiq misses demand for two quarters.

7. Operating cash use exceeds $60 million over any trailing twelve months without a matching jump in durable customer demand.

8. Shares outstanding grow more than 15% in a year again without per-share sales and gross profit growing faster.

9. A competitor matches Ambiq's power advantage at similar cost and with better developer tools.

The verdict

Ambiq earned every growth point. It lost the candidate badge on valuation, concentration, and unfinished profit proof. Great company candidate. Bad place to get impatient.

The DBTech read

Customer concentration tells you how much of a company's dinner depends on a few tables. Ambiq's top three customers supplied 77.5% of Q2 sales, so one lost product slot can wreck the quarter. High concentration is not automatically bad when customers are strong, but it makes results fragile and gives buyers leverage on price. Good looks like new customers growing until the top three fall below 60% without old sales disappearing. Always ask who pays the company before admiring how fast revenue grew.

Sources and as-of date

As of: August 13, 2026, 10:54 a.m. ET. Price is the last verified trade at 10:54:39 a.m. ET from the Robinhood broker feed. The official August 12 close was $65.59 from the SIP listing-exchange close.

1. Ambiq Q2 2026 Form 10-Q, filed August 11, 2026:

https://www.sec.gov/Archives/edgar/data/1500412/000119312526344145/ambq-20260630.htm

2. Ambiq Q2 2026 earnings release, Exhibit 99.1, August 11, 2026:

https://www.sec.gov/Archives/edgar/data/1500412/000119312526343443/ambq-ex99_1.htm

3. Ambiq Form 8-K, August 11, 2026:

https://www.sec.gov/Archives/edgar/data/1500412/000119312526343443/ambq-20260811.htm

4. Ambiq June 2026 follow-on prospectus, including business and concentration disclosures:

https://www.sec.gov/Archives/edgar/data/1500412/000119312526281376/d152097d424b4.htm

5. Ambiq official company site and product positioning:

https://ambiq.com/

6. Nasdaq August 2026 IPO calendar, used in the candidate screen:

https://api.nasdaq.com/api/ipo/calendar?date=2026-08

7. Robinhood broker feed and official prior-session close, August 13, 2026:

https://robinhood.com/us/en/stocks/AMBQ/

Calculation notes

All growth rates, margins, market values, valuation multiples, customer concentration, cash runway, and scenarios were calculated with Python Decimal arithmetic from the cited inputs. Enterprise value equals market value less cash because the filed balance sheet showed no financial debt. Scenario sales are my assumptions, not company guidance or analyst forecasts. Potential dilution can make per-share outcomes worse.

Bottom line: Ambiq has the technology, growth, and balance sheet of a future edge-AI leader. At $67.84, the market already knows. Watch it, demand diversification and profit, and make the price come to us.

Research only. Derek makes the allocation decision.

DBTech45