On August 4, 2026, AMD reported the best quarter in its history.
Then the stock fell about 9%.
That gap between the numbers and the reaction is the most useful thing in the whole report, because it points at the one line most coverage skipped: free cash flow went down while revenue went up.
This article works through what AMD actually reported, where the twelve-month analyst targets sit, and what a 2030 outcome would require — using AMD's own stated earnings goal rather than a proprietary model.
Key Takeaways
AMD's Q2 2026 revenue reached a record $11.5 billion, up 50% year over year.
Data center revenue more than doubled to $6.7 billion and now makes up 58% of the company.
Free cash flow fell 39% from the previous quarter to $1.56 billion as capital spending more than doubled.
Shares closed the August 4 session up about 7%, then dropped roughly 9% in after-hours trading to $472.
Ten named Wall Street firms hold twelve-month targets between $500 and $730, against a $579 average across 51 analysts.
AMD's own goal of more than $20 in annual earnings per share is the cleanest anchor available for any 2030 estimate.
Here is the full quarter in one place.
Metric | Q2 2026 | Q1 2026 | Q2 2025 | Change vs Q2 2025 |
Revenue | $11,536M | $10,253M | $7,685M | Up 50% |
Data Center revenue | $6,718M | $5,775M | $3,240M | Up 107% |
Client revenue | $3,062M | $2,885M | $2,499M | Up 23% |
Gaming revenue | $779M | $720M | $1,122M | Down 31% |
Embedded revenue | $977M | $873M | $824M | Up 19% |
GAAP gross margin | 54% | 53% | 40% | Up 14 points |
Non-GAAP gross margin | 56% | 55% | 43% | Up 13 points |
GAAP diluted EPS | $1.38 | $0.84 | $0.54 | Up 156% |
Non-GAAP diluted EPS | $1.66 | $1.37 | $0.48 | Up 246% |
Free cash flow | $1,558M | $2,566M | $1,180M | Up 32% |
Capital expenditure | $808M | $389M | $282M | Up 186% |
Source: AMD Q2 2026 financial results, August 4, 2026.
One footnote matters when reading the year-over-year columns.
Data center revenue reached $6,718 million, up 107% from a year earlier and 16% from the first quarter.
That segment alone now accounts for 58% of everything AMD sells, and it produced $2,103 million of operating income against a $155 million loss in the same quarter last year.
Two years ago this was the growth story running alongside the main business.
It is now the main business, and the client, gaming and embedded segments together contribute less than half of revenue for the first time.
The mix shift is the single most durable fact in the report, because it changes which end markets AMD's earnings are actually exposed to.
Third-quarter guidance came in at about $13 billion, plus or minus $300 million, above the roughly $12.5 billion the Street had been modelling before the report — though several outlets described the guide as merely in line with expectations once estimates were updated.
What did not look good sits in the cash flow statement.
Free cash flow margin went from 25% to 14%, capital expenditure more than doubled to $808 million, and accounts receivable consumed another $1,246 million of cash.
The read is straightforward: AMD is winning this market, and the capital required to keep winning it is rising faster than the revenue it brings in.
One more line is worth separating out — GAAP net income of $2,297 million includes a $483 million non-cash gain on long-term investments, so roughly a fifth of the headline profit did not come from selling chips.
All ten of the named firm targets below sit above where the stock trades, though the full analyst pool is wider on both sides.
Firm | Analyst | 12-month target | Date |
UBS | | $730 | July 24, 2026 |
Benchmark | | $685 | Reaffirmed August 4, 2026 |
Barclays | | $665 | June 1, 2026 |
Goldman Sachs | | $640 | July 5, 2026 |
Jefferies | | $640 | July 24, 2026 |
Stifel | | $635 | July 2026 |
Mizuho | | $625 | July 27, 2026 |
Wedbush | | $600 | July 27, 2026 |
Citi | | $575 | June 2026 |
Susquehanna | | $500 | July 30, 2026 |
These are not all built the same way, which is worth saying plainly rather than presenting the spread as one settled number.
The higher targets are mostly built on the agentic AI server CPU thesis, the lower ones on near-term GPU shipment estimates, and those two methods can disagree for years before either is proven wrong.
Scenario ranges below are anchored to documented price levels rather than to a single forecast.
Scenario | 12-month range | What would have to happen |
Bear | $320 to $470 | Q3 revenue lands below the $12.7 billion low end of guidance, or Helios volume slips out of the fourth quarter; the $472 post-earnings level fails to hold |
Base | $470 to $580 | Q3 delivers inside the guided $12.7 to $13.3 billion range and the stock works back toward the $579 consensus average |
Bull | $580 to $730 | The June 30 high near $585 is reclaimed on evidence the 2027 data center doubling is on track, opening the path toward the highest published target |
The honest read is that these three sit closer together in likelihood than the target table alone suggests.
Demand is accelerating and management raised its own long-term bar on the same call, but cash conversion weakened and a well-funded competitor is entering the segment the whole thesis rests on.
When two first-order signals point in opposite directions like this, treating one scenario as clearly dominant would be manufacturing a view the evidence does not support.
A five-year number is only as good as the assumption underneath it, so it is worth showing the assumption rather than the model.
That figure is AMD's own, it is on the record, and it is the cleanest available anchor for a 2030 estimate.
The rest is a multiple, and the multiple is the part nobody can forecast.
Scenario | 2030 range | What would have to be true |
Bear | Below $500 | Annual earnings stall near or below the $20 goal and the market pays about 25 times |
Base | $700 to $900 | Earnings run modestly above $20 a share and the multiple settles in the mid-30s to low 40s |
Bull | $1,100 to $1,300 | Earnings run well past $20 and the market keeps paying 55 to 60 times |
This is arithmetic on a stated company goal, not a proprietary model, and it can be recalculated by anyone as the earnings picture changes.
Baird's $1,250, derived from its own preliminary 2030 earnings estimate, lands inside that bull band. It is one firm's view, but it is built on the same kind of arithmetic used here rather than a different method.
The bull case is a chain, and every link is something management has already put on the record.
Those commitments give revenue visibility that did not exist a year ago, and they are the reason ten separate firms raised targets into the print.
None of it is a result yet, and each step assumes the one before it lands on time.
The clearest risk is competitive, and it is newer than most AMD coverage reflects.
Nvidia CFO Colette Kress told analysts on the company's most recent earnings call that Nvidia has visibility to nearly $20 billion in total CPU revenue this fiscal year across its Grace and Vera processors. Nvidia detailed the Vera chip itself in July 2026, claiming roughly 50% faster performance than comparable x86 processors on its own benchmarks. Both companies now point at the same number. AMD raised its own server CPU forecast at Advancing AI 2026 in July to more than $200 billion by 2030, up from about $25 billion today and from the $120 billion it had given just two months earlier. Nvidia CFO Colette Kress described the same opportunity as a brand-new $200 billion market for her company. The market size is not in dispute. Who supplies it is. Valuation is the second risk, and the reported multiple varies enormously depending on how it is calculated.
Published figures for AMD through mid-2026 range from about 54 times forward earnings to more than 200 times trailing GAAP earnings, because some use adjusted numbers and some do not, and because forward estimates were rising fast all year.
Anyone quoting a single P/E for AMD without saying which basis they used is not telling you much.
Third is capital intensity, already visible in the Q2 numbers — if free cash flow keeps compressing while capex climbs, the earnings that the entire 2030 arithmetic depends on arrive later and cost more to produce.
When MEXC Research weighs a price scenario for a US stock, it puts verifiable first-party data first — real-time Level 1 quotes and how a stock actually traded in the extended and overnight sessions — above secondhand commentary and above single-number model outputs.
That weighting stays the same from one article to the next, including when the conclusion points the other way from the last one.
August 4, 2026 is a clean demonstration of why the session structure matters more than most price commentary admits.
AMD released results at 4:15 PM Eastern Time, fifteen minutes after the US market closed.
By the time the next regular session opened, the entire 7% gain from that day's trading had been erased and then some, with shares at roughly $472 against a $513 close.
An investor who could only act between 9:30 AM and 4:00 PM Eastern was not in the market for any part of that move.
This is a structural feature of earnings-driven stocks, not a one-off: US companies release results outside regular hours precisely so the market has time to digest them, which means the repricing happens when most retail access is shut.
Signals currently point in different directions, and that is the honest read rather than a hedge.
Demand is accelerating and management raised its own long-term bar on the same call that triggered the selloff, while cash conversion weakened sharply in the same filing — so the bear, base and bull scenarios above should be treated as closer to comparable in likelihood than the analyst target spread implies.
What is the AMD stock price prediction for 2030?
Using AMD's own stated goal of more than $20 in annual earnings per share, a 2030 share price lands roughly between $500 and $1,300 depending entirely on what multiple the market pays.
What is the AMD stock price target for the next 12 months?
Published targets from named firms run from $500 to $730, against a $579.11 average across 51 analysts tracked by S&P Global Market Intelligence.
Why did AMD stock fall after record Q2 2026 earnings?
Free cash flow dropped 39% to $1.56 billion while capital expenditure more than doubled to $808 million, so the quarter showed rising costs to sustain the growth alongside the record revenue.
How much revenue did AMD report in Q2 2026?
AMD reported $11.54 billion for the quarter ended June 27, 2026, up 50% year over year, with data center revenue of $6.72 billion.
Is AMD's data center growth expected to continue into 2027?
Lisa Su told analysts on August 4, 2026 that AMD expects data center sales to more than double in 2027 and server revenue to grow more than 70%.
What is AMD's P/E ratio?
Published figures for AMD in mid-2026 range from about 54 times forward earnings to over 200 times trailing GAAP earnings, so the basis matters more than the number.
What is the biggest risk to AMD's server CPU business?
Nvidia's Vera CPU projects more than $20 billion in server CPU revenue in its first full year, and Arm-based designs are forecast to take a large share of a market where AMD's EPYC is x86.
Can you trade AMD stock outside US market hours?
Extended and overnight sessions exist for US equities, and AMD's Q2 2026 move happened almost entirely in them — results were released at 4:15 PM Eastern, after the regular session closed.
AMD is no longer a chipmaker with a data center business attached.
Data center is 58% of revenue, it more than doubled year over year, and management has guided it to double again in 2027.
The reason the stock fell anyway is in the same filing: free cash flow went down 39% in the quarter revenue went up 13%, because the capital needed to hold this position is climbing faster than the revenue it generates.
Both of those things are true at once, and any AMD stock price prediction that only carries one of them is telling half the story.
For 2030, the useful question is not what a model outputs but whether AMD earns meaningfully more than the $20 a share it has told investors to expect — and what the market decides that is worth.