Originally published 2013 · Completely rewritten and updated October 2026
The APU — AMD’s Accelerated Processing Unit, a CPU and GPU on a single chip — is one of the most consequential ideas in modern computing. It took fifteen years to fully work. This is the complete story, from the humble beginnings of Llano to the Ryzen AI 300 series of 2026, where the original dream finally, completely, came true.
2011 — Llano: The Beginning
The first APU, codenamed Llano, combined a K10-era CPU (the aging Phenom II architecture) with Radeon HD 6000-series-derived graphics. Chips like the A8-3850 looked odd on paper — a dated CPU paired with surprisingly capable graphics — but the concept was immediately validated: Llano’s integrated graphics demolished Intel’s HD Graphics of the era. For budget desktops and laptops, it was a revelation. The CPU side was already behind, though, and everyone knew it.
2012–2013 — Trinity and Richland: Piledriver Steps In
Trinity (2012) replaced the old K10 cores with Piledriver modules and upgraded graphics to VLIW4-based Radeon HD 7000-class. Mobile Trinity chips like the A10-4600M brought genuinely playable gaming to mainstream laptops — our own archives from 2013 cover the A10-5750M era. Richland (2013) was a refined Trinity: better power management, higher clocks, same architecture. Solid, iterative, unglamorous — and selling well in budget laptops.
2014 — Kaveri: The HSA Gamble
Kaveri was AMD’s big swing: Steamroller CPU cores, GCN graphics, and HSA (Heterogeneous System Architecture) — the vision of CPU and GPU sharing memory seamlessly. The A10-7850K was a good chip, but HSA’s software revolution never arrived. Kaveri is remembered as the APU that aimed highest and landed… fine.
2015–2016 — Carrizo and Bristol Ridge: Survival Mode
These were tough years. Carrizo (2015) brought meaningful efficiency gains to laptops, and Bristol Ridge (2016) carried the torch while AMD bet the company on Zen. Few mourned them, but they kept AMD in the laptop game during its darkest period.
2017 — Raven Ridge: Zen Changes Everything
Raven Ridge — the Ryzen 5 2400G and Ryzen 3 2200G — was the moment the APU grew up. Zen CPU cores (finally competitive with Intel) paired with Vega graphics. The 2400G could genuinely game at 720p/1080p low settings without a graphics card. It sold enormously well and remains a beloved budget-classic. Mobile Raven Ridge (Ryzen 5 2500U) put real performance in thin laptops.
2019–2020 — Picasso and Renoir: Refinement
Picasso (2019) was a Zen+ refresh. Renoir (2020) was the quiet masterpiece: Zen 2 cores, up to 8 of them, in a 15W laptop chip (Ryzen 7 4800U) that embarrassed Intel’s mobile lineup. Renoir proved the APU could be a premium product, not just a budget compromise.
2021 — Cezanne: Zen 3 Arrives
Cezanne (Ryzen 5000 mobile, e.g., 5800U/5900HX) brought Zen 3’s superb efficiency and IPC to APUs. The 5900HX became a favorite in gaming laptops paired with discrete GPUs — the APU as the ideal companion chip.
2022 — Rembrandt: RDNA 2 Ignites Integrated Graphics
Rembrandt (Ryzen 6000 mobile) was the biggest iGPU leap in years: RDNA 2 graphics with up to 12 compute units, plus DDR5 support. The Ryzen 7 6800U’s 680M graphics could play AAA titles at 1080p low — the “no graphics card needed” dream getting real. This generation also introduced Microsoft Pluton and modern power management.
2023 — Phoenix: Zen 4 + RDNA 3
Phoenix (Ryzen 7040 series) combined Zen 4 with RDNA 3 graphics and added the first NPUs for AI workloads (XDNA). The 7840U/7840HS became the darling of handheld gaming PCs — devices like the ROG Ally proved an APU alone could power a real gaming handheld.
2024 — Hawk Point and Strix Point: The AI Era
Hawk Point (Ryzen 8040) refined Phoenix with a faster NPU. Then Strix Point changed the naming forever: Ryzen AI 300 series. Zen 5 CPU cores, RDNA 3.5 graphics (up to 16 CUs in the top chips), and a 50-TOPS NPU meeting Microsoft’s Copilot+ requirements. The Ryzen AI 9 HX 370’s integrated graphics rivaled entry-level discrete GPUs — a sentence that would have been absurd in 2011.
2025–2026 — Ryzen AI 300 Series: The Dream, Realized
As of late 2026, Strix Point derivatives (including Strix Halo-class high-end APUs with vastly more GPU compute) define the state of the art. A modern premium APU plays AAA games at 1080p medium-high settings with no discrete GPU, handles AI workloads locally, and sips power. The original 2011 promise — “good enough graphics for everyone, on one chip” — hasn’t just been met; it’s been exceeded. Today’s APUs don’t merely suffice; in thin-and-light laptops, they’re the preferred solution.
What’s Next?
AMD’s trajectory points toward ever-tighter CPU/GPU/NPU integration, with future APUs expected to blur the line between “integrated” and “discrete” graphics entirely. The chiplet era may yet bring modular APUs. Whatever comes, the arc is clear: fifteen years from Llano’s compromise to Strix Point’s triumph.
Further reading: gaming laptop buying guide · best gaming laptops of 2026 · RTX 5060 vs RTX 5070
Frequently Asked Questions
What was the first AMD APU?
Llano, launched in 2011 — starting with desktop chips like the A8-3850, followed by mobile Llano parts. It paired K10 CPU cores with Radeon HD 6000-series-derived graphics.
Do AMD APUs still exist in 2026?
Yes — though AMD now markets them as Ryzen AI mobile processors. Every Ryzen AI 300-series laptop chip is an APU in the original sense: CPU, GPU, and now NPU on one die.
What's the best APU for gaming without a graphics card?
As of 2026, the top Strix Point / Ryzen AI 300 chips (and especially the higher-tier 'Halo'-class APUs) offer the best integrated gaming performance ever shipped — comfortably playing modern titles at 1080p.
APU vs CPU + discrete GPU: which for a gaming laptop?
For maximum gaming performance, a discrete GPU (like an RTX 5060 or better) still wins. But the gap has narrowed so dramatically that many buyers are now better served by a premium APU's efficiency, battery life, and lower cost. Our buying guide breaks down the decision.
The Console Connection: APUs That Conquered Living Rooms
No history of the APU is complete without the consoles. In 2013, both Sony and Microsoft bet their next generation on AMD semi-custom APUs: the PlayStation 4 and Xbox One each shipped with an AMD Jaguar-based APU combining x86 CPU cores with GCN graphics on one die.
This was the APU’s quiet triumph. While PC enthusiasts debated Kaveri’s HSA ambitions, tens of millions of consoles proved the concept at scale: a single AMD chip could power an entire generation of gaming. The PS4’s APU delivered remarkably capable 1080p gaming from a ~$400 box — arguably the best price-to-performance gaming device of its era.
The console wins funded AMD through its darkest years and validated the APU thesis so thoroughly that the PS5 and Xbox Series X|S (2020) doubled down with Zen 2 + RDNA 2 semi-custom APUs. In a very real sense, the APU won the living room a decade before it won the laptop.
2026: Two Paths — Efficiency Kings and Halo-Class Monsters
The APU market of 2026 has split into two philosophies. On one side, chips like the Ryzen AI 300 series prioritize efficiency: all-day battery life, silent operation, and integrated graphics that handle 1080p gaming gracefully — the premium thin-and-light APU perfected.
On the other side, AMD’s high-end “Halo”-class APUs pursue a wilder goal: integrated graphics that rival mid-range discrete GPUs. With vastly more GPU compute units, wide memory buses, and massive on-package cache, these chips ask a provocative question: if the integrated GPU is this good, who needs the discrete one at all?
For gaming-laptop buyers, the practical upshot is choice. A Halo-class machine offers near-discrete gaming performance without the discrete GPU’s power, heat, or cost penalties — while a mainstream Ryzen AI laptop delivers the best battery life the category has ever seen. The APU is no longer the compromise option. In 2026, it’s increasingly the smart option.
Key Takeaways: 15 Years of the APU
- 2011–2013: The concept years — Llano proves the idea; Trinity and Richland make it affordable.
- 2014–2016: The wilderness years — Kaveri swings big; Carrizo keeps the lights on.
- 2017–2020: The Zen turnaround — Raven Ridge and Renoir make APUs genuinely desirable.
- 2021–2024: The acceleration — Cezanne through Strix Point bring RDNA graphics and AI NPUs.
- 2025–2026: The payoff — Ryzen AI 300 and Halo-class chips fulfill the original promise.
The through-line is patience. AMD spent fifteen years shrinking the gap between “integrated” and “good enough,” then kept going until “good enough” became “great.”
Why did AMD call them “APUs” instead of just CPUs with graphics?
Marketing, but with a point. “APU” signaled that the GPU wasn’t a bonus feature — it was a first-class citizen of the architecture, sharing memory and working alongside the CPU. In the HSA era, AMD genuinely believed heterogeneous computing would reshape software. The name stuck long after the industry moved on, and “Ryzen AI” is its modern successor branding.
